File size: 1,261 Bytes
ee6793a
 
 
 
9169ca7
ee6793a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9169ca7
ee6793a
 
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
python -m torch.distributed.launch --nproc_per_node=1 run_speech_recognition_seq2seq_streaming.py \
	--model_name_or_path="openai/whisper-large-v2" \
	--dataset_name="NbAiLab/NCC_S" \
	--language="Norwegian" \
	--resume_from_checkpoint="./checkpoint-5000" \
	--streaming \
	--train_split_name="train" \
	--eval_split_name="validation" \
	--model_index_name="Whisper Large Norwegian" \
	--output_dir="./" \
	--max_steps="5000" \
	--per_device_train_batch_size="12" \
	--per_device_eval_batch_size="6" \
	--gradient_accumulation_steps="1" \
	--logging_steps="25" \
	--learning_rate="1e-5" \
	--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="text" \
	--freeze_feature_encoder="false" \
	--report_to="tensorboard" \
	--metric_for_best_model="wer" \
	--greater_is_better="false" \
	--load_best_model_at_end \
	--gradient_checkpointing="true" \
	--fp16="true" \
	--do_train \
	--do_eval \
	--predict_with_generate \
	--do_normalize_eval \
	--use_auth_token \
	--overwrite_output_dir="false" \
	--push_to_hub_model_id="whisper-large-nob-ncc-s" \
	--push_to_hub