manual submit

run.sh

@@ -1,15 +1,15 @@
-python run_speech_recognition_ctc.py \
+WANDB_ENTITY=NbAiLab WANDB_PROJECT=wav2vec2 python run_speech_recognition_ctc.py \
         --dataset_name="NbAiLab/NPSC" \
         --model_name_or_path="KBLab/wav2vec2-large-voxrex" \
         --hub_model_id="NbAiLab/wav2vec2-large-voxrex-npsc-nynorsk" \
-        --dataset_config_name="16K_mp3_nynorsk" \
+        --dataset_config_name="16K_mp3" \
         --output_dir="./" \
         --overwrite_output_dir \
-        --num_train_epochs="40" \
+        --num_train_epochs="80" \
         --per_device_train_batch_size="16" \
         --per_device_eval_batch_size="16" \
         --gradient_accumulation_steps="2" \
-        --learning_rate="7.5e-5" \
+        --learning_rate="1e-4" \
         --warmup_steps="2000" \
         --length_column_name="input_length" \
         --evaluation_strategy="steps" \
@@ -29,7 +29,7 @@ python run_speech_recognition_ctc.py \
         --mask_feature_prob="0.25" \
         --mask_feature_length="64" \
         --gradient_checkpointing \
-        --min_duration_in_seconds="0.8" \
+        --min_duration_in_seconds="0.5" \
         --max_duration_in_seconds="30.0" \
         --use_auth_token \
         --seed="42" \
@@ -37,5 +37,4 @@ python run_speech_recognition_ctc.py \
         --group_by_length \
         --do_train --do_eval \
         --push_to_hub \
-        --preprocessing_num_workers="32"\
-	--ctc_zero_infinity=True
+        --preprocessing_num_workers="32"

run_freddy.sh

@@ -0,0 +1,41 @@
+python run_speech_recognition_ctc.py \
+        --dataset_name="NbAiLab/NPSC" \
+        --model_name_or_path="KBLab/wav2vec2-large-voxrex" \
+        --hub_model_id="NbAiLab/wav2vec2-large-voxrex-npsc-nynorsk" \
+        --dataset_config_name="16K_mp3_nynorsk" \
+        --output_dir="./" \
+        --overwrite_output_dir \
+        --num_train_epochs="40" \
+        --per_device_train_batch_size="16" \
+        --per_device_eval_batch_size="16" \
+        --gradient_accumulation_steps="2" \
+        --learning_rate="7.5e-5" \
+        --warmup_steps="2000" \
+        --length_column_name="input_length" \
+        --evaluation_strategy="steps" \
+        --text_column_name="text" \
+        --save_steps="500" \
+        --eval_steps="500" \
+        --logging_steps="100" \
+        --layerdrop="0.041" \
+        --attention_dropout="0.094" \
+        --activation_dropout="0.055" \
+        --hidden_dropout="0.047" \
+        --save_total_limit="3" \
+        --freeze_feature_encoder \
+        --feat_proj_dropout="0.04" \
+        --mask_time_prob="0.082" \
+        --mask_time_length="10" \
+        --mask_feature_prob="0.25" \
+        --mask_feature_length="64" \
+        --gradient_checkpointing \
+        --min_duration_in_seconds="0.8" \
+        --max_duration_in_seconds="30.0" \
+        --use_auth_token \
+        --seed="42" \
+        --fp16 \
+        --group_by_length \
+        --do_train --do_eval \
+        --push_to_hub \
+        --preprocessing_num_workers="32"\
+	--ctc_zero_infinity=True

vocab.json

@@ -1 +1 @@
-{"a": 1, "b": 2, "c": 3, "d": 4, "e": 5, "f": 6, "g": 7, "h": 8, "i": 9, "j": 10, "k": 11, "l": 12, "m": 13, "n": 14, "o": 15, "p": 16, "q": 17, "r": 18, "s": 19, "t": 20, "u": 21, "v": 22, "w": 23, "x": 24, "y": 25, "z": 26, "å": 27, "æ": 28, "ø": 29, "|": 0, "[UNK]": 30, "[PAD]": 31}
+{"a": 1, "b": 2, "c": 3, "d": 4, "e": 5, "f": 6, "g": 7, "h": 8, "i": 9, "j": 10, "k": 11, "l": 12, "m": 13, "n": 14, "o": 15, "p": 16, "q": 17, "r": 18, "s": 19, "t": 20, "u": 21, "v": 22, "w": 23, "x": 24, "y": 25, "z": 26, "\u00e5": 27, "\u00e6": 28, "\u00f8": 29, "|": 0, "[UNK]": 30, "[PAD]": 31}