diff --git "a/wandb/run-20220504_142129-1tmxz74i/files/output.log" "b/wandb/run-20220504_142129-1tmxz74i/files/output.log" --- "a/wandb/run-20220504_142129-1tmxz74i/files/output.log" +++ "b/wandb/run-20220504_142129-1tmxz74i/files/output.log" @@ -28592,5 +28592,3075 @@ To disable this warning, you can either: - Avoid using `tokenizers` before the fork if possible - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible +Could not estimate the number of tokens of the input, floating-point operations will not be computed `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████▍ | 4501/9720 [11:47:21<1168:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.625, 'learning_rate': 0.000186161834514522, 'epoch': 1.39} +Could not estimate the number of tokens of the input, floating-point operations will not be computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6791, 'learning_rate': 0.00018612620545528476, 'epoch': 1.39} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.718, 'learning_rate': 0.0001860905763960475, 'epoch': 1.39} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed8:18:49, 805.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████▉ | 4505/9720 [11:47:44<286:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7382, 'learning_rate': 0.00018601931827757308, 'epoch': 1.39} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6261, 'learning_rate': 0.00018598368921833586, 'epoch': 1.39} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7113, 'learning_rate': 0.0001857699148629125, 'epoch': 1.39} +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed6:26:08, 197.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|█████████████████████████████████▉ | 4515/9720 [11:48:34<14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed14:59:15, 10.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|██████████████████████████████████▍ | 4519/9720 [11:48:53<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3762, 'learning_rate': 0.00018552051144825189, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4799, 'learning_rate': 0.00018548488238901464, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6488, 'learning_rate': 0.0001854492533297774, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2391, 'learning_rate': 0.0001854136242705402, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1481, 'learning_rate': 0.00018537799521130296, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:40:10, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▍ | 4525/9720 [11:49:19<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:49:19<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:49:19<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3026, 'learning_rate': 0.00018527110803359125, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2073, 'learning_rate': 0.00018523547897435406, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.8497, 'learning_rate': 0.00018512859179664235, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.0001, 'learning_rate': 0.00018505733367816791, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:27:29, 4.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▌ | 4534/9720 [11:49:51<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:49:51<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:49:51<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.8208, 'learning_rate': 0.00018498607555969342, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:49:51<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:49:51<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.4614, 'learning_rate': 0.0001849504465004562, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:49:51<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.2211, 'learning_rate': 0.00018480793026350728, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.6913, 'learning_rate': 0.0001847723012042701, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.5808, 'learning_rate': 0.00018473667214503284, 'epoch': 1.4} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:03:35, 3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▌ | 4543/9720 [11:50:32<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:12:07, 5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▌ | 4547/9720 [11:50:56<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▌ | 4547/9720 [11:50:56<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8666, 'learning_rate': 0.00018452289778960945, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9719, 'learning_rate': 0.00018448726873037223, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8653, 'learning_rate': 0.00018441601061189774, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7699, 'learning_rate': 0.00018438038155266055, 'epoch': 1.4} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3755, 'learning_rate': 0.0001843447524934233, 'epoch': 1.41} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8257, 'learning_rate': 0.0001843091234341861, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9191, 'learning_rate': 0.00018427349437494884, 'epoch': 1.41} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7524, 'learning_rate': 0.00018416660719723716, 'epoch': 1.41} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:36:12, 5.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▋ | 4564/9720 [11:52:23<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6342, 'learning_rate': 0.00018391720378257658, 'epoch': 1.41} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5529, 'learning_rate': 0.00018388157472333934, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2995, 'learning_rate': 0.00018384594566410212, 'epoch': 1.41} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4198, 'learning_rate': 0.00018381031660486487, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4463, 'learning_rate': 0.00018377468754562763, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4324, 'learning_rate': 0.00018373905848639044, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2894, 'learning_rate': 0.0001837034294271532, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1692, 'learning_rate': 0.00018366780036791595, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.101, 'learning_rate': 0.00018359654224944148, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:49:09, 4.76s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▊ | 4578/9720 [11:53:20<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.0664, 'learning_rate': 0.00018341839695325534, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:29:40, 3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▉ | 4582/9720 [11:53:34<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.6797, 'learning_rate': 0.00018327588071630644, 'epoch': 1.41} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.5416, 'learning_rate': 0.00018324025165706922, 'epoch': 1.41} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.1332, 'learning_rate': 0.00018309773542012032, 'epoch': 1.42} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:07:08, 3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▉ | 4589/9720 [11:53:57<4:28:02, 3.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:53:57<4:28:02, 3.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:53:57<4:28:02, 3.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:28:02, 3.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:28:02, 3.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:28:02, 3.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:28:02, 3.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|██████████████████████████████████▉ | 4591/9720 [11:54:07<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:54:07<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7613, 'learning_rate': 0.00018277707388698522, 'epoch': 1.42} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8745, 'learning_rate': 0.0001827058157685108, 'epoch': 1.42} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5588, 'learning_rate': 0.00018263455765003632, 'epoch': 1.42} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.534, 'learning_rate': 0.00018259892859079908, 'epoch': 1.42} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6018, 'learning_rate': 0.00018256329953156186, 'epoch': 1.42} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6259, 'learning_rate': 0.00018252767047232464, 'epoch': 1.42} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5809, 'learning_rate': 0.00018231389611690125, 'epoch': 1.42} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3967, 'learning_rate': 0.000182278267057664, 'epoch': 1.42} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6, 'learning_rate': 0.00018220700893918957, 'epoch': 1.42} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2941, 'learning_rate': 0.00018210012176147786, 'epoch': 1.42} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6056, 'learning_rate': 0.00018199323458376618, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4139, 'learning_rate': 0.00018195760552452896, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3395, 'learning_rate': 0.00018192197646529172, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.0048, 'learning_rate': 0.00018185071834681728, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.9755, 'learning_rate': 0.00018181508928758004, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.0569, 'learning_rate': 0.00018177946022834282, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.259, 'learning_rate': 0.00018174383116910557, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.9022, 'learning_rate': 0.0001816369439913939, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.8732, 'learning_rate': 0.00018160131493215667, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.675, 'learning_rate': 0.0001815300568136822, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.6791, 'learning_rate': 0.0001813875405767333, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.3058, 'learning_rate': 0.00018135191151749606, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:17:39, 4.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|███████████████████████████████████▎ | 4638/9720 [11:57:41<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:57:41<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:57:41<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 1.9923, 'learning_rate': 0.0001812806533990216, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:57:41<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.6725, 'learning_rate': 0.00018120939528054716, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.3977, 'learning_rate': 0.00018117376622130992, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.1266, 'learning_rate': 0.00018110250810283546, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9568, 'learning_rate': 0.00018103124998436102, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.945, 'learning_rate': 0.00018099562092512377, 'epoch': 1.43} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.0293, 'learning_rate': 0.00018095999186588656, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5411, 'learning_rate': 0.0001809243628066493, 'epoch': 1.43} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6484, 'learning_rate': 0.00018085310468817487, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9226, 'learning_rate': 0.00018081747562893763, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:36:42, 3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|███████████████████████████████████▍ | 4653/9720 [11:59:07<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [11:59:07<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5263, 'learning_rate': 0.0001807462175104632, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9934, 'learning_rate': 0.00018071058845122595, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.592, 'learning_rate': 0.00018063933033275148, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6615, 'learning_rate': 0.00018049681409580259, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8338, 'learning_rate': 0.00018046118503656534, 'epoch': 1.44} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5183, 'learning_rate': 0.0001803899269180909, 'epoch': 1.44} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3088, 'learning_rate': 0.00018021178162190476, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2715, 'learning_rate': 0.00018014052350343027, 'epoch': 1.44} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4788, 'learning_rate': 0.00018010489444419305, 'epoch': 1.44} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.242, 'learning_rate': 0.0001800692653849558, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2025, 'learning_rate': 0.0001800336363257186, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.9685, 'learning_rate': 0.00017999800726648137, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.9887, 'learning_rate': 0.0001799267491480069, 'epoch': 1.44} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.8831, 'learning_rate': 0.0001798911200887697, 'epoch': 1.44} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.9632, 'learning_rate': 0.0001797486038518208, 'epoch': 1.45} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.6556, 'learning_rate': 0.00017971297479258354, 'epoch': 1.45} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.6813, 'learning_rate': 0.0001796773457333463, 'epoch': 1.45} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.3847, 'learning_rate': 0.00017964171667410908, 'epoch': 1.45} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:55, 5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|███████████████████████████████████▋ | 4686/9720 [12:01:27<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:01:27<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:01:27<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.3641, 'learning_rate': 0.0001795704585556346, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:01:27<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:01:27<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.7918, 'learning_rate': 0.0001794279423186857, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.4306, 'learning_rate': 0.00017939231325944844, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.145, 'learning_rate': 0.000179321055140974, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.3339, 'learning_rate': 0.00017924979702249954, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.0556, 'learning_rate': 0.00017921416796326233, 'epoch': 1.45} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:37:45, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|███████████████████████████████████▊ | 4700/9720 [12:02:42<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8096, 'learning_rate': 0.00017907165172631343, 'epoch': 1.45} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7621, 'learning_rate': 0.00017903602266707618, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7445, 'learning_rate': 0.00017896476454860172, 'epoch': 1.45} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7266, 'learning_rate': 0.00017892913548936447, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.69, 'learning_rate': 0.00017889350643012728, 'epoch': 1.45} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6179, 'learning_rate': 0.0001788222483116528, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:14:51, 5.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|███████████████████████████████████▊ | 4709/9720 [12:03:28<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4449, 'learning_rate': 0.00017875099019317833, 'epoch': 1.45} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:58:26, 5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|███████████████████████████████████▊ | 4712/9720 [12:03:43<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:03:43<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:03:43<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4747, 'learning_rate': 0.000178537215837755, 'epoch': 1.46} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.475, 'learning_rate': 0.0001784659577192805, 'epoch': 1.46} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2605, 'learning_rate': 0.00017843032866004328, 'epoch': 1.46} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1157, 'learning_rate': 0.00017835907054156882, 'epoch': 1.46} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:41:36, 4.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████▉ | 4722/9720 [12:04:26<5:43:58, 4.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:04:26<5:43:58, 4.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:04:26<5:43:58, 4.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████▉ | 4723/9720 [12:04:30<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:04:30<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:04:30<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.7454, 'learning_rate': 0.00017814529618614546, 'epoch': 1.46} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:36:53, 4.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████ | 4730/9720 [12:04:56<5:08:08, 3.71s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:08, 3.71s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:08, 3.71s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████ | 4731/9720 [12:04:59<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:04:59<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:04:59<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.6924, 'learning_rate': 0.00017793152183072207, 'epoch': 1.46} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.3836, 'learning_rate': 0.00017786026371224763, 'epoch': 1.46} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:02:36, 3.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████ | 4738/9720 [12:05:22<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████ | 4738/9720 [12:05:22<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.6222, 'learning_rate': 0.00017764648935682424, 'epoch': 1.46} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.401, 'learning_rate': 0.00017761086029758702, 'epoch': 1.46} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.027, 'learning_rate': 0.00017753960217911256, 'epoch': 1.46} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.0846, 'learning_rate': 0.0001774683440606381, 'epoch': 1.46} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:25:22, 3.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▏ | 4748/9720 [12:06:19<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:06:19<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:06:19<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:06:19<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9156, 'learning_rate': 0.00017729019876445195, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8832, 'learning_rate': 0.0001772545697052147, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5671, 'learning_rate': 0.00017721894064597751, 'epoch': 1.47} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6431, 'learning_rate': 0.00017718331158674027, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:08:36, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▏ | 4755/9720 [12:06:57<7:21:29, 5.34s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:06:57<7:21:29, 5.34s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:21:29, 5.34s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:21:29, 5.34s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6364, 'learning_rate': 0.0001771120534682658, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:21:29, 5.34s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:21:29, 5.34s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▏ | 4757/9720 [12:07:07<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6083, 'learning_rate': 0.00017700516629055412, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6549, 'learning_rate': 0.00017696953723131688, 'epoch': 1.47} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4365, 'learning_rate': 0.00017686265005360523, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3905, 'learning_rate': 0.00017679139193513073, 'epoch': 1.47} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3342, 'learning_rate': 0.00017672013381665627, 'epoch': 1.47} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:08:40, 5.