/usr/data/condor/execute/dir_27392/whisper_slu 01/04/2024 22:52:21 - WARNING - __main__ - Process rank: 0, device: cuda:0, n_gpu: 1distributed training: True, 16-bits training: True 01/04/2024 22:52:21 - INFO - __main__ - Training/evaluation parameters Seq2SeqTrainingArguments( _n_gpu=1, adafactor=False, adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, auto_find_batch_size=False, bf16=False, bf16_full_eval=False, data_seed=None, dataloader_drop_last=False, dataloader_num_workers=0, dataloader_persistent_workers=False, dataloader_pin_memory=True, ddp_backend=None, ddp_broadcast_buffers=None, ddp_bucket_cap_mb=None, ddp_find_unused_parameters=None, ddp_timeout=1800, debug=[], deepspeed=None, disable_tqdm=True, dispatch_batches=None, do_eval=True, do_predict=False, do_train=True, eval_accumulation_steps=None, eval_delay=0, eval_steps=200, evaluation_strategy=steps, fp16=True, fp16_backend=auto, fp16_full_eval=False, fp16_opt_level=O1, fsdp=[], fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, fsdp_min_num_params=0, fsdp_transformer_layer_cls_to_wrap=None, full_determinism=False, generation_config=None, generation_max_length=225, generation_num_beams=None, gradient_accumulation_steps=8, gradient_checkpointing=False, gradient_checkpointing_kwargs=None, greater_is_better=None, group_by_length=False, half_precision_backend=auto, hub_always_push=False, hub_model_id=None, hub_private_repo=False, hub_strategy=every_save, hub_token=, ignore_data_skip=False, include_inputs_for_metrics=False, include_num_input_tokens_seen=False, include_tokens_per_second=False, jit_mode_eval=False, label_names=None, label_smoothing_factor=0.0, learning_rate=0.0001, length_column_name=length, load_best_model_at_end=False, local_rank=0, log_level=info, log_level_replica=warning, log_on_each_node=True, logging_dir=/esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner/runs/Jan04_22-52-21_etrog.esat.kuleuven.be, logging_first_step=False, logging_nan_inf_filter=True, logging_steps=100, logging_strategy=steps, lr_scheduler_kwargs={}, lr_scheduler_type=cosine, max_grad_norm=1.0, max_steps=5000, metric_for_best_model=None, mp_parameters=, neftune_noise_alpha=None, no_cuda=False, num_train_epochs=3.0, optim=adamw_torch, optim_args=None, output_dir=/esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner, overwrite_output_dir=True, past_index=-1, per_device_eval_batch_size=8, per_device_train_batch_size=16, predict_with_generate=True, prediction_loss_only=False, push_to_hub=False, push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=, ray_scope=last, remove_unused_columns=True, report_to=['wandb'], resume_from_checkpoint=None, run_name=/esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner, save_on_each_node=False, save_only_model=False, save_safetensors=True, save_steps=200, save_strategy=steps, save_total_limit=None, seed=42, skip_memory_metrics=True, sortish_sampler=False, split_batches=False, tf32=None, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, torchdynamo=None, tpu_metrics_debug=False, tpu_num_cores=None, use_cpu=False, use_ipex=False, use_legacy_prediction_loop=False, use_mps_device=False, warmup_ratio=0.0, warmup_steps=500, weight_decay=0.0, ) 01/04/2024 22:52:21 - INFO - __main__ - Training/evaluation parameters Seq2SeqTrainingArguments( _n_gpu=1, adafactor=False, adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, auto_find_batch_size=False, bf16=False, bf16_full_eval=False, data_seed=None, dataloader_drop_last=False, dataloader_num_workers=0, dataloader_persistent_workers=False, dataloader_pin_memory=True, ddp_backend=None, ddp_broadcast_buffers=None, ddp_bucket_cap_mb=None, ddp_find_unused_parameters=None, ddp_timeout=1800, debug=[], deepspeed=None, disable_tqdm=True, dispatch_batches=None, do_eval=True, do_predict=False, do_train=True, eval_accumulation_steps=None, eval_delay=0, eval_steps=200, evaluation_strategy=steps, fp16=True, fp16_backend=auto, fp16_full_eval=False, fp16_opt_level=O1, fsdp=[], fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, fsdp_min_num_params=0, fsdp_transformer_layer_cls_to_wrap=None, full_determinism=False, generation_config=None, generation_max_length=225, generation_num_beams=None, gradient_accumulation_steps=8, gradient_checkpointing=False, gradient_checkpointing_kwargs=None, greater_is_better=None, group_by_length=False, half_precision_backend=auto, hub_always_push=False, hub_model_id=None, hub_private_repo=False, hub_strategy=every_save, hub_token=, ignore_data_skip=False, include_inputs_for_metrics=False, include_num_input_tokens_seen=False, include_tokens_per_second=False, jit_mode_eval=False, label_names=None, label_smoothing_factor=0.0, learning_rate=0.0001, length_column_name=length, load_best_model_at_end=False, local_rank=0, log_level=info, log_level_replica=warning, log_on_each_node=True, logging_dir=/esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner/runs/Jan04_22-52-21_etrog.esat.kuleuven.be, logging_first_step=False, logging_nan_inf_filter=True, logging_steps=100, logging_strategy=steps, lr_scheduler_kwargs={}, lr_scheduler_type=cosine, max_grad_norm=1.0, max_steps=5000, metric_for_best_model=None, mp_parameters=, neftune_noise_alpha=None, no_cuda=False, num_train_epochs=3.0, optim=adamw_torch, optim_args=None, output_dir=/esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner, overwrite_output_dir=True, past_index=-1, per_device_eval_batch_size=8, per_device_train_batch_size=16, predict_with_generate=True, prediction_loss_only=False, push_to_hub=False, push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=, ray_scope=last, remove_unused_columns=True, report_to=['wandb'], resume_from_checkpoint=None, run_name=/esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner, save_on_each_node=False, save_only_model=False, save_safetensors=True, save_steps=200, save_strategy=steps, save_total_limit=None, seed=42, skip_memory_metrics=True, sortish_sampler=False, split_batches=False, tf32=None, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, torchdynamo=None, tpu_metrics_debug=False, tpu_num_cores=None, use_cpu=False, use_ipex=False, use_legacy_prediction_loop=False, use_mps_device=False, warmup_ratio=0.0, warmup_steps=500, weight_decay=0.0, ) 01/04/2024 22:52:27 - WARNING - __main__ - Resize embeddings to account for the new tokens 01/04/2024 22:52:27 - INFO - __main__ - WhisperForConditionalGeneration( 01/04/2024 22:52:27 - INFO - __main__ - (model): WhisperModel( 01/04/2024 22:52:27 - INFO - __main__ - (encoder): WhisperEncoder( 01/04/2024 22:52:27 - INFO - __main__ - (conv1): Conv1d(80, 768, kernel_size=(3,), stride=(1,), padding=(1,)) 01/04/2024 22:52:27 - INFO - __main__ - (conv2): Conv1d(768, 768, kernel_size=(3,), stride=(2,), padding=(1,)) 01/04/2024 22:52:27 - INFO - __main__ - (embed_positions): Embedding(1500, 768) 01/04/2024 22:52:27 - INFO - __main__ - (layers): ModuleList( 01/04/2024 22:52:27 - INFO - __main__ - (0-11): 12 x WhisperEncoderLayer( 01/04/2024 22:52:27 - INFO - __main__ - (self_attn): WhisperAttention( 01/04/2024 22:52:27 - INFO - __main__ - (k_proj): Linear(in_features=768, out_features=768, bias=False) 01/04/2024 22:52:27 - INFO - __main__ - (v_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (q_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (out_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - (self_attn_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) 01/04/2024 22:52:27 - INFO - __main__ - (activation_fn): GELUActivation() 01/04/2024 22:52:27 - INFO - __main__ - (fc1): Linear(in_features=768, out_features=3072, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (fc2): Linear(in_features=3072, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (final_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - (layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - (decoder): WhisperDecoder( 01/04/2024 22:52:27 - INFO - __main__ - (embed_tokens): Embedding(51885, 768) 01/04/2024 22:52:27 - INFO - __main__ - (embed_positions): WhisperPositionalEmbedding(448, 768) 01/04/2024 22:52:27 - INFO - __main__ - (layers): ModuleList( 01/04/2024 22:52:27 - INFO - __main__ - (0-11): 12 x WhisperDecoderLayer( 01/04/2024 22:52:27 - INFO - __main__ - (self_attn): WhisperAttention( 01/04/2024 22:52:27 - INFO - __main__ - (k_proj): Linear(in_features=768, out_features=768, bias=False) 01/04/2024 22:52:27 - INFO - __main__ - (v_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (q_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (out_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - (activation_fn): GELUActivation() 01/04/2024 22:52:27 - INFO - __main__ - (self_attn_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) 01/04/2024 22:52:27 - INFO - __main__ - (encoder_attn): WhisperAttention( 01/04/2024 22:52:27 - INFO - __main__ - (k_proj): Linear(in_features=768, out_features=768, bias=False) 01/04/2024 22:52:27 - INFO - __main__ - (v_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (q_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (out_proj): Linear(in_features=768, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - (encoder_attn_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) 01/04/2024 22:52:27 - INFO - __main__ - (fc1): Linear(in_features=768, out_features=3072, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (fc2): Linear(in_features=3072, out_features=768, bias=True) 01/04/2024 22:52:27 - INFO - __main__ - (final_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - (layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - (proj_out): Linear(in_features=768, out_features=51885, bias=False) 01/04/2024 22:52:27 - INFO - __main__ - ) 01/04/2024 22:52:27 - INFO - __main__ - Loaded model with 241,750,272 parameters, of which 153,596,160 require gradients 01/04/2024 22:52:41 - INFO - __main__ - Loaded preprocessed dataset from /usr/data/condor/execute/dir_27392/data/end2end/vectorized_dataset {'loss': 1.1831, 'learning_rate': 2e-05, 'epoch': 0.18} {'loss': 0.3562, 'learning_rate': 4e-05, 'epoch': 0.36} {'eval_loss': 0.3264639973640442, 'eval_wer': 0.19195022054190297, 'eval_runtime': 194.3024, 'eval_samples_per_second': 5.147, 'eval_steps_per_second': 0.643, 'epoch': 0.36} {'loss': 0.3149, 'learning_rate': 6e-05, 'epoch': 0.54} {'loss': 0.3149, 'learning_rate': 8e-05, 'epoch': 0.71} {'eval_loss': 0.313632994890213, 'eval_wer': 0.18415248897290484, 'eval_runtime': 190.5904, 'eval_samples_per_second': 5.247, 'eval_steps_per_second': 0.656, 'epoch': 0.71} {'loss': 0.3164, 'learning_rate': 0.0001, 'epoch': 0.89} {'loss': 0.2778, 'learning_rate': 9.987820251299122e-05, 'epoch': 1.07} {'eval_loss': 0.3204321563243866, 'eval_wer': 0.1786389413988658, 'eval_runtime': 190.0545, 'eval_samples_per_second': 5.262, 'eval_steps_per_second': 0.658, 'epoch': 1.07} {'loss': 0.2254, 'learning_rate': 9.951340343707852e-05, 'epoch': 1.25} {'loss': 0.2288, 'learning_rate': 9.890738003669029e-05, 'epoch': 1.43} {'eval_loss': 0.31555524468421936, 'eval_wer': 0.17166824196597352, 'eval_runtime': 189.1302, 'eval_samples_per_second': 5.287, 'eval_steps_per_second': 0.661, 'epoch': 1.43} {'loss': 0.2295, 'learning_rate': 9.806308479691595e-05, 'epoch': 1.61} {'loss': 0.2307, 'learning_rate': 9.698463103929542e-05, 'epoch': 1.79} {'eval_loss': 0.30555686354637146, 'eval_wer': 0.17080182734719596, 'eval_runtime': 186.4985, 'eval_samples_per_second': 5.362, 'eval_steps_per_second': 0.67, 'epoch': 1.79} {'loss': 0.2281, 'learning_rate': 9.567727288213005e-05, 'epoch': 1.97} {'loss': 0.1482, 'learning_rate': 9.414737964294636e-05, 'epoch': 2.14} {'eval_loss': 0.3138311207294464, 'eval_wer': 0.16816320100819154, 'eval_runtime': 188.367, 'eval_samples_per_second': 5.309, 'eval_steps_per_second': 0.664, 'epoch': 2.14} {'loss': 0.1326, 'learning_rate': 9.24024048078213e-05, 'epoch': 2.32} {'loss': 0.1368, 'learning_rate': 9.045084971874738e-05, 'epoch': 2.5} {'eval_loss': 0.3136378824710846, 'eval_wer': 0.16560333963453056, 'eval_runtime': 188.5481, 'eval_samples_