File size: 1,248 Bytes
ed3e91a |
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 |
deepspeed run_speech_recognition_seq2seq_streaming.py \
--deepspeed="ds_config.json" \
--model_name_or_path="openai/whisper-medium" \
--datasets_config_file="/home/ubuntu/whisper-medium-is/ds_config.json" \
--language="icelandic" \
--train_split_name="train" \
--eval_dataset_name="language-and-voice-lab/samromur_asr" \
--eval_split_name="validation" \
--eval_steps="1000" \
--max_eval_samples="1000" \
--model_index_name="Whisper medium Icelandic" \
--max_steps="20000" \
--output_dir="./" \
--per_device_train_batch_size="64" \
--per_device_eval_batch_size="32" \
--logging_steps="25" \
--learning_rate="1e-5" \
--warmup_steps="500" \
--evaluation_strategy="steps" \
--save_strategy="steps" \
--save_steps="6000" \
--generation_max_length="225" \
--length_column_name="input_length" \
--max_duration_in_seconds="30" \
--text_column_name="normalized_text" \
--freeze_feature_encoder="False" \
--report_to="tensorboard" \
--metric_for_best_model="wer" \
--greater_is_better="False" \
--load_best_model_at_end \
--gradient_checkpointing \
--fp16 \
--ignore_data_skip \
--do_train \
--do_eval \
--predict_with_generate \
--do_normalize_eval \
--streaming \
--push_to_hub \
--use_auth_token \
--mapper="/home/ubuntu/whisper/text_mapping"
|