File size: 2,133 Bytes
5c23c43 497a54e 5c23c43 497a54e 5c23c43 497a54e 5c23c43 497a54e 60b3f40 497a54e 60b3f40 497a54e |
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 43 44 45 46 47 48 49 50 |
python run_interleave.py \
--model_name_or_path 'emilios/whisper-medium-el' \
--model_revision main \
--do_train True \
--do_eval True \
--freeze_feature_encoder False \
--freeze_encoder False \
--model_index_name 'Whisper Medium El Greco' \
--dataset_name 'mozilla-foundation/common_voice_11_0,google/fleurs' \
--dataset_config_name 'el,el_gr' \
--train_split_name 'train+validation,train+validation' \
--eval_split_name 'test,-' \
--text_column_name 'sentence,transcription' \
--audio_column_name 'audio,audio' \
--streaming False \
--max_duration_in_seconds 30 \
--do_lower_case False \
--do_remove_punctuation False \
--do_normalize_eval True \
--language greek \
--task transcribe \
--shuffle_buffer_size 500 \
--output_dir './' \
--overwrite_output_dir True \
--per_device_train_batch_size 32 \
--gradient_accumulation_steps 1 \
--per_device_eval_batch_size 16 \
--learning_rate 1e-5 \
--dropout 0.1 \
--warmup_steps 500 \
--max_steps 5000 \
--resume_from_checkpoint="4000" \
--eval_steps 1000 \
--gradient_checkpointing True \
--cache_dir '~/.cache' \
--fp16 True \
--evaluation_strategy steps \
--predict_with_generate True \
--generation_max_length 225 \
--save_steps 1000 \
--logging_steps 25 \
--report_to tensorboard \
--load_best_model_at_end True \
--metric_for_best_model wer \
--greater_is_better False \
--push_to_hub True
|