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▎ | 4768/9720 [12:07:58<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.0734, 'learning_rate': 0.00017664887569818184, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1943, 'learning_rate': 0.0001766132466389446, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2742, 'learning_rate': 0.0001765776175797074, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2435, 'learning_rate': 0.00017654198852047015, 'epoch': 1.47} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2274, 'learning_rate': 0.0001765063594612329, 'epoch': 1.47} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1006, 'learning_rate': 0.0001764707304019957, 'epoch': 1.47} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:06:06, 4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▎ | 4776/9720 [12:08:30<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▎ | 4776/9720 [12:08:30<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:31:32, 4.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▍ | 4782/9720 [12:08:52<4:58:35, 3.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:35, 3.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:35, 3.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:35, 3.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:35, 3.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:35, 3.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.7093, 'learning_rate': 0.0001761144398096234, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:35, 3.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▍ | 4785/9720 [12:09:02<4:42:50, 3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:02<4:42:50, 3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:02<4:42:50, 3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.4928, 'learning_rate': 0.0001760431816911489, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:02<4:42:50, 3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▍ | 4787/9720 [12:09:08<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:08<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:08<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.2693, 'learning_rate': 0.00017597192357267447, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:08<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:08<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.2265, 'learning_rate': 0.00017593629451343723, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:08<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 1.8167, 'learning_rate': 0.0001759006654542, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:08<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:08<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.4275, 'learning_rate': 0.0001758650363949628, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.1344, 'learning_rate': 0.0001757225201580139, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:26:40, 3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▌ | 4796/9720 [12:09:57<8:12:32, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:57<8:12:32, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:57<8:12:32, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:09:57<8:12:32, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:32, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:12:32, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:32, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:12:32, 6.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▌ | 4798/9720 [12:10:09<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:10:09<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:10:09<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6262, 'learning_rate': 0.00017558000392106494, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:10:09<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6807, 'learning_rate': 0.00017554437486182775, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6051, 'learning_rate': 0.0001755087458025905, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:06:01, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▌ | 4802/9720 [12:10:32<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:10:32<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:10:32<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7049, 'learning_rate': 0.00017543748768411604, 'epoch': 1.48} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8645, 'learning_rate': 0.0001754018586248788, 'epoch': 1.48} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6909, 'learning_rate': 0.0001753662295656416, 'epoch': 1.48} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:42:13, 5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▌ | 4808/9720 [12:11:02<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:11:02<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:11:02<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:08, 5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|████████████████████████████████████▋ | 4811/9720 [12:11:16<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3999, 'learning_rate': 0.00017504556803250653, 'epoch': 1.49} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2533, 'learning_rate': 0.00017493868085479482, 'epoch': 1.49} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3549, 'learning_rate': 0.0001748674227363204, 'epoch': 1.49} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4212, 'learning_rate': 0.00017483179367708314, 'epoch': 1.49} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3165, 'learning_rate': 0.00017479616461784592, 'epoch': 1.49} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1172, 'learning_rate': 0.00017476053555860868, 'epoch': 1.49} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1252, 'learning_rate': 0.0001747249064993715, 'epoch': 1.49} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1517, 'learning_rate': 0.00017468927744013424, 'epoch': 1.49} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.9612, 'learning_rate': 0.000174653648380897, 'epoch': 1.49} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.8512, 'learning_rate': 0.0001745467612031853, 'epoch': 1.49} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:38:35, 4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████▊ | 4830/9720 [12:12:35<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.4531, 'learning_rate': 0.00017436861590699914, 'epoch': 1.49} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:05:30, 3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████▊ | 4836/9720 [12:12:55<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 1.9537, 'learning_rate': 0.0001741548415515758, 'epoch': 1.49} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 1.9982, 'learning_rate': 0.00017411921249233856, 'epoch': 1.49} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.7958, 'learning_rate': 0.00017408358343310132, 'epoch': 1.49} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:29:21, 3.31s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████▊ | 4842/9720 [12:13:22<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.4349, 'learning_rate': 0.00017401232531462688, 'epoch': 1.49} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.724, 'learning_rate': 0.00017390543813691517, 'epoch': 1.5} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7774, 'learning_rate': 0.00017383418001844074, 'epoch': 1.5} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9569, 'learning_rate': 0.0001737985509592035, 'epoch': 1.5} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:59:59, 5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████▉ | 4850/9720 [12:14:10<8:03:15, 5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:14:10<8:03:15, 5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:14:10<8:03:15, 5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:14:10<8:03:15, 5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6954, 'learning_rate': 0.00017372729284072903, 'epoch': 1.5} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:14:10<8:03:15, 5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:03:15, 5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:03:15, 5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4838, 'learning_rate': 0.00017369166378149184, 'epoch': 1.5} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:03:15, 5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████▉ | 4853/9720 [12:14:27<7:46:33, 5.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████▉ | 4853/9720 [12:14:27<7:46:33, 5.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:14:27<7:46:33, 5.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:14:27<7:46:33, 5.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:14:27<7:46:33, 5.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████▉ | 4854/9720 [12:14:33<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:14:33<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8501, 'learning_rate': 0.00017358477660378013, 'epoch': 1.5} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8488, 'learning_rate': 0.00017354914754454288, 'epoch': 1.5} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4773, 'learning_rate': 0.0001735135184853057, 'epoch': 1.5} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6138, 'learning_rate': 0.00017340663130759398, 'epoch': 1.5} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7711, 'learning_rate': 0.0001732997441298823, 'epoch': 1.5} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1394, 'learning_rate': 0.00017312159883369616, 'epoch': 1.5} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1978, 'learning_rate': 0.00017294345353751, 'epoch': 1.5} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2887, 'learning_rate': 0.00017290782447827277, 'epoch': 1.5} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3596, 'learning_rate': 0.00017287219541903552, 'epoch': 1.5} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1373, 'learning_rate': 0.00017283656635979833, 'epoch': 1.5} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:41:21, 5.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|█████████████████████████████████████▏ | 4877/9720 [12:16:20<5:36:33, 4.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:33, 4.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:36:33, 4.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|█████████████████████████████████████▏ | 4878/9720 [12:16:24<5:32:07, 4.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:24<5:32:07, 4.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:24<5:32:07, 4.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|█████████████████████████████████████▏ | 4879/9720 [12:16:28<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:28<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:28<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.4158, 'learning_rate': 0.0001725871629451377, 'epoch': 1.51} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:27:25, 4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|█████████████████████████████████████▏ | 4884/9720 [12:16:46<4:57:18, 3.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:57:18, 3.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:57:18, 3.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:57:18, 3.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:57:18, 3.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.1929, 'learning_rate': 0.00017248027576742604, 'epoch': 1.51} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:57:18, 3.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:57:18, 3.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|█████████████████████████████████████▏ | 4887/9720 [12:16:56<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:56<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:56<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 1.9975, 'learning_rate': 0.00017240901764895155, 'epoch': 1.51} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:56<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.1176, 'learning_rate': 0.00017237338858971436, 'epoch': 1.51} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:56<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:16:56<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.5649, 'learning_rate': 0.0001723021304712399, 'epoch': 1.51} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9344, 'learning_rate': 0.0001720883561158165, 'epoch': 1.51} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:38:44, 3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|█████████████████████████████████████▎ | 4901/9720 [12:18:18<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:18:18<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:18:18<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:18:18<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8397, 'learning_rate': 0.00017191021081963036, 'epoch': 1.51} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:18:18<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:18:18<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:18:18<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7436, 'learning_rate': 0.0001718389527011559, 'epoch': 1.51} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5739, 'learning_rate': 0.00017176769458268143, 'epoch': 1.51} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7032, 'learning_rate': 0.00017173206552344422, 'epoch': 1.51} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:54:11, 5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|█████████████████████████████████████▎ | 4908/9720 [12:18:54<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3831, 'learning_rate': 0.00017166080740496975, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:54:13, 5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▍ | 4914/9720 [12:19:23<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▍ | 4914/9720 [12:19:23<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3821, 'learning_rate': 0.0001713757749310719, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3301, 'learning_rate': 0.00017126888775336024, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.0893, 'learning_rate': 0.000171233258694123, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.0963, 'learning_rate': 0.00017119762963488575, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.8193, 'learning_rate': 0.00017112637151641132, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2229, 'learning_rate': 0.0001710907424571741, 'epoch': 1.52} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1508, 'learning_rate': 0.00017105511339793685, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.8698, 'learning_rate': 0.00017098385527946242, 'epoch': 1.52} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:23:43, 4.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▌ | 4930/9720 [12:20:27<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.4449, 'learning_rate': 0.00017080570998327627, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.5278, 'learning_rate': 0.00017073445186480178, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.3266, 'learning_rate': 0.0001706988228055646, 'epoch': 1.52} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.1352, 'learning_rate': 0.00017066319374632735, 'epoch': 1.52} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.1713, 'learning_rate': 0.00017062756468709013, 'epoch': 1.52} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:58:23, 3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▌ | 4942/9720 [12:21:14<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▌ | 4942/9720 [12:21:14<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9741, 'learning_rate': 0.0001703781612724295, 'epoch': 1.53} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7937, 'learning_rate': 0.0001703425322131923, 'epoch': 1.53} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:56:41, 5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▋ | 4949/9720 [12:21:57<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:21:57<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:21:57<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:21:57<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.8489, 'learning_rate': 0.00017016438691700613, 'epoch': 1.53} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.6257, 'learning_rate': 0.00017009312879853167, 'epoch': 1.53} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.5836, 'learning_rate': 0.00017005749973929445, 'epoch': 1.53} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:44:26, 5.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▋ | 4955/9720 [12:22:30<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.819, 'learning_rate': 0.00016998624162081999, 'epoch': 1.53} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.7567, 'learning_rate': 0.00016987935444310833, 'epoch': 1.53} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.2229, 'learning_rate': 0.00016966558008768494, 'epoch': 1.53} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:06:46, 5.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▊ | 4966/9720 [12:23:22<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3442, 'learning_rate': 0.00016959432196921048, 'epoch': 1.53} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.1062, 'learning_rate': 0.000169523063850736, 'epoch': 1.53} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.4486, 'learning_rate': 0.0001694874347914988, 'epoch': 1.53} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.055, 'learning_rate': 0.00016945180573226155, 'epoch': 1.53} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.3777, 'learning_rate': 0.00016941617667302433, 'epoch': 1.53} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.9605, 'learning_rate': 0.0001693805476137871, 'epoch': 1.53} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<6:03:19, 4.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▉ | 4977/9720 [12:24:07<5:10:34, 3.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:24:07<5:10:34, 3.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:24:07<5:10:34, 3.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▉ | 4978/9720 [12:24:11<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.7263, 'learning_rate': 0.00016909551513988926, 'epoch': 1.54} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:08:14, 3.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▉ | 4982/9720 [12:24:25<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:24:25<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:24:25<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.4867, 'learning_rate': 0.00016898862796217758, 'epoch': 1.54} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.4298, 'learning_rate': 0.00016895299890294033, 'epoch': 1.54} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 2.0809, 'learning_rate': 0.00016884611172522868, 'epoch': 1.54} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<4:44:13, 3.60s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████▉ | 4991/9720 [12:24:57<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`... [12:24:57<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 3.9955, 'learning_rate': 0.0001686323373698053, 'epoch': 1.54} +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +{'loss': 4.1576, 'learning_rate': 0.00016852545019209358, 'epoch': 1.54} +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<5:52:25, 4.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|██████████████████████████████████████ | 4999/9720 [12:25:48<7:45:38, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:38, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:45:38, 5.