File size: 1,866 Bytes
2bd3cb5
 
 
1f8d800
 
 
 
2bd3cb5
 
 
 
 
 
1f8d800
 
 
2bd3cb5
1f8d800
2bd3cb5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu
python run_speech_recognition_seq2seq_streaming.py \
	--model_name_or_path="openai/whisper-small" \
	--train_dataset_names  "Finnish-NLP/youtube_fi_110k_preprocessed" "Finnish-NLP/fleurs_asr_fi_audio_processed" "Finnish-NLP/css10_audio_processed" "Finnish-NLP/voxpopuli_fi_audio_processed" "Finnish-NLP/parliament_audio_processed" "Finnish-NLP/aalto_eduskunta_asr_audio_processed" "Finnish-NLP/fbc_dialog_audio_processed" "Finnish-NLP/fbc_monolog_audio_processed" "mozilla-foundation/common_voice_11_0" \
	--train_dataset_config_names  "" "" "" "" "" "" "" "" "fi" \
    --train_dataset_split_name  "train" "train" "train" "train" "train" "train" "train" "train" "train" \
    --train_dataset_text_column_names "texts" "sentence" "sentence" "sentence" "sentence" "sentence" "sentence" "sentence" "sentence" \
    --eval_dataset_name="mozilla-foundation/common_voice_11_0" \
    --eval_dataset_config_name="fi" \
    --eval_split_name="test" \
	--language="finnish" \
	--model_index_name="Whisper Small Finnish" \
	--max_steps="10000" \
	--output_dir="./" \
	--per_device_train_batch_size="64" \
	--per_device_eval_batch_size="32" \
	--logging_steps="25" \
	--learning_rate="3e-6" \
	--warmup_steps="500" \
	--evaluation_strategy="steps" \
	--eval_steps="1000" \
	--save_strategy="steps" \
	--save_steps="1000" \
	--generation_max_length="225" \
	--length_column_name="input_length" \
	--max_duration_in_seconds="30" \
	--text_column_name="sentence" \
	--freeze_feature_encoder="False" \
	--report_to="tensorboard" \
	--metric_for_best_model="wer" \
	--greater_is_better="False" \
    --load_best_model_at_end \
    --optim="adamw_bnb_8bit" \
	--gradient_checkpointing \
	--fp16 \
	--overwrite_output_dir \
	--do_train \
	--do_eval \
	--predict_with_generate \
	--do_normalize_eval \
	--use_auth_token \
	--push_to_hub