Training in progress, step 500

.ipynb_checkpoints/run-checkpoint.sh

@@ -3,11 +3,12 @@ python run_speech_recognition_ctc_bnb.py \
 	--model_name_or_path="facebook/wav2vec2-xls-r-1b" \
 	--dataset_config_name="ar" \
 	--output_dir="./" \
-	--num_train_epochs="20" \
-	--per_device_train_batch_size="32" \
+	--overwrite_output_dir \
+	--num_train_epochs="10" \
+	--per_device_train_batch_size="16" \
 	--per_device_eval_batch_size="8" \
 	--gradient_accumulation_steps="4" \
-	--learning_rate="1.5e-4" \
+	--learning_rate="6.5e-5" \
 	--warmup_steps="2000" \
 	--length_column_name="input_length" \
 	--evaluation_strategy="steps" \
@@ -30,4 +31,5 @@ python run_speech_recognition_ctc_bnb.py \
 	--fp16 \
 	--group_by_length \
 	--do_train --do_eval \
-	--push_to_hub
+	--push_to_hub
+    

previous_runs/.ipynb_checkpoints/run-checkpoint.sh

@@ -0,0 +1,35 @@
+python run_speech_recognition_ctc_bnb.py \
+	--dataset_name="mozilla-foundation/common_voice_8_0" \
+	--model_name_or_path="facebook/wav2vec2-xls-r-1b" \
+	--dataset_config_name="ar" \
+	--output_dir="./" \
+	--num_train_epochs="20" \
+	--per_device_train_batch_size="32" \
+	--per_device_eval_batch_size="8" \
+	--gradient_accumulation_steps="4" \
+	--learning_rate="1.5e-4" \
+	--warmup_steps="2000" \
+	--length_column_name="input_length" \
+	--evaluation_strategy="steps" \
+	--text_column_name="sentence" \
+	--chars_to_ignore , ? . ! \- \; \: \" “ % ‘ ” � — ’ … – \
+	--save_steps="500" \
+	--eval_steps="500" \
+	--logging_steps="100" \
+	--layerdrop="0.0" \
+	--activation_dropout="0.1" \
+	--save_total_limit="4" \
+	--freeze_feature_encoder \
+	--feat_proj_dropout="0.0" \
+	--mask_time_prob="0.75" \
+	--mask_time_length="10" \
+	--mask_feature_prob="0.25" \
+	--mask_feature_length="64" \
+	--gradient_checkpointing \
+	--use_auth_token \
+	--fp16 \
+	--group_by_length \
+	--do_train --do_eval \
+	--push_to_hub
+    
+# Poor performance. Stopped at {'eval_loss': 1.338714361190796, 'eval_wer': 0.897404591449766, 'eval_runtime': 422.6845, 'eval_samples_per_second': 24.576, 'eval_steps_per_second': 3.073, 'epoch': 10.07}

previous_runs/run.sh

@@ -0,0 +1,35 @@
+python run_speech_recognition_ctc_bnb.py \
+	--dataset_name="mozilla-foundation/common_voice_8_0" \
+	--model_name_or_path="facebook/wav2vec2-xls-r-1b" \
+	--dataset_config_name="ar" \
+	--output_dir="./" \
+	--num_train_epochs="20" \
+	--per_device_train_batch_size="32" \
+	--per_device_eval_batch_size="8" \
+	--gradient_accumulation_steps="4" \
+	--learning_rate="1.5e-4" \
+	--warmup_steps="2000" \
+	--length_column_name="input_length" \
+	--evaluation_strategy="steps" \
+	--text_column_name="sentence" \
+	--chars_to_ignore , ? . ! \- \; \: \" “ % ‘ ” � — ’ … – \
+	--save_steps="500" \
+	--eval_steps="500" \
+	--logging_steps="100" \
+	--layerdrop="0.0" \
+	--activation_dropout="0.1" \
+	--save_total_limit="4" \
+	--freeze_feature_encoder \
+	--feat_proj_dropout="0.0" \
+	--mask_time_prob="0.75" \
+	--mask_time_length="10" \
+	--mask_feature_prob="0.25" \
+	--mask_feature_length="64" \
+	--gradient_checkpointing \
+	--use_auth_token \
+	--fp16 \
+	--group_by_length \
+	--do_train --do_eval \
+	--push_to_hub
+    
+# Poor performance. Stopped at {'eval_loss': 1.338714361190796, 'eval_wer': 0.897404591449766, 'eval_runtime': 422.6845, 'eval_samples_per_second': 24.576, 'eval_steps_per_second': 3.073, 'epoch': 10.07}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6813b2b0872c64e7a3717070cac4d515a66932195b3780d1a534af395051a571
+oid sha256:72110b18b9f181a1796af0610a022c116fbc3b53316387df30a797bd09e691ad
 size 3850799473

run.sh

@@ -3,11 +3,12 @@ python run_speech_recognition_ctc_bnb.py \
 	--model_name_or_path="facebook/wav2vec2-xls-r-1b" \
 	--dataset_config_name="ar" \
 	--output_dir="./" \
-	--num_train_epochs="20" \
-	--per_device_train_batch_size="32" \
+	--overwrite_output_dir \
+	--num_train_epochs="10" \
+	--per_device_train_batch_size="16" \
 	--per_device_eval_batch_size="8" \
 	--gradient_accumulation_steps="4" \
-	--learning_rate="1.5e-4" \
+	--learning_rate="6.5e-5" \
 	--warmup_steps="2000" \
 	--length_column_name="input_length" \
 	--evaluation_strategy="steps" \
@@ -30,4 +31,5 @@ python run_speech_recognition_ctc_bnb.py \
 	--fp16 \
 	--group_by_length \
 	--do_train --do_eval \
-	--push_to_hub
+	--push_to_hub
+    

special_tokens_map.json

@@ -1 +1 @@
-{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76e7d1cf613331f4d89062dcf3e088ac0c1ea8e4317d97dfec009e3fad701c1e
+oid sha256:dd4fb4cf661cbec66ff39edfd085a48937505029ebdfd04e974b45d65bbc7c64
 size 2991