File size: 1,175 Bytes
37f8bf6
 
 
 
 
c58dbaf
37f8bf6
 
7d61ebe
7132467
 
37f8bf6
7718d64
ef5fb31
37f8bf6
 
 
 
 
 
 
b0b32a4
 
37f8bf6
 
7d61ebe
37f8bf6
5200b6e
37f8bf6
 
 
 
ef5fb31
7d61ebe
37f8bf6
 
 
 
 
 
 
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
WANDB_PROJECT=auto-speech-recognition-french
python run_speech_recognition_ctc.py \
--dataset_name="mozilla-foundation/common_voice_8_0" \
--model_name_or_path="facebook/wav2vec2-xls-r-300m" \
--dataset_config_name="fr" \
--tokenizer_name_or_path="./" \
--output_dir="./" \
--overwrite_output_dir \
--num_train_epochs="4" \
--per_device_train_batch_size="64" \
--per_device_eval_batch_size="64" \
--gradient_accumulation_steps="1" \
--learning_rate="1e-4" \
--warmup_steps="2700" \
--length_column_name="input_length" \
--evaluation_strategy="steps" \
--text_column_name="sentence" \
--save_steps="500" \
--eval_steps="500" \
--logging_steps="100" \
--layerdrop="0.0" \
--activation_dropout="0.05" \
--save_total_limit="2" \
--freeze_feature_encoder \
--feat_proj_dropout="0.0" \
--mask_time_prob="0.6" \
--mask_time_length="10" \
--mask_feature_prob="0.25" \
--mask_feature_length="10" \
--gradient_checkpointing \
--report_to="wandb" \
--run_name="xls-r-300m-fr" \
--max_eval_samples="4500" \
--max_duration_in_seconds="10" \
--use_auth_token \
--fp16 \
--group_by_length \
--preprocessing_num_workers="64" \
--do_train --do_eval \
--load_best_model_at_end \
--push_to_hub