wav2vec2-german / run_dummy.sh
patrickvonplaten's picture
adapt config to robust wav2vec2; play around with dummy run
1a847ef
raw
history blame contribute delete
707 Bytes
#!/usr/bin/env bash
./run_wav2vec2_pretrain_flax.py \
--output_dir="./dummy" \
--speech_file_column="path" \
--num_train_epochs="1" \
--per_device_train_batch_size="4" \
--per_device_eval_batch_size="4" \
--validation_split_percentage="50" \
--learning_rate="5e-4" \
--weight_decay="0.01" \
--warmup_steps="200" \
--model_name_or_path="./" \
--dataset_name="common_voice" \
--dataset_config_name="cnh" \
--preprocessing_num_workers="4" \
--max_duration_in_seconds="20.0" \
--adam_beta1="0.9" \
--adam_beta2="0.98" \
--dtype="bfloat16" \
--cache_dir="/home/wav2vec2-experiments/data_cache/" \
--pad_to_multiple_of="32768" \
--push_to_hub