File size: 1,085 Bytes
1b49ff8 |
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 |
python run_speech_recognition_ctc.py \
--dataset_name="mozilla-foundation/common_voice_8_0" \
--model_name_or_path="facebook/wav2vec2-xls-r-1b" \
--dataset_config_name="nl" \
--output_dir="./" \
--overwrite_output_dir \
--num_train_epochs="50" \
--per_device_train_batch_size="8" \
--per_device_eval_batch_size="8" \
--gradient_accumulation_steps="5" \
--learning_rate="5e-5" \
--warmup_steps="500" \
--length_column_name="input_length" \
--evaluation_strategy="steps" \
--text_column_name="sentence" \
--save_steps="500" \
--chars_to_ignore , ´ ? \= . ! \- \; \: \" “ % ‘ ” � — ’ … – \
--eval_steps="500" \
--logging_steps="100" \
--layerdrop="0.0" \
--activation_dropout="0.1" \
--save_total_limit="3" \
--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 \
--report_to="tensorboard" \
--push_to_hub |