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...hEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|██████████████████████████████████████ | 5000/9720 [12:25:54<7:49:42, 5.97s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|██████████████████████████████████████ | 5000/9720 [12:25:54<7:49:42, 5.97s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 0%| | 2/1845 [00:01<22:52, 1.34it/s]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 0%|▏ | 3/1845 [00:03<33:08, 1.08s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 0%|▏ | 5/1845 [00:05<39:25, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 0%|▎ | 6/1845 [00:07<41:23, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 0%|▎ | 8/1845 [00:10<41:36, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 0%|▍ | 9/1845 [00:11<41:02, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▍ | 11/1845 [00:14<42:04, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▌ | 12/1845 [00:15<44:17, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▌ | 13/1845 [00:17<44:18, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▋ | 15/1845 [00:20<44:01, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▋ | 16/1845 [00:21<43:14, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▊ | 18/1845 [00:24<42:01, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▊ | 19/1845 [00:25<43:26, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▉ | 21/1845 [00:28<41:44, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|▉ | 22/1845 [00:29<41:20, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|█ | 23/1845 [00:31<41:52, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|█ | 25/1845 [00:34<44:18, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|█▏ | 26/1845 [00:35<43:01, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 1%|█▏ | 27/1845 [00:37<44:39, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▎ | 29/1845 [00:40<45:55, 1.52s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▎ | 30/1845 [00:41<46:30, 1.54s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▎ | 31/1845 [00:43<45:43, 1.51s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▍ | 33/1845 [00:46<43:40, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▍ | 34/1845 [00:47<43:48, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▌ | 36/1845 [00:50<44:50, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▌ | 37/1845 [00:51<43:45, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▋ | 38/1845 [00:53<44:44, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▊ | 40/1845 [00:56<43:08, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▊ | 41/1845 [00:57<43:10, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▉ | 43/1845 [01:00<40:45, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|█▉ | 44/1845 [01:01<41:03, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 2%|██ | 46/1845 [01:04<41:07, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██ | 47/1845 [01:05<40:48, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▏ | 49/1845 [01:08<43:12, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▏ | 50/1845 [01:10<43:27, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▎ | 52/1845 [01:13<41:55, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▎ | 53/1845 [01:14<41:58, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▎ | 54/1845 [01:16<42:46, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▍ | 56/1845 [01:18<41:43, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▌ | 57/1845 [01:19<39:39, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▌ | 59/1845 [01:22<40:40, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▋ | 60/1845 [01:24<40:52, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▋ | 62/1845 [01:27<41:57, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▊ | 63/1845 [01:28<41:31, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 3%|██▊ | 64/1845 [01:29<42:05, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|██▉ | 66/1845 [01:32<41:57, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|██▉ | 68/1845 [01:35<41:29, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███ | 69/1845 [01:36<42:09, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███ | 70/1845 [01:38<43:08, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▏ | 72/1845 [01:41<41:02, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▏ | 73/1845 [01:42<41:19, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▏ | 74/1845 [01:43<41:20, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▎ | 76/1845 [01:46<41:05, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▍ | 77/1845 [01:48<40:38, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▍ | 79/1845 [01:50<41:19, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▌ | 80/1845 [01:52<42:14, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▌ | 82/1845 [01:55<43:44, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 4%|███▋ | 83/1845 [01:56<41:16, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|███▋ | 85/1845 [01:59<40:11, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|███▊ | 86/1845 [02:01<42:39, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|███▊ | 87/1845 [02:02<42:40, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|███▉ | 89/1845 [02:05<41:08, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|███▉ | 90/1845 [02:06<40:57, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|███▉ | 91/1845 [02:08<40:29, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|████ | 93/1845 [02:11<42:34, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|████▏ | 94/1845 [02:12<41:32, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|████▏ | 96/1845 [02:15<41:36, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|████▎ | 97/1845 [02:16<41:03, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|████▎ | 99/1845 [02:19<39:35, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 5%|████▎ | 100/1845 [02:20<38:03, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▍ | 102/1845 [02:23<40:42, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▍ | 103/1845 [02:24<39:54, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▌ | 104/1845 [02:26<41:30, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▌ | 106/1845 [02:28<39:00, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▋ | 107/1845 [02:30<39:15, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▋ | 109/1845 [02:33<40:38, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▊ | 110/1845 [02:34<41:34, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▊ | 111/1845 [02:36<42:52, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▉ | 113/1845 [02:39<42:41, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|████▉ | 114/1845 [02:40<42:56, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|█████ | 116/1845 [02:43<39:50, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|█████ | 117/1845 [02:44<39:19, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 6%|█████▏ | 119/1845 [02:47<41:40, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▏ | 120/1845 [02:49<41:49, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▏ | 121/1845 [02:50<41:59, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▎ | 123/1845 [02:53<41:09, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▍ | 124/1845 [02:54<38:55, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▍ | 126/1845 [02:57<39:24, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▌ | 127/1845 [02:58<38:47, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▌ | 129/1845 [03:01<38:28, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▋ | 130/1845 [03:02<39:29, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▋ | 132/1845 [03:05<40:45, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▊ | 133/1845 [03:07<40:12, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▊ | 134/1845 [03:08<41:06, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▉ | 136/1845 [03:11<41:09, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 7%|█████▉ | 137/1845 [03:12<40:46, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████ | 139/1845 [03:15<40:25, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████ | 140/1845 [03:17<40:14, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████ | 141/1845 [03:18<41:51, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▏ | 143/1845 [03:21<41:23, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▏ | 144/1845 [03:23<42:10, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▎ | 146/1845 [03:25<39:05, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▎ | 147/1845 [03:27<39:05, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▍ | 149/1845 [03:29<39:47, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▌ | 150/1845 [03:31<39:20, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▌ | 152/1845 [03:34<40:03, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▋ | 153/1845 [03:35<39:51, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▋ | 155/1845 [03:38<38:13, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 8%|██████▊ | 156/1845 [03:39<39:26, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|██████▊ | 158/1845 [03:42<36:59, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|██████▉ | 159/1845 [03:43<36:58, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|██████▉ | 161/1845 [03:46<38:02, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████ | 162/1845 [03:47<38:16, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████ | 163/1845 [03:49<40:57, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████▏ | 165/1845 [03:52<40:14, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████▏ | 166/1845 [03:53<38:13, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████▎ | 168/1845 [03:56<39:55, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████▎ | 169/1845 [03:57<40:34, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████▍ | 171/1845 [04:00<38:53, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████▍ | 172/1845 [04:01<37:46, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████▌ | 174/1845 [04:04<38:18, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 9%|███████▌ | 175/1845 [04:05<38:02, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|███████▋ | 177/1845 [04:08<37:14, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|███████▋ | 178/1845 [04:09<36:26, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|███████▊ | 180/1845 [04:12<36:04, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|███████▊ | 181/1845 [04:13<36:59, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|███████▉ | 182/1845 [04:15<38:26, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|███████▉ | 184/1845 [04:18<39:50, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|████████ | 185/1845 [04:19<38:11, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|████████ | 187/1845 [04:22<40:53, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|████████▏ | 188/1845 [04:23<39:43, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|████████▏ | 189/1845 [04:25<38:33, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|████████▎ | 191/1845 [04:28<39:22, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 10%|████████▎ | 192/1845 [04:29<37:13, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▍ | 194/1845 [04:32<37:27, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▍ | 195/1845 [04:33<36:47, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▌ | 197/1845 [04:36<37:47, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▌ | 198/1845 [04:37<38:48, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▋ | 200/1845 [04:40<38:06, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▋ | 201/1845 [04:42<39:03, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▊ | 202/1845 [04:43<40:04, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▊ | 204/1845 [04:46<38:16, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▉ | 205/1845 [04:47<40:01, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|████████▉ | 207/1845 [04:50<39:52, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|█████████ | 208/1845 [04:52<37:48, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|█████████ | 209/1845 [04:53<37:06, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|█████████▏ | 211/1845 [04:56<39:57, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 11%|█████████▏ | 212/1845 [04:57<38:27, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▎ | 214/1845 [05:00<40:01, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▎ | 215/1845 [05:02<40:22, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▎ | 216/1845 [05:03<41:10, 1.52s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▍ | 218/1845 [05:06<39:12, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▍ | 219/1845 [05:08<40:39, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▌ | 221/1845 [05:10<37:30, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▋ | 222/1845 [05:12<38:03, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▋ | 223/1845 [05:13<38:16, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▊ | 225/1845 [05:16<39:06, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▊ | 226/1845 [05:17<36:31, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▉ | 228/1845 [05:20<37:39, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 12%|█████████▉ | 229/1845 [05:22<38:45, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████ | 231/1845 [05:24<36:52, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████ | 232/1845 [05:26<35:25, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▏ | 234/1845 [05:28<37:28, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▏ | 235/1845 [05:30<36:37, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▎ | 237/1845 [05:33<37:03, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▎ | 238/1845 [05:34<37:44, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▍ | 240/1845 [05:37<36:17, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▍ | 241/1845 [05:38<36:52, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▍ | 242/1845 [05:40<37:38, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▌ | 244/1845 [05:42<35:47, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▌ | 245/1845 [05:44<36:26, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▋ | 247/1845 [05:46<37:33, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 13%|██████████▊ | 248/1845 [05:48<36:56, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|██████████▊ | 250/1845 [05:50<35:54, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|██████████▉ | 251/1845 [05:52<35:18, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|██████████▉ | 253/1845 [05:55<38:32, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████ | 254/1845 [05:56<38:14, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████ | 255/1845 [05:58<39:04, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████▏ | 257/1845 [06:01<38:46, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████▏ | 258/1845 [06:02<38:53, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████▎ | 260/1845 [06:05<37:09, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████▎ | 261/1845 [06:06<37:16, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████▎ | 262/1845 [06:08<38:44, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████▍ | 264/1845 [06:11<39:47, 1.51s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████▍ | 265/1845 [06:12<37:48, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 14%|███████████▌ | 267/1845 [06:15<37:05, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|███████████▌ | 268/1845 [06:17<37:43, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|███████████▋ | 269/1845 [06:18<37:49, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|███████████▊ | 271/1845 [06:21<38:47, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|███████████▊ | 272/1845 [06:23<38:24, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|███████████▉ | 274/1845 [06:25<36:56, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|███████████▉ | 275/1845 [06:27<37:28, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|███████████▉ | 276/1845 [06:28<37:20, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|████████████ | 278/1845 [06:31<37:00, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|████████████ | 279/1845 [06:32<37:03, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|████████████▏ | 281/1845 [06:35<36:44, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|████████████▏ | 282/1845 [06:37<37:34, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|████████████▎ | 283/1845 [06:38<37:58, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 15%|████████████▎ | 285/1845 [06:41<36:04, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▍ | 286/1845 [06:42<36:42, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▍ | 288/1845 [06:45<36:37, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▌ | 289/1845 [06:47<36:08, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▌ | 291/1845 [06:49<37:23, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▋ | 292/1845 [06:51<37:24, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▋ | 294/1845 [06:54<35:49, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▊ | 295/1845 [06:55<35:41, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▉ | 297/1845 [06:58<36:06, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▉ | 298/1845 [06:59<37:35, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|████████████▉ | 299/1845 [07:01<35:59, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|█████████████ | 301/1845 [07:03<36:31, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|█████████████ | 302/1845 [07:05<36:48, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 16%|█████████████▏ | 304/1845 [07:08<36:36, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▏ | 305/1845 [07:09<34:28, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▎ | 306/1845 [07:10<35:56, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▎ | 308/1845 [07:13<36:10, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▍ | 309/1845 [07:15<35:57, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▍ | 311/1845 [07:17<35:09, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▌ | 312/1845 [07:19<36:07, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▌ | 314/1845 [07:22<36:29, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▋ | 315/1845 [07:23<36:36, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▋ | 316/1845 [07:25<35:38, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▊ | 318/1845 [07:27<34:20, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▉ | 320/1845 [07:30<34:30, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▉ | 321/1845 [07:31<35:45, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 17%|█████████████▉ | 322/1845 [07:33<36:20, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████ | 324/1845 [07:36<35:37, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████ | 325/1845 [07:37<34:47, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▏ | 327/1845 [07:40<35:05, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▏ | 328/1845 [07:41<33:42, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▎ | 330/1845 [07:44<34:03, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▎ | 331/1845 [07:45<35:24, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▍ | 332/1845 [07:47<36:32, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|��█████████████▍ | 334/1845 [07:50<36:49, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▌ | 335/1845 [07:51<35:13, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▌ | 337/1845 [07:54<35:07, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▋ | 338/1845 [07:55<36:02, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▋ | 339/1845 [07:57<36:34, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 18%|██████████████▊ | 341/1845 [08:00<35:09, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|██████████████▊ | 343/1845 [08:02<34:02, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|██████████████▉ | 344/1845 [08:03<32:48, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|██████████████▉ | 345/1845 [08:05<34:50, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████ | 347/1845 [08:08<34:20, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████ | 348/1845 [08:09<34:10, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████▏ | 350/1845 [08:12<36:07, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████▏ | 351/1845 [08:14<36:30, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████▎ | 352/1845 [08:15<36:25, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████▎ | 354/1845 [08:18<36:13, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████▍ | 355/1845 [08:19<37:08, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████▍ | 357/1845 [08:22<36:00, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████▌ | 358/1845 [08:24<35:21, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 19%|███████████████▌ | 359/1845 [08:25<33:38, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|███████████████▋ | 361/1845 [08:28<36:10, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|███████████████▋ | 362/1845 [08:29<34:41, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|███████████████▊ | 364/1845 [08:32<34:19, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|███████████████▊ | 366/1845 [08:35<34:00, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|███████████████▉ | 367/1845 [08:36<35:03, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|███████████████▉ | 368/1845 [08:38<34:16, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|████████████████ | 370/1845 [08:40<33:07, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|████████████████ | 371/1845 [08:42<33:24, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|████████████████▏ | 372/1845 [08:43<34:49, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|████████████████▏ | 374/1845 [08:46<33:24, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|████████████████▎ | 375/1845 [08:47<34:47, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|████████████████▎ | 377/1845 [08:50<32:49, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 20%|████████████████▍ | 378/1845 [08:52<34:24, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▍ | 379/1845 [08:53<35:55, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▌ | 381/1845 [08:56<37:02, 1.52s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▌ | 382/1845 [08:58<36:15, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▌ | 383/1845 [08:59<35:53, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▋ | 385/1845 [09:02<35:20, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▋ | 386/1845 [09:03<34:35, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▊ | 388/1845 [09:06<34:02, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▊ | 389/1845 [09:07<32:11, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▉ | 391/1845 [09:10<32:09, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|████████████████▉ | 392/1845 [09:11<32:23, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|█████████████████ | 394/1845 [09:14<33:46, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 21%|█████████████████▏ | 395/1845 [09:16<34:34, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▏ | 397/1845 [09:19<35:38, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▎ | 398/1845 [09:20<34:55, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▎ | 399/1845 [09:22<33:58, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▍ | 401/1845 [09:25<35:13, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▍ | 402/1845 [09:26<34:39, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▌ | 404/1845 [09:29<35:32, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▌ | 405/1845 [09:31<36:30, 1.52s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▌ | 406/1845 [09:32<36:01, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▋ | 408/1845 [09:35<33:13, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▋ | 409/1845 [09:36<33:29, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▊ | 411/1845 [09:39<33:51, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▊ | 412/1845 [09:40<34:18, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▉ | 413/1845 [09:42<34:54, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 22%|█████████████████▉ | 415/1845 [09:45<33:56, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████ | 416/1845 [09:46<34:17, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████ | 418/1845 [09:49<32:00, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▏ | 419/1845 [09:50<31:04, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▎ | 421/1845 [09:53<34:50, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▎ | 422/1845 [09:55<34:54, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▍ | 424/1845 [09:57<32:20, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▍ | 425/1845 [09:59<33:02, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▍ | 426/1845 [10:00<32:18, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▌ | 428/1845 [10:03<33:27, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▌ | 429/1845 [10:04<33:08, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▋ | 431/1845 [10:07<32:04, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 23%|██████████████████▋ | 432/1845 [10:08<31:51, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|██████████████████▊ | 434/1845 [10:11<32:57, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|██████████████████▊ | 435/1845 [10:13<32:49, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|██████████████████▉ | 436/1845 [10:14<34:04, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|██████████████████▉ | 438/1845 [10:17<33:16, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████ | 440/1845 [10:19<30:51, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████ | 441/1845 [10:21<31:18, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████▏ | 442/1845 [10:22<31:20, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████▎ | 444/1845 [10:25<33:22, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████▎ | 445/1845 [10:26<32:39, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████▍ | 447/1845 [10:29<31:58, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████▍ | 448/1845 [10:31<32:56, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████▍ | 449/1845 [10:32<30:53, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 24%|███████████████████▌ | 451/1845 [10:35<35:18, 1.52s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|███████████████████▋ | 453/1845 [10:38<32:19, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|███████████████████▋ | 454/1845 [10:39<32:13, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|███████████████████▋ | 455/1845 [10:41<33:23, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|███████████████████▊ | 457/1845 [10:43<31:54, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|███████████████████▊ | 458/1845 [10:45<32:07, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|███████████████████▉ | 460/1845 [10:47<31:05, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|███████████████████▉ | 461/1845 [10:49<32:33, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|████████████████████ | 462/1845 [10:51<32:47, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|████████████████████ | 464/1845 [10:53<32:03, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|████████████████████▏ | 465/1845 [10:55<33:21, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|████████████████████▏ | 467/1845 [10:57<30:37, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|████████████████████▎ | 468/1845 [10:59<31:54, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 25%|████████████████████▎ | 469/1845 [11:00<32:24, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▍ | 471/1845 [11:03<33:46, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▍ | 472/1845 [11:05<33:29, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▌ | 474/1845 [11:07<32:07, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▌ | 475/1845 [11:09<32:05, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▋ | 477/1845 [11:12<31:54, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▋ | 478/1845 [11:13<32:13, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▊ | 479/1845 [11:15<33:03, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▊ | 481/1845 [11:18<33:01, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▉ | 482/1845 [11:19<30:00, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|████████████████████▉ | 484/1845 [11:21<28:46, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|█████████████████████ | 486/1845 [11:24<29:09, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 26%|█████████████████████ | 487/1845 [11:25<31:23, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▏ | 489/1845 [11:28<29:46, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▏ | 490/1845 [11:29<29:06, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▎ | 492/1845 [11:32<29:44, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▍ | 493/1845 [11:33<30:47, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▍ | 495/1845 [11:36<30:15, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▌ | 496/1845 [11:37<29:37, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▌ | 498/1845 [11:40<28:53, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▋ | 499/1845 [11:41<28:56, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▋ | 501/1845 [11:44<30:15, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▊ | 502/1845 [11:45<32:06, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▊ | 503/1845 [11:47<33:02, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▉ | 505/1845 [11:50<32:29, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 27%|█████████████████████▉ | 506/1845 [11:51<32:02, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████ | 508/1845 [11:54<29:55, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████ | 509/1845 [11:55<30:39, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▏ | 511/1845 [11:58<31:33, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▏ | 512/1845 [12:00<32:31, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▏ | 513/1845 [12:01<32:21, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▎ | 515/1845 [12:04<32:42, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▎ | 516/1845 [12:05<32:00, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▍ | 518/1845 [12:08<30:10, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▌ | 519/1845 [12:09<29:55, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▌ | 521/1845 [12:12<29:10, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▋ | 522/1845 [12:13<29:46, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▋ | 524/1845 [12:16<30:59, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 28%|██████████████████████▊ | 525/1845 [12:17<29:18, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|██████████████████████▊ | 527/1845 [12:20<29:34, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|██████████████████████▉ | 528/1845 [12:22<30:45, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|██████████████████████▉ | 530/1845 [12:24<29:25, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████ | 531/1845 [12:26<29:55, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████ | 532/1845 [12:27<29:47, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████▏ | 534/1845 [12:30<31:47, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████▏ | 535/1845 [12:32<31:10, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████▏ | 536/1845 [12:33<31:51, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████▎ | 538/1845 [12:36<31:38, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████▎ | 539/1845 [12:38<32:19, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|█████���█████████████████▍ | 541/1845 [12:40<32:05, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████▌ | 542/1845 [12:42<30:56, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 29%|███████████████████████▌ | 544/1845 [12:44<29:51, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|███████████████████████▋ | 545/1845 [12:46<31:21, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|███████████████████████▋ | 547/1845 [12:48<28:58, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|███████████████████████▊ | 548/1845 [12:50<29:27, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|███████████████████████▊ | 550/1845 [12:53<31:34, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|███████████████████████▉ | 551/1845 [12:54<31:25, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|███████████████████████▉ | 553/1845 [12:57<30:09, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|████████████████████████ | 554/1845 [12:59<30:06, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|████████████████████████ | 555/1845 [13:00<30:31, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|████████████████████████▏ | 557/1845 [13:03<29:35, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|████████████████████████▏ | 558/1845 [13:04<29:19, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|████████████████████████▎ | 560/1845 [13:07<29:36, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 30%|████████████████████████▎ | 561/1845 [13:08<30:01, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▍ | 563/1845 [13:11<29:59, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▍ | 564/1845 [13:12<29:20, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▌ | 566/1845 [13:15<27:07, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▌ | 567/1845 [13:16<27:23, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▋ | 569/1845 [13:19<30:15, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▋ | 570/1845 [13:21<30:30, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▊ | 572/1845 [13:23<29:20, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▊ | 573/1845 [13:25<29:08, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▉ | 575/1845 [13:27<28:44, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|████████████████████████▉ | 576/1845 [13:29<29:33, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|█████████████████████████ | 577/1845 [13:30<29:56, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|█████████████████████████ | 579/1845 [13:33<31:08, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 31%|█████████████████████████▏ | 580/1845 [13:35<31:09, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▏ | 582/1845 [13:37<26:18, 1.25s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▎ | 584/1845 [13:40<26:57, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▎ | 585/1845 [13:41<27:11, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▍ | 587/1845 [13:44<26:58, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▍ | 588/1845 [13:45<27:09, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▌ | 590/1845 [13:47<27:26, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▋ | 591/1845 [13:49<28:47, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████���███████▋ | 592/1845 [13:50<28:46, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▊ | 594/1845 [13:53<27:30, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▊ | 595/1845 [13:54<28:45, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▉ | 597/1845 [13:57<28:57, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 32%|█████████████████████████▉ | 598/1845 [13:59<29:02, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████ | 600/1845 [14:02<29:49, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████ | 601/1845 [14:03<30:48, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████ | 602/1845 [14:05<31:28, 1.52s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▏ | 604/1845 [14:08<29:57, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▏ | 605/1845 [14:09<29:31, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▎ | 606/1845 [14:11<29:37, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▎ | 608/1845 [14:13<29:20, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▍ | 609/1845 [14:15<27:41, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▍ | 611/1845 [14:17<28:20, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▌ | 612/1845 [14:19<28:46, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▌ | 614/1845 [14:22<28:55, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▋ | 615/1845 [14:23<29:45, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▋ | 616/1845 [14:25<29:13, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 33%|██████████████████████████▊ | 618/1845 [14:28<29:48, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|██████████████████████████▉ | 620/1845 [14:30<27:16, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|██████████████████████████▉ | 621/1845 [14:32<28:25, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|██████████████████████████▉ | 622/1845 [14:33<28:52, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████ | 624/1845 [14:36<29:06, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████ | 625/1845 [14:38<30:08, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████▏ | 626/1845 [14:39<30:36, 1.51s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████▏ | 628/1845 [14:42<30:14, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████▎ | 629/1845 [14:43<29:30, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████▎ | 631/1845 [14:46<28:04, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████▍ | 632/1845 [14:48<29:30, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████▍ | 633/1845 [14:49<29:40, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████▌ | 635/1845 [14:52<28:52, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 34%|███████████████████████████▌ | 636/1845 [14:53<29:08, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|███████████████████████████▌ | 637/1845 [14:55<28:31, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|███████████████████████████▋ | 639/1845 [14:58<28:36, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|███████████████████████████▊ | 640/1845 [14:59<28:59, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|███████████████████████████▊ | 642/1845 [15:02<27:35, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|███████████████████████████▉ | 643/1845 [15:03<26:54, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|███████████████████████████▉ | 645/1845 [15:06<29:23, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|████████████████████████████ | 646/1845 [15:08<29:27, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|████████████████████████████ | 648/1845 [15:10<27:13, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|████████████████████████████▏ | 649/1845 [15:12<26:55, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|████████████████████████████▏ | 651/1845 [15:14<26:34, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|████████████████████████████▎ | 652/1845 [15:16<26:52, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 35%|████████████████████████████▎ | 654/1845 [15:18<27:29, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▍ | 655/1845 [15:20<28:44, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▍ | 656/1845 [15:22<28:43, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▌ | 658/1845 [15:24<27:58, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▌ | 659/1845 [15:26<27:16, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▋ | 661/1845 [15:28<27:35, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▋ | 662/1845 [15:30<27:08, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▊ | 664/1845 [15:33<28:38, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▊ | 665/1845 [15:34<28:04, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▉ | 667/1845 [15:37<27:14, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|████████████████████████████▉ | 668/1845 [15:38<26:30, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|█████████████████████████████ | 670/1845 [15:41<26:51, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|█████████████████████████████ | 671/1845 [15:42<26:24, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 36%|█████████████████████████████▏ | 673/1845 [15:45<26:40, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▏ | 674/1845 [15:46<26:40, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▎ | 675/1845 [15:48<26:32, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▎ | 677/1845 [15:50<26:03, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▍ | 678/1845 [15:52<25:42, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▍ | 680/1845 [15:54<26:50, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▌ | 681/1845 [15:56<28:16, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▌ | 683/1845 [15:59<26:47, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▋ | 684/1845 [16:00<26:57, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▋ | 686/1845 [16:03<25:41, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▊ | 687/1845 [16:04<26:02, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▉ | 689/1845 [16:07<25:15, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 37%|█████████████████████████████▉ | 690/1845 [16:08<25:12, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████ | 692/1845 [16:11<25:46, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████ | 693/1845 [16:12<26:18, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▏ | 695/1845 [16:15<27:08, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▏ | 696/1845 [16:17<28:03, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▏ | 697/1845 [16:18<28:23, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▎ | 699/1845 [16:21<26:31, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▎ | 700/1845 [16:22<27:14, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▍ | 702/1845 [16:25<26:25, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▍ | 703/1845 [16:26<26:33, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▌ | 704/1845 [16:28<26:51, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▌ | 706/1845 [16:31<26:37, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▋ | 707/1845 [16:32<27:39, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▋ | 709/1845 [16:35<27:22, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 38%|██████████████████████████████▊ | 710/1845 [16:37<27:07, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|██████████████████████████████▊ | 712/1845 [16:39<25:22, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|██████████████████████████████▉ | 713/1845 [16:41<27:02, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|██████████████████████████████▉ | 714/1845 [16:42<26:09, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████ | 716/1845 [16:45<26:17, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████▏ | 718/1845 [16:47<24:52, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████▏ | 719/1845 [16:49<25:53, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████▎ | 721/1845 [16:51<23:47, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████▎ | 722/1845 [16:53<24:33, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████▍ | 724/1845 [16:55<25:26, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████▍ | 725/1845 [16:57<25:28, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████▍ | 726/1845 [16:58<26:02, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 39%|███████████████████████████████▌ | 728/1845 [17:01<26:50, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|███████████████████████████████▌ | 729/1845 [17:03<27:40, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|███████████████████████████████▋ | 730/1845 [17:04<28:08, 1.51s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|███████████████████████████████▋ | 732/1845 [17:07<26:20, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|███████████████████████████████▊ | 733/1845 [17:09<26:23, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|███████████████████████████████▊ | 735/1845 [17:11<26:07, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|███████████████████████████████▉ | 736/1845 [17:13<25:34, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|████████████████████████████████ | 738/1845 [17:15<24:43, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|████████████████████████████████ | 739/1845 [17:17<25:22, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|████████████████████████████████ | 740/1845 [17:18<26:36, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|████████████████████████████████▏ | 742/1845 [17:21<26:42, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|████████████████████████████████▏ | 743/1845 [17:23<27:27, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|████████████████████████████████▎ | 745/1845 [17:26<26:17, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 40%|████████████████████████████████▎ | 746/1845 [17:27<25:09, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▍ | 748/1845 [17:30<24:57, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▍ | 749/1845 [17:31<25:27, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▌ | 750/1845 [17:33<26:01, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▌ | 752/1845 [17:35<25:54, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▋ | 753/1845 [17:37<25:26, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▋ | 755/1845 [17:39<24:55, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▊ | 756/1845 [17:41<26:00, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▊ | 758/1845 [17:44<25:18, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▉ | 759/1845 [17:45<25:19, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|████████████████████████████████▉ | 760/1845 [17:47<25:33, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|█████████████████████████████████ | 762/1845 [17:50<26:45, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|█████████████████████████████████ | 763/1845 [17:51<27:08, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 41%|█████████████████████████████████▏ | 764/1845 [17:53<26:34, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▏ | 766/1845 [17:56<27:07, 1.51s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▎ | 767/1845 [17:57<26:56, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▎ | 769/1845 [18:00<24:32, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▍ | 770/1845 [18:01<24:31, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▍ | 772/1845 [18:03<23:15, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▌ | 773/1845 [18:05<23:50, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▌ | 775/1845 [18:08<25:05, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▋ | 776/1845 [18:09<24:37, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▋ | 778/1845 [18:12<24:10, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▊ | 779/1845 [18:13<23:41, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▊ | 780/1845 [18:15<24:08, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▉ | 782/1845 [18:17<24:08, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 42%|█████████████████████████████████▉ | 783/1845 [18:19<23:58, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████ | 785/1845 [18:22<25:02, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████ | 786/1845 [18:23<24:42, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▏ | 788/1845 [18:26<24:19, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|████████████��█████████████████████▏ | 789/1845 [18:27<25:53, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▎ | 790/1845 [18:29<26:22, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▎ | 792/1845 [18:32<25:07, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▍ | 793/1845 [18:33<24:18, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▍ | 794/1845 [18:35<25:10, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▌ | 796/1845 [18:38<25:40, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▌ | 798/1845 [18:40<24:37, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▋ | 799/1845 [18:42<24:19, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▋ | 800/1845 [18:43<24:14, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 43%|██████████████████████████████████▊ | 802/1845 [18:46<24:16, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|██████████████████████████████████▊ | 803/1845 [18:47<24:02, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|██████████████████████████████████▉ | 805/1845 [18:50<24:41, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|██████████████████████████████████▉ | 806/1845 [18:51<24:31, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████ | 808/1845 [18:54<23:19, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████ | 809/1845 [18:56<24:14, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████▏ | 811/1845 [18:58<24:08, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████▏ | 812/1845 [19:00<23:44, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████▎ | 813/1845 [19:01<23:41, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████▎ | 815/1845 [19:04<24:50, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████▍ | 816/1845 [19:05<24:40, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████▍ | 818/1845 [19:08<24:30, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████▌ | 819/1845 [19:10<25:33, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 44%|███████████████████████████████████▌ | 820/1845 [19:11<24:18, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|███████████████████████████████████▋ | 822/1845 [19:14<23:36, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|███████████████████████████████████▋ | 824/1845 [19:17<23:11, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|███████████████████████████████████▊ | 825/1845 [19:18<22:46, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|███████████████████████████████████▊ | 826/1845 [19:19<23:18, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|███████████████████████████████████▉ | 828/1845 [19:22<23:32, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|███████████████████████████████████▉ | 830/1845 [19:25<23:05, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|████████████████████████████████████ | 831/1845 [19:26<22:45, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|████████████████████████████████████ | 833/1845 [19:29<22:13, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|████████████████████████████████████▏ | 834/1845 [19:30<23:41, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|████████████████████████████████████▏ | 836/1845 [19:33<23:05, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|████████████████████████████████████▎ | 837/1845 [19:35<24:01, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 45%|████████████████████████████████████▎ | 838/1845 [19:36<23:32, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▍ | 840/1845 [19:39<24:07, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▍ | 841/1845 [19:40<24:02, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▌ | 842/1845 [19:42<23:07, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▌ | 844/1845 [19:45<23:43, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▋ | 845/1845 [19:46<23:57, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▋ | 846/1845 [19:47<23:55, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▊ | 848/1845 [19:50<22:54, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▊ | 849/1845 [19:51<22:53, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▉ | 851/1845 [19:54<22:12, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|████████████████████████████████████▉ | 852/1845 [19:56<22:46, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|█████████████████████████████████████ | 854/1845 [19:58<23:16, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|█████████████████████████████████████ | 855/1845 [20:00<22:42, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 46%|█████████████████████████████████████▏ | 857/1845 [20:02<22:05, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▏ | 859/1845 [20:05<20:30, 1.25s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▎ | 860/1845 [20:06<20:29, 1.25s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▍ | 862/1845 [20:09<22:14, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▍ | 863/1845 [20:10<21:27, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▌ | 865/1845 [20:13<21:51, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▌ | 866/1845 [20:14<21:28, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▋ | 868/1845 [20:17<22:02, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▋ | 869/1845 [20:18<23:17, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▋ | 870/1845 [20:20<23:08, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▊ | 872/1845 [20:22<22:04, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▊ | 873/1845 [20:24<22:23, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▉ | 875/1845 [20:27<22:20, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 47%|█████████████████████████████████████▉ | 876/1845 [20:28<22:58, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████ | 878/1845 [20:31<23:06, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████ | 879/1845 [20:32<21:59, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▏ | 880/1845 [20:34<22:55, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▏ | 882/1845 [20:37<22:45, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▎ | 883/1845 [20:38<22:48, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▎ | 885/1845 [20:41<22:12, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▍ | 886/1845 [20:42<21:24, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▌ | 888/1845 [20:45<21:45, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▌ | 889/1845 [20:46<21:31, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▋ | 891/1845 [20:49<21:52, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▋ | 892/1845 [20:50<21:43, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 48%|██████████████████████████████████████▊ | 894/1845 [20:53<21:52, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|██████████████████████████████████████▊ | 895/1845 [20:54<22:05, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|██████████████████████████████████████▉ | 897/1845 [20:57<22:05, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|██████████████████████████████████████▉ | 898/1845 [20:59<22:07, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|██████████████████████████████████████▉ | 899/1845 [21:00<22:28, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████ | 901/1845 [21:03<21:52, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████ | 902/1845 [21:04<21:46, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████▏ | 904/1845 [21:07<20:25, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████▏ | 905/1845 [21:08<20:52, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████▎ | 907/1845 [21:11<21:18, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████▎ | 908/1845 [21:12<21:10, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████▍ | 910/1845 [21:15<22:31, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████▌ | 911/1845 [21:17<23:24, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 49%|███████████████████████████████████████▌ | 912/1845 [21:18<21:53, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|███████████████████████████████████████▋ | 914/1845 [21:21<21:01, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|███████████████████████████████████████▋ | 915/1845 [21:22<21:19, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|███████████████████████████████████████▊ | 917/1845 [21:25<21:26, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|███████████████████████████████████████▊ | 918/1845 [21:27<21:57, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|███████████████████████████████████████▉ | 920/1845 [21:29<21:57, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|███████████████████████████████████████▉ | 921/1845 [21:31<22:53, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|███████████████████████████████████████▉ | 922/1845 [21:32<22:23, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████████ | 924/1845 [21:35<21:13, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████████ | 925/1845 [21:37<21:24, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████████▏ | 927/1845 [21:39<21:34, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████████▏ | 928/1845 [21:41<21:05, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████████▎ | 930/1845 [21:44<21:14, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 50%|████████████████████████████████████████▎ | 931/1845 [21:45<21:24, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▍ | 933/1845 [21:48<21:17, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▍ | 934/1845 [21:49<21:57, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▌ | 936/1845 [21:52<22:43, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▋ | 937/1845 [21:54<21:45, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▋ | 939/1845 [21:56<20:58, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▊ | 940/1845 [21:58<21:35, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▊ | 941/1845 [21:59<21:07, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▉ | 943/1845 [22:02<20:55, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|████████████████████████████████████████▉ | 944/1845 [22:04<21:09, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████████ | 946/1845 [22:06<20:54, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████████ | 947/1845 [22:08<20:52, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████████▏ | 949/1845 [22:11<20:48, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 51%|█████████████████████████████████████████▏ | 950/1845 [22:12<20:11, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▏ | 951/1845 [22:13<21:30, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▎ | 953/1845 [22:16<20:48, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▎ | 954/1845 [22:18<21:13, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▍ | 955/1845 [22:19<21:56, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▍ | 957/1845 [22:22<21:47, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▌ | 958/1845 [22:24<21:20, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▌ | 959/1845 [22:25<22:21, 1.51s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▋ | 961/1845 [22:28<21:51, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▋ | 962/1845 [22:30<21:37, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▊ | 964/1845 [22:32<20:56, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▊ | 965/1845 [22:34<21:18, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▉ | 967/1845 [22:36<19:57, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 52%|█████████████████████████████████████████▉ | 968/1845 [22:38<20:32, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████ | 969/1845 [22:40<21:07, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████ | 971/1845 [22:43<21:38, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▏ | 972/1845 [22:44<20:51, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▏ | 974/1845 [22:46<19:45, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▎ | 975/1845 [22:48<19:28, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▎ | 977/1845 [22:50<18:14, 1.26s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▍ | 978/1845 [22:52<19:37, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████��███████████████████████████▍ | 980/1845 [22:55<19:41, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▌ | 981/1845 [22:56<19:51, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▌ | 983/1845 [22:59<19:06, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▋ | 984/1845 [23:00<20:03, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▋ | 985/1845 [23:02<20:18, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 53%|██████████████████████████████████████████▊ | 987/1845 [23:04<19:33, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|██████████████████████████████████████████▊ | 988/1845 [23:06<19:49, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|██████████████████████████████████████████▉ | 990/1845 [23:09<20:19, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|██████████████████████████████████████████▉ | 991/1845 [23:10<20:37, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|███████████████████████████████████████████ | 993/1845 [23:13<20:12, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|███████████████████████████████████████████ | 994/1845 [23:14<20:01, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|███████████████████████████████████████████▏ | 996/1845 [23:17<19:41, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|███████████████████████████████████████████▏ | 997/1845 [23:18<19:18, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|███████████████████████████████████████████▎ | 999/1845 [23:21<19:18, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|██████████████████████████████████████████▊ | 1000/1845 [23:23<19:29, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|██████████████████████████████████████████▊ | 1001/1845 [23:24<19:28, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|██████████████████████████████████████████▉ | 1003/1845 [23:27<19:04, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 54%|██████████████████████████████████████████▉ | 1004/1845 [23:28<20:10, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████ | 1006/1845 [23:31<19:56, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████ | 1007/1845 [23:32<19:48, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▏ | 1009/1845 [23:35<19:17, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▏ | 1010/1845 [23:37<20:00, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▎ | 1011/1845 [23:38<20:24, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▍ | 1013/1845 [23:41<19:31, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▍ | 1014/1845 [23:42<18:54, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▌ | 1016/1845 [23:45<18:17, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▌ | 1017/1845 [23:46<19:04, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▋ | 1019/1845 [23:49<19:14, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▋ | 1020/1845 [23:50<19:03, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▋ | 1021/1845 [23:52<19:52, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 55%|███████████████████████████████████████████▊ | 1023/1845 [23:55<19:41, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|███████████████████████████████████████████▊ | 1024/1845 [23:56<19:37, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|███████████████████████████████████████████▉ | 1026/1845 [23:59<18:59, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████ | 1028/1845 [24:01<17:01, 1.25s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████ | 1029/1845 [24:03<17:32, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████▏ | 1031/1845 [24:05<17:51, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████▏ | 1032/1845 [24:07<18:53, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████▏ | 1033/1845 [24:08<19:40, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████▎ | 1035/1845 [24:11<18:46, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████▎ | 1036/1845 [24:13<18:56, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████▍ | 1038/1845 [24:15<18:12, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████▍ | 1039/1845 [24:17<18:14, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████████████████████████▌ | 1041/1845 [24:19<16:59, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 56%|████████████████████████��███████████████████▌ | 1042/1845 [24:20<17:26, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|████████████████████████████████████████████▋ | 1044/1845 [24:23<17:35, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|████████████████████████████████████████████▋ | 1045/1845 [24:24<17:55, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|████████████████████████████████████████████▊ | 1047/1845 [24:28<19:55, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|████████████████████████████████████████████▊ | 1048/1845 [24:29<20:35, 1.55s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|████████████████████████████████████████████▉ | 1049/1845 [24:31<20:09, 1.52s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|█████████████████████████████████████████████ | 1051/1845 [24:33<18:27, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|█████████████████████████████████████████████ | 1052/1845 [24:35<17:41, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|█████████████████████████████████████████████▏ | 1054/1845 [24:38<18:45, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|█████████████████████████████████████████████▏ | 1055/1845 [24:39<19:00, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|█████████████████████████████████████████████▎ | 1057/1845 [24:42<18:12, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|█████████████████████████████████████████████▎ | 1058/1845 [24:43<18:02, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 57%|█████████████████████████████████████████████▍ | 1060/1845 [24:46<17:31, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▍ | 1061/1845 [24:47<18:46, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▍ | 1062/1845 [24:49<18:12, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▌ | 1064/1845 [24:52<18:59, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▌ | 1065/1845 [24:53<17:37, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▋ | 1067/1845 [24:56<18:43, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▋ | 1068/1845 [24:57<17:00, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▊ | 1070/1845 [25:00<16:59, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▊ | 1071/1845 [25:01<16:47, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▉ | 1073/1845 [25:04<17:38, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|█████████████████████████████████████████████▉ | 1074/1845 [25:05<17:51, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|██████████████████████████████████████████████ | 1076/1845 [25:08<17:42, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|██████████████████████████████████████████████ | 1077/1845 [25:09<17:31, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 58%|██████████████████████████████████████████████▏ | 1079/1845 [25:12<16:41, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▏ | 1080/1845 [25:13<16:58, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▎ | 1082/1845 [25:16<18:06, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▎ | 1083/1845 [25:18<18:12, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▍ | 1084/1845 [25:19<18:01, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▌ | 1086/1845 [25:22<17:50, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▌ | 1087/1845 [25:23<18:28, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|█████████████████████████���████████████████████▌ | 1088/1845 [25:25<18:06, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▋ | 1090/1845 [25:28<18:35, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▋ | 1091/1845 [25:29<18:02, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▊ | 1092/1845 [25:30<17:38, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▊ | 1094/1845 [25:33<17:36, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▉ | 1095/1845 [25:35<17:11, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 59%|██████████████████████████████████████████████▉ | 1097/1845 [25:37<17:02, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████ | 1098/1845 [25:39<17:53, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████ | 1100/1845 [25:42<17:45, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▏ | 1101/1845 [25:43<17:26, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▏ | 1102/1845 [25:45<17:30, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▎ | 1104/1845 [25:47<17:29, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▎ | 1105/1845 [25:49<17:14, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▍ | 1107/1845 [25:51<16:57, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▍ | 1109/1845 [25:54<16:12, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▌ | 1110/1845 [25:55<16:24, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▌ | 1112/1845 [25:58<16:14, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▋ | 1113/1845 [26:00<16:51, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▋ | 1115/1845 [26:02<16:04, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 60%|███████████████████████████████████████████████▊ | 1116/1845 [26:04<16:51, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|███████████████████████████████████████████████▊ | 1117/1845 [26:05<17:14, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|███████████████████████████████████████████████▉ | 1119/1845 [26:08<17:37, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|███████████████████████████████████████████████▉ | 1120/1845 [26:09<17:13, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████ | 1122/1845 [26:12<16:30, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████ | 1123/1845 [26:13<16:35, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████▏ | 1124/1845 [26:15<16:58, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████▏ | 1126/1845 [26:18<16:27, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████▎ | 1128/1845 [26:20<16:01, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████▎ | 1129/1845 [26:22<16:45, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████▍ | 1131/1845 [26:24<15:46, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████▍ | 1132/1845 [26:26<15:44, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 61%|████████████████████████████████████████████████▌ | 1134/1845 [26:28<15:52, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|████████████████████████████████████████████████▌ | 1135/1845 [26:30<15:59, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|████████████████████████████████████████████████▋ | 1137/1845 [26:32<15:07, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|████████████████████████████████████████████████▋ | 1138/1845 [26:33<15:21, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|████████████████████████████████████████████████▊ | 1140/1845 [26:36<15:04, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|████████████████████████████████████████████████▊ | 1141/1845 [26:38<15:55, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|████████████████████████████████████████████████▉ | 1142/1845 [26:39<15:48, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|████████████████████████████████████████████████▉ | 1144/1845 [26:41<15:16, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|█████████████████████████████████████████████████ | 1145/1845 [26:43<15:24, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|█████████████████████████████████████████████████ | 1147/1845 [26:46<16:14, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|█████████████████████████████████████████████████▏ | 1148/1845 [26:47<16:15, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|█████████████████████████████████████████████████▏ | 1150/1845 [26:50<16:38, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|█████████████████████████████████████████████████▎ | 1151/1845 [26:51<16:25, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 62%|█████████████████████████████████████████████████▎ | 1152/1845 [26:53<16:18, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▍ | 1154/1845 [26:56<16:01, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▍ | 1155/1845 [26:57<16:20, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▌ | 1157/1845 [27:00<15:55, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▌ | 1158/1845 [27:01<16:01, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▋ | 1160/1845 [27:04<16:44, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▋ | 1161/1845 [27:06<16:23, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▊ | 1163/1845 [27:08<15:53, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▊ | 1164/1845 [27:10<15:40, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▉ | 1165/1845 [27:11<15:44, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|█████████████████████████████████████████████████▉ | 1167/1845 [27:14<16:33, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|██████████████████████████████████████████████████ | 1168/1845 [27:16<16:46, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|██████████████████████████████████████████████████ | 1169/1845 [27:17<16:55, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 63%|██████████████████████████████████████████████████▏ | 1171/1845 [27:20<16:30, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▏ | 1172/1845 [27:22<16:35, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▎ | 1174/1845 [27:25<16:08, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▎ | 1175/1845 [27:26<15:31, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▎ | 1176/1845 [27:27<15:36, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▍ | 1178/1845 [27:30<15:10, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▍ | 1179/1845 [27:31<15:02, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▌ | 1181/1845 [27:34<14:51, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▌ | 1182/1845 [27:35<15:05, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▋ | 1184/1845 [27:38<15:37, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▋ | 1185/1845 [27:40<15:44, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▊ | 1187/1845 [27:43<15:48, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▊ | 1188/1845 [27:44<14:54, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 64%|██████████████████████████████████████████████████▉ | 1190/1845 [27:46<14:49, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|██████████████████████████████████████████████████▉ | 1191/1845 [27:48<15:33, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████ | 1192/1845 [27:49<15:26, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▏ | 1194/1845 [27:52<15:24, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▏ | 1195/1845 [27:54<15:38, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▎ | 1197/1845 [27:56<14:36, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▎ | 1199/1845 [27:59<13:54, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▍ | 1200/1845 [28:00<14:20, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▍ | 1201/1845 [28:02<14:56, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▌ | 1203/1845 [28:05<14:47, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▌ | 1204/1845 [28:06<14:53, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▋ | 1206/1845 [28:09<14:33, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 65%|███████████████████████████████████████████████████▋ | 1207/1845 [28:10<14:13, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|███████████████████████████████████████████████████▊ | 1209/1845 [28:13<14:26, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|███████████████████████████████████████���███████████▊ | 1210/1845 [28:14<14:09, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|███████████████████████████████████████████████████▉ | 1212/1845 [28:17<14:09, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|███████████████████████████████████████████████████▉ | 1213/1845 [28:18<14:04, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|████████████████████████████████████████████████████ | 1215/1845 [28:21<14:23, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|████████████████████████████████████████████████████ | 1217/1845 [28:23<13:25, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|████████████████████████████████████████████████████▏ | 1218/1845 [28:25<13:57, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|████████████████████████████████████████████████████▏ | 1219/1845 [28:26<14:00, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|████████████████████████████████████████████████████▎ | 1221/1845 [28:29<14:35, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|████████████████████████████████████████████████████▎ | 1222/1845 [28:30<14:19, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|████████████████████████████████████████████████████▍ | 1224/1845 [28:33<14:32, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 66%|████████████████████████████████████████████████████▍ | 1225/1845 [28:35<14:32, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|████████████████████████████████████████████████████▌ | 1227/1845 [28:37<13:47, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|████████████████████████████████████████████████████▌ | 1228/1845 [28:39<14:15, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|████████████████████████████████████████████████████▌ | 1229/1845 [28:40<14:03, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|████████████████████████████████████████████████████▋ | 1231/1845 [28:43<13:29, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|████████████████████████████████████████████████████▊ | 1232/1845 [28:44<13:43, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|████████████████████████████████████████████████████▊ | 1234/1845 [28:47<13:12, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|████████████████████████████████████████████████████▉ | 1235/1845 [28:48<13:43, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|████████████████████████████████████████████████████▉ | 1237/1845 [28:51<14:22, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|█████████████████████████████████████████████████████ | 1238/1845 [28:52<14:24, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|█████████████████████████████████████████████████████ | 1240/1845 [28:55<14:05, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|█████████████████████████████████████████████████████▏ | 1242/1845 [28:58<13:26, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|█████████████████████████████████████████████████████▏ | 1243/1845 [28:59<13:35, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 67%|█████████████████████████████████████████████████████▎ | 1245/1845 [29:02<13:37, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▎ | 1246/1845 [29:03<13:20, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▍ | 1247/1845 [29:04<13:14, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▍ | 1249/1845 [29:07<13:56, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▌ | 1250/1845 [29:08<13:08, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▌ | 1252/1845 [29:12<14:06, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▋ | 1253/1845 [29:13<14:14, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▋ | 1255/1845 [29:16<13:15, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|███████████████████████████████████████████████████���█▊ | 1256/1845 [29:17<13:52, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▊ | 1257/1845 [29:19<13:53, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▉ | 1259/1845 [29:21<14:00, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|█████████████████████████████████████████████████████▉ | 1260/1845 [29:23<13:46, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|██████████████████████████████████████████████████████ | 1262/1845 [29:26<14:06, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 68%|██████████████████████████████████████████████████████ | 1263/1845 [29:27<13:47, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████ | 1264/1845 [29:29<13:57, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▏ | 1266/1845 [29:32<14:37, 1.51s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▎ | 1267/1845 [29:33<13:37, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▎ | 1269/1845 [29:36<13:43, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▍ | 1270/1845 [29:37<13:58, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▍ | 1272/1845 [29:40<13:28, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▌ | 1273/1845 [29:41<12:39, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▌ | 1275/1845 [29:44<13:10, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▋ | 1276/1845 [29:46<13:24, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▋ | 1277/1845 [29:47<12:41, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▊ | 1279/1845 [29:50<13:04, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▊ | 1280/1845 [29:51<12:55, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 69%|██████████████████████████████████████████████████████▉ | 1282/1845 [29:54<13:06, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|██████████████████████████████████████████████████████▉ | 1283/1845 [29:55<13:11, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████ | 1285/1845 [29:58<13:37, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████ | 1286/1845 [30:00<13:15, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▏ | 1288/1845 [30:02<12:08, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▏ | 1289/1845 [30:03<12:32, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▎ | 1291/1845 [30:06<12:48, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▎ | 1292/1845 [30:08<12:51, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▍ | 1294/1845 [30:10<12:31, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▍ | 1295/1845 [30:12<13:14, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▍ | 1296/1845 [30:13<13:09, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▌ | 1298/1845 [30:16<13:02, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 70%|███████████████████████████████████████████████████████▌ | 1299/1845 [30:18<12:46, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|███████████████████████████████████████████████████████▋ | 1301/1845 [30:21<12:50, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|███████████████████████████████████████████████████████▋ | 1302/1845 [30:22<13:10, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|███████████████████████████████████████████████████████▊ | 1303/1845 [30:23<13:00, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|███████████████████████████████████████████████████████▉ | 1305/1845 [30:26<12:28, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|███████████████████████████████████████████████████████▉ | 1306/1845 [30:27<12:03, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████ | 1308/1845 [30:30<12:12, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████ | 1309/1845 [30:32<12:53, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████ | 1310/1845 [30:33<12:53, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████▏ | 1312/1845 [30:36<12:55, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████▏ | 1313/1845 [30:38<13:13, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████▎ | 1315/1845 [30:41<12:51, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████▎ | 1316/1845 [30:42<12:36, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████▍ | 1317/1845 [30:43<12:46, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 71%|████████████████████████████████████████████████████████▍ | 1319/1845 [30:46<12:19, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|████████████████████████████████████████████████████████▌ | 1320/1845 [30:48<12:02, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|████████████████████████████████████████████████████████▌ | 1322/1845 [30:50<12:15, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|████████████████████████████████████████████████████████▋ | 1323/1845 [30:52<12:30, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|████████████████████████████████████████████████████████▋ | 1325/1845 [30:55<12:10, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|████████████████████████████████████████████████████████▊ | 1326/1845 [30:56<12:42, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|████████████████████████████████████████████████████████▊ | 1327/1845 [30:58<12:45, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|████████████████████████████████████████████████████████▉ | 1329/1845 [31:01<12:26, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|████████████████████████████████████████████████████████▉ | 1330/1845 [31:02<11:50, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|█████████████████████████████████████████████████████████ | 1332/1845 [31:04<11:07, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|█████████████████████████████████████████████████████████ | 1333/1845 [31:06<10:54, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|█████████████████████████████████████████████████████████▏ | 1335/1845 [31:09<11:51, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 72%|█████████████████████████████████████████████████████████▏ | 1336/1845 [31:10<12:17, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▎ | 1338/1845 [31:12<10:49, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▎ | 1339/1845 [31:14<11:16, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▍ | 1341/1845 [31:17<11:22, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▍ | 1342/1845 [31:18<11:17, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▌ | 1344/1845 [31:21<11:54, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▌ | 1345/1845 [31:23<12:15, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▋ | 1347/1845 [31:25<11:09, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▋ | 1348/1845 [31:26<10:50, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▊ | 1350/1845 [31:29<10:35, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▊ | 1351/1845 [31:30<10:46, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|█████████████████████████████████████████████████████████▉ | 1353/1845 [31:33<11:21, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|██████████████████████████████████████████████████████████ | 1355/1845 [31:35<10:22, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 73%|██████████████████████████████████████████████████████████ | 1356/1845 [31:36<10:08, 1.24s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▏ | 1358/1845 [31:39<10:07, 1.25s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▏ | 1360/1845 [31:42<10:25, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▎ | 1361/1845 [31:43<10:14, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▎ | 1363/1845 [31:46<10:36, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▍ | 1364/1845 [31:47<10:45, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▍ | 1365/1845 [31:48<10:44, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▌ | 1367/1845 [31:51<10:49, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▌ | 1368/1845 [31:53<10:41, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▋ | 1370/1845 [31:55<10:14, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▋ | 1372/1845 [31:58<10:22, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▊ | 1373/1845 [31:59<10:17, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 74%|██████████████████████████████████████████████████████████▊ | 1374/1845 [32:01<10:36, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|██��███████████████████████████████████████████████████████▉ | 1376/1845 [32:04<11:12, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|██████████████████████████████████████████████████████████▉ | 1377/1845 [32:05<10:50, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|███████████████████████████████████████████████████████████ | 1379/1845 [32:07<10:22, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|███████████████████████████████████████████████████████████ | 1380/1845 [32:09<10:29, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|███████████████████████████████████████████████████████████▏ | 1382/1845 [32:11<10:17, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|███████████████████████████████████████████████████████████▏ | 1383/1845 [32:13<10:14, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|███████████████████████████████████████████████████████████▎ | 1385/1845 [32:16<10:14, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|███████████████████████████████████████████████████████████▍ | 1387/1845 [32:18<09:49, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|███████████████████████████████████████████████████████████▍ | 1388/1845 [32:19<10:02, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|███████████████████████████████████████████████████████████▍ | 1389/1845 [32:21<09:55, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 75%|██████████████████��████████████████████████████████████████▌ | 1391/1845 [32:24<10:31, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|███████████████████████████████████████████████████████████▋ | 1393/1845 [32:26<09:56, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|███████████████████████████████████████████████████████████▋ | 1394/1845 [32:27<09:24, 1.25s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|███████████████████████████████████████████████████████████▊ | 1396/1845 [32:30<09:49, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|███████████████████████████████████████████████████████████▊ | 1397/1845 [32:31<09:29, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|███████████████████████████████████████████████████████████▉ | 1399/1845 [32:34<10:12, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|███████████████████████████████████████████████████████████▉ | 1400/1845 [32:35<09:49, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|███████████████████████████████████████████████████████████▉ | 1401/1845 [32:37<10:26, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|████████████████████████████████████████████████████████████ | 1403/1845 [32:40<10:45, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|████████████████████████████████████████████████████████████ | 1404/1845 [32:41<10:09, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|████████████████████████████████████████████████████████████▏ | 1406/1845 [32:44<10:36, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|████████████████████████████████████████████████████████████▏ | 1407/1845 [32:45<10:07, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|████████████████████████████████████████████████████████████▎ | 1409/1845 [32:48<10:11, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|████████████████████████████████████████████████████████████▎ | 1410/1845 [32:50<10:15, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 76%|████████████████████████████████████████████████████████████▍ | 1411/1845 [32:51<10:12, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|████████████████████████████████████████████████████████████▌ | 1413/1845 [32:54<09:47, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|████████████████████████████████████████████████████████████▌ | 1414/1845 [32:55<09:51, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|████████████████████████████████████████████████████████████▋ | 1416/1845 [32:58<09:51, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|████████████████████████████████████████████████████████████▋ | 1418/1845 [33:00<08:49, 1.24s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|████████████████████████████████████████████████████████████▊ | 1419/1845 [33:01<08:56, 1.26s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|██████████████████████████████████████��█████████████████████▊ | 1421/1845 [33:04<09:40, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|████████████████████████████████████████████████████████████▉ | 1422/1845 [33:06<09:40, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|████████████████████████████████████████████████████████████▉ | 1423/1845 [33:07<09:48, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|█████████████████████████████████████████████████████████████ | 1425/1845 [33:10<09:52, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|█████████████████████████████████████████████████████████████ | 1426/1845 [33:11<09:40, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|█████████████████████████████████████████████████████████████▏ | 1428/1845 [33:14<09:29, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 77%|█████████████████████████████████████████████████████████████▏ | 1429/1845 [33:16<09:50, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▎ | 1431/1845 [33:18<09:41, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▎ | 1432/1845 [33:20<10:05, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▎ | 1433/1845 [33:22<10:09, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▍ | 1435/1845 [33:24<09:35, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▍ | 1436/1845 [33:26<09:24, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▌ | 1438/1845 [33:28<09:23, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▌ | 1439/1845 [33:30<09:06, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▋ | 1441/1845 [33:32<08:56, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▋ | 1442/1845 [33:34<09:09, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▊ | 1443/1845 [33:35<09:12, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▊ | 1445/1845 [33:38<09:00, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|█████████████████████████████████████████████████████████████▉ | 1446/1845 [33:39<09:06, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 78%|██████████████████████████████████████████████████████████████ | 1448/1845 [33:42<09:00, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████ | 1450/1845 [33:45<09:06, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▏ | 1451/1845 [33:46<08:56, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▏ | 1453/1845 [33:49<08:47, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▎ | 1454/1845 [33:50<08:59, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▎ | 1456/1845 [33:53<08:48, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▍ | 1457/1845 [33:54<08:51, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▍ | 1458/1845 [33:56<09:19, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▌ | 1460/1845 [33:59<09:16, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▌ | 1461/1845 [34:00<09:26, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▋ | 1463/1845 [34:03<08:38, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▋ | 1464/1845 [34:04<08:50, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 79%|██████████████████████████████████████████████████████████████▋ | 1465/1845 [34:06<08:52, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|██████████████████████████████████████████████████████████████▊ | 1467/1845 [34:09<08:48, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|██████████████████████████████████████████████████████████████▊ | 1468/1845 [34:10<08:43, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|██████████████████████████████████████████████████████████████▉ | 1470/1845 [34:13<08:40, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|██████████████████████████████████████████████████████████████▉ | 1471/1845 [34:14<08:37, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████ | 1473/1845 [34:17<08:35, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████ | 1474/1845 [34:18<08:35, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████▏ | 1476/1845 [34:21<08:28, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████▏ | 1477/1845 [34:22<08:18, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████▎ | 1479/1845 [34:25<08:11, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████▎ | 1480/1845 [34:27<08:18, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████▍ | 1481/1845 [34:28<08:30, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████▍ | 1483/1845 [34:31<08:35, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 80%|███████████████████████████████████████████████████████████████▌ | 1484/1845 [34:32<08:14, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|███████████████████████████████████████████████████████████████▋ | 1486/1845 [34:35<07:52, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|███████████████████████████████████████████████████████████████▋ | 1487/1845 [34:36<08:06, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|███████████████████████████████████████████████████████████████▊ | 1489/1845 [34:39<08:24, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|███████████████████████████████████████████████████████████████▊ | 1490/1845 [34:40<08:17, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|███████████████████████████████████████████████████████████████▊ | 1491/1845 [34:42<08:25, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|██████████████��████████████████████████████████████████████████▉ | 1493/1845 [34:45<08:03, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|███████████████████████████████████████████████████████████████▉ | 1494/1845 [34:46<07:32, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|████████████████████████████████████████████████████████████████ | 1496/1845 [34:48<07:34, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|████████████████████████████████████████████████████████████████ | 1497/1845 [34:50<07:41, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|████████████████████████████████████████████████████████████████▏ | 1499/1845 [34:53<07:55, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|████████████████████████████████████████████████████████████████▏ | 1500/1845 [34:54<07:52, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|████████████████████████████████████████████████████████████████▎ | 1502/1845 [34:57<08:15, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 81%|████████████████████████████████████████████████████████████████▎ | 1503/1845 [34:58<08:11, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▍ | 1504/1845 [35:00<08:19, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▍ | 1506/1845 [35:03<08:04, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▌ | 1507/1845 [35:04<08:05, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▌ | 1509/1845 [35:07<07:58, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▋ | 1510/1845 [35:09<08:19, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▋ | 1511/1845 [35:10<08:21, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▊ | 1513/1845 [35:13<07:54, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▊ | 1514/1845 [35:14<07:26, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▉ | 1516/1845 [35:17<07:42, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|████████████████████████████████████████████████████████████████▉ | 1517/1845 [35:18<07:27, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|█████████████████████████████████████████████████████████████████ | 1519/1845 [35:21<06:50, 1.26s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|█████████████████████████████████████████████████████████████████ | 1520/1845 [35:22<07:01, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 82%|█████████████████████████████████████████████████████████████████▏ | 1522/1845 [35:25<07:09, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▏ | 1523/1845 [35:26<07:14, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▎ | 1525/1845 [35:29<07:09, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▍ | 1527/1845 [35:31<06:51, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▍ | 1528/1845 [35:33<07:02, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▌ | 1530/1845 [35:35<06:59, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▌ | 1531/1845 [35:37<06:48, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▋ | 1533/1845 [35:39<06:52, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▋ | 1534/1845 [35:41<07:08, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▋ | 1535/1845 [35:42<07:02, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▊ | 1537/1845 [35:45<07:11, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▊ | 1538/1845 [35:46<07:14, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 83%|█████████████████████████████████████████████████████████████████▉ | 1540/1845 [35:49<07:02, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|█████████████████████████████████████████████████████████████████▉ | 1541/1845 [35:50<06:51, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████ | 1543/1845 [35:53<06:46, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████ | 1544/1845 [35:55<06:50, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▏ | 1546/1845 [35:57<06:48, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▏ | 1547/1845 [35:59<06:48, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▎ | 1549/1845 [36:01<06:39, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▎ | 1550/1845 [36:03<07:00, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▍ | 1552/1845 [36:06<07:05, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▍ | 1553/1845 [36:07<07:06, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▌ | 1554/1845 [36:09<07:04, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▋ | 1556/1845 [36:11<06:39, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▋ | 1558/1845 [36:14<06:44, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 84%|██████████████████████████████████████████████████████████████████▊ | 1559/1845 [36:16<06:35, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|██████████████████████████████████████████████████████████████████▊ | 1560/1845 [36:17<06:28, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|██████████████████████████████████████████████████████████████████▉ | 1562/1845 [36:19<06:17, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|██████████████████████████████████████████████████████████████████▉ | 1564/1845 [36:22<06:12, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████ | 1565/1845 [36:24<06:22, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████ | 1566/1845 [36:25<06:32, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████▏ | 1568/1845 [36:28<06:34, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████▏ | 1569/1845 [36:29<06:40, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████▏ | 1570/1845 [36:31<06:33, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████▎ | 1571/1845 [36:32<06:44, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████▎ | 1573/1845 [36:35<06:36, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████▍ | 1574/1845 [36:37<06:37, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████▍ | 1576/1845 [36:40<06:40, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 85%|███████████████████████████████████████████████████████████████████▌ | 1577/1845 [36:41<06:31, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|███████████████████████���███████████████████████████████████████████▌ | 1578/1845 [36:43<06:36, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|███████████████████████████████████████████████████████████████████▋ | 1580/1845 [36:46<06:18, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|███████████████████████████████████████████████████████████████████▋ | 1581/1845 [36:47<06:01, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|███████████████████████████████████████████████████████████████████▊ | 1583/1845 [36:50<06:15, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|███████████████████████████████████████████████████████████████████▊ | 1584/1845 [36:51<06:04, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|███████████████████████████████████████████████████████████████████▉ | 1586/1845 [36:54<06:05, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|███████████████████████████████████████████████████████████████████▉ | 1587/1845 [36:55<05:50, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|████████████████████████████████████████████████████████████████████ | 1589/1845 [36:58<06:12, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|████████████████████████████████████████████████████████████████████ | 1590/1845 [37:00<06:05, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|████████████████████████████████████████████████████████████████████ | 1591/1845 [37:01<05:58, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|████████████████████████████████████████████████████████████████████▏ | 1593/1845 [37:04<05:43, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 86%|████████████████████████████████████████████████████████████████████▎ | 1595/1845 [37:06<05:37, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▎ | 1596/1845 [37:08<05:32, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▍ | 1598/1845 [37:10<05:28, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▍ | 1599/1845 [37:12<05:25, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▌ | 1601/1845 [37:14<05:22, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▌ | 1602/1845 [37:16<05:29, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▋ | 1604/1845 [37:18<05:32, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▋ | 1605/1845 [37:20<05:41, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▊ | 1607/1845 [37:23<05:17, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▊ | 1608/1845 [37:24<05:10, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▉ | 1609/1845 [37:25<05:20, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|████████████████████████████████████████████████████████████████████▉ | 1611/1845 [37:28<05:30, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|█████████████████████████████████████████████████████████████████████ | 1612/1845 [37:30<05:36, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 87%|█████████████████████████████████████████████████████████████████████ | 1614/1845 [37:33<05:27, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▏ | 1615/1845 [37:34<05:20, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▏ | 1617/1845 [37:36<05:05, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▎ | 1618/1845 [37:38<05:05, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▎ | 1619/1845 [37:39<05:11, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|██████████████████████████████████████████████████████████████��██████▍ | 1621/1845 [37:42<05:28, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▍ | 1622/1845 [37:44<05:37, 1.51s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▍ | 1623/1845 [37:45<05:33, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▌ | 1625/1845 [37:48<05:14, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▌ | 1626/1845 [37:50<05:09, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▋ | 1628/1845 [37:52<05:03, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▊ | 1629/1845 [37:54<05:14, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▊ | 1630/1845 [37:55<05:14, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 88%|█████████████████████████████████████████████████████████████████████▉ | 1632/1845 [37:58<04:54, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|█████████████████████████████████████████████████████████████████████▉ | 1634/1845 [38:00<04:34, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|█████████���████████████████████████████████████████████████████████████ | 1635/1845 [38:02<04:26, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████ | 1637/1845 [38:04<04:32, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▏ | 1638/1845 [38:06<04:25, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▏ | 1640/1845 [38:08<04:34, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▎ | 1641/1845 [38:10<04:39, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▎ | 1643/1845 [38:12<04:28, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▍ | 1644/1845 [38:14<04:43, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▍ | 1645/1845 [38:16<04:47, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▌ | 1647/1845 [38:18<04:40, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▌ | 1648/1845 [38:20<04:33, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▋ | 1650/1845 [38:23<04:47, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 89%|██████████████████████████████████████████████████████████████████████▋ | 1651/1845 [38:24<04:35, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|██████████████████████████████████████████████████████████████████████▊ | 1653/1845 [38:27<04:13, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|██████████████████████████████████████████████████████████████████████▊ | 1654/1845 [38:28<04:18, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|██████████████████████████████████████████████████████████████████████▉ | 1656/1845 [38:31<04:18, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|██████████████████████████████████████████████████████████████████████▉ | 1657/1845 [38:32<04:08, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|███████████████████████████████████████████████████████████████████████ | 1659/1845 [38:34<03:58, 1.28s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|███████████████████████████████████████████████████████████████████████ | 1661/1845 [38:37<04:02, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|███████████████████████████████████████████████████████████████████████▏ | 1662/1845 [38:39<04:07, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|██████████████████████████████████████████████████████████████████��████▏ | 1664/1845 [38:41<03:58, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|███████████████████████████████████████████████████████████████████████▎ | 1665/1845 [38:42<03:56, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|███████████████████████████████████████████████████████████████████████▎ | 1666/1845 [38:44<04:00, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|███████████████████████████████████████████████████████████████████████▍ | 1668/1845 [38:47<04:15, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 90%|███████████████████████████████████████████████████████████████████████▍ | 1669/1845 [38:48<03:56, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|███████████████████████████████████████████████████████████████████████▌ | 1671/1845 [38:51<04:16, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|███████████████████████████████████████████████████████████████████████▌ | 1672/1845 [38:53<04:12, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|███████████████████████████████████████████████████████████████████████▋ | 1674/1845 [38:55<03:57, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|███████████████████████████████████████████████████████████████████████▋ | 1675/1845 [38:57<03:57, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|███████████████████████████████████████████████████████████████████████▊ | 1676/1845 [38:58<03:53, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|███████████████████████████████████████████████████████████████████████▊ | 1678/1845 [39:01<04:00, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|███████████████████████████████████████████████████████████████████████▉ | 1679/1845 [39:02<03:52, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|███████████████████████████████████████████████████████████████████████▉ | 1681/1845 [39:05<03:47, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|████████████████████████████████████████████████████████████████████████ | 1682/1845 [39:06<03:36, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|████████████████████████████████████████████████████████████████████████ | 1684/1845 [39:09<03:38, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|████████████████████████████████████████████████████████████████████████▏ | 1685/1845 [39:10<03:39, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|████████████████████████████████████████████████████████████████████████▏ | 1687/1845 [39:13<03:39, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 91%|████████████████████████████████████████████████████████████████████████▎ | 1688/1845 [39:15<03:47, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▎ | 1690/1845 [39:17<03:13, 1.25s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▍ | 1691/1845 [39:18<03:18, 1.29s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▍ | 1693/1845 [39:21<03:19, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▌ | 1694/1845 [39:23<03:34, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▌ | 1695/1845 [39:24<03:29, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▋ | 1697/1845 [39:27<03:28, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▋ | 1698/1845 [39:28<03:25, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▊ | 1700/1845 [39:31<03:25, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▉ | 1702/1845 [39:34<03:20, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▉ | 1703/1845 [39:36<03:24, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|████████████████████████████████████████████████████████████████████████▉ | 1704/1845 [39:37<03:23, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 92%|██████████████████████████████████��██████████████████████████████████████ | 1706/1845 [39:40<03:20, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████ | 1707/1845 [39:41<03:22, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▏ | 1709/1845 [39:44<03:03, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▏ | 1710/1845 [39:45<02:58, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▎ | 1712/1845 [39:48<02:58, 1.34s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▎ | 1713/1845 [39:49<02:57, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▍ | 1715/1845 [39:52<02:59, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▍ | 1716/1845 [39:53<02:56, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▌ | 1717/1845 [39:55<02:59, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▌ | 1719/1845 [39:58<02:50, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▋ | 1720/1845 [39:59<02:55, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▋ | 1722/1845 [40:02<02:50, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▊ | 1723/1845 [40:03<02:47, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 93%|█████████████████████████████████████████████████████████████████████████▊ | 1725/1845 [40:06<02:48, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|█████████████████████████████████████████████████████████████████████████▉ | 1726/1845 [40:08<02:55, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|█████████████████████████████████████████████████████████████████████████▉ | 1728/1845 [40:10<02:43, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████ | 1729/1845 [40:12<02:38, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████ | 1731/1845 [40:14<02:33, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████▏ | 1732/1845 [40:15<02:28, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████▏ | 1734/1845 [40:18<02:35, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████▎ | 1735/1845 [40:20<02:38, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████▎ | 1736/1845 [40:21<02:30, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████▍ | 1738/1845 [40:24<02:32, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████▍ | 1739/1845 [40:25<02:27, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████▌ | 1741/1845 [40:28<02:29, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 94%|██████████████████████████████████████████████████████████████████████████▌ | 1742/1845 [40:30<02:24, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|██████████████████████████████████████████████████████████████████████████▋ | 1744/1845 [40:32<02:18, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|██████████████████████████████████████████████████████████████████████████▋ | 1745/1845 [40:33<02:12, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|██████████████████████████████████████████████████████████████████████████▊ | 1747/1845 [40:36<02:10, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|██████████████████████████████████████████████████████████████████████████▉ | 1749/1845 [40:39<02:01, 1.26s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|██████████████████████████████████████████████████████████████████████████▉ | 1750/1845 [40:40<02:04, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|██████████████████████████████████████████████████████████████████████████▉ | 1751/1845 [40:41<02:02, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|███████████████████████████████████████████████████████████████████████████ | 1753/1845 [40:44<02:12, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|███████████████████████████████████████████████████████████████████████████ | 1754/1845 [40:46<02:16, 1.50s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|███████████████████████████████████████████████████████████████████████████▏ | 1755/1845 [40:47<02:12, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|███████████████████████████████████████████████████████████████████████████▏ | 1757/1845 [40:50<02:08, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|███████████████████████████████████████████████████████████████████████████▎ | 1758/1845 [40:52<02:04, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|███████████████████████████████████████████████████████████████████████████▎ | 1760/1845 [40:55<02:00, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 95%|█████████████████████████████████████████████████████████���█████████████████▍ | 1761/1845 [40:56<02:04, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▍ | 1762/1845 [40:58<02:01, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▌ | 1764/1845 [41:01<01:58, 1.47s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▌ | 1765/1845 [41:02<01:54, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▋ | 1767/1845 [41:05<01:47, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▋ | 1768/1845 [41:06<01:48, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▋ | 1769/1845 [41:07<01:44, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▊ | 1771/1845 [41:10<01:40, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▊ | 1772/1845 [41:12<01:41, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|███████████████████████████████████████████████████████████████████████████▉ | 1774/1845 [41:14<01:33, 1.32s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|████████████████████████████████████████████████████████████████████████████ | 1776/1845 [41:17<01:33, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|████████████████████████████████████████████████████████████████████████████ | 1777/1845 [41:18<01:31, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|████████████████████████████████████████████████████████████████████████████▏ | 1779/1845 [41:21<01:29, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 96%|████████████████████████████████████████████████████████████████████████████▏ | 1780/1845 [41:22<01:31, 1.41s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▎ | 1781/1845 [41:24<01:35, 1.49s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▎ | 1783/1845 [41:27<01:27, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▍ | 1784/1845 [41:28<01:29, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▍ | 1785/1845 [41:30<01:24, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▌ | 1787/1845 [41:33<01:24, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▌ | 1788/1845 [41:34<01:24, 1.48s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▌ | 1789/1845 [41:36<01:26, 1.54s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▋ | 1791/1845 [41:39<01:18, 1.46s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▋ | 1792/1845 [41:40<01:12, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▊ | 1794/1845 [41:42<01:06, 1.31s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▊ | 1795/1845 [41:44<01:09, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▉ | 1797/1845 [41:46<01:02, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 97%|████████████████████████████████████████████████████████████████████████████▉ | 1798/1845 [41:48<01:03, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████ | 1800/1845 [41:50<01:00, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████ | 1801/1845 [41:52<00:58, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▏ | 1803/1845 [41:54<00:55, 1.33s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▏ | 1804/1845 [41:56<00:55, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▎ | 1806/1845 [41:59<00:55, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▎ | 1807/1845 [42:00<00:53, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▍ | 1809/1845 [42:03<00:51, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▌ | 1810/1845 [42:04<00:48, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▌ | 1811/1845 [42:06<00:47, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▋ | 1813/1845 [42:09<00:45, 1.43s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▋ | 1814/1845 [42:10<00:44, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|████████████████████��████████████████████████████████████████████████████████▊ | 1816/1845 [42:13<00:42, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 98%|█████████████████████████████████████████████████████████████████████████████▊ | 1817/1845 [42:14<00:40, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|█████████████████████████████████████████████████████████████████████████████▊ | 1818/1845 [42:16<00:37, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|█████████████████████████████████████████████████████████████████████████████▉ | 1820/1845 [42:19<00:36, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|█████████████████████████████████████████████████████████████████████████████▉ | 1821/1845 [42:20<00:34, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████ | 1823/1845 [42:23<00:30, 1.38s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████ | 1824/1845 [42:24<00:28, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████▏| 1826/1845 [42:27<00:26, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████▏| 1827/1845 [42:28<00:24, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████▎| 1828/1845 [42:30<00:24, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████▎| 1830/1845 [42:33<00:21, 1.45s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████▍| 1831/1845 [42:34<00:20, 1.44s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████▍| 1833/1845 [42:37<00:16, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. + 99%|██████████████████████████████████████████████████████████████████████████████▌| 1835/1845 [42:39<00:12, 1.27s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +100%|██████████████████████████████████████████████████████████████████████████████▌| 1836/1845 [42:40<00:11, 1.30s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +100%|██████████████████████████████████████████████████████████████████████████████▋| 1837/1845 [42:42<00:10, 1.35s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +100%|██████████████████████████████████████████████████████████████████████████████▋| 1839/1845 [42:45<00:08, 1.37s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +100%|██████████████████████████████████████████████████████████████████████████████▊| 1840/1845 [42:46<00:06, 1.36s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +100%|██████████████████████████████████████████████████████████████████████████████▊| 1842/1845 [42:49<00:04, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +100%|██████████████████████████████████████████████████████████████████████████████▉| 1843/1845 [42:50<00:02, 1.39s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +100%|███████████████████████████████████████████████████████████████████████████████| 1845/1845 [42:53<00:00, 1.40s/it]The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Configuration saved in ./checkpoint-5000/config.json The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Configuration saved in ./checkpoint-5000/config.json The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Feature extractor saved in ./checkpoint-5000/preprocessor_config.json The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +Feature extractor saved in ./checkpoint-5000/preprocessor_config.json The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +Feature extractor saved in ./checkpoint-5000/preprocessor_config.json The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +Feature extractor saved in ./checkpoint-5000/preprocessor_config.json The following columns in the evaluation set don't have a corresponding argument in `SpeechEncoderDecoderModel.forward` and have been ignored: length, lang. If length, lang are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message. +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +05/05/2022 03:31:44 - WARNING - huggingface_hub.repository - Adding files tracked by Git LFS: ['wandb/run-20220504_142129-1tmxz74i/logs/debug-internal.log']. This may take a bit of time if the files are large. +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... To disable this warning, you can either: - Avoid using `tokenizers` before the fork if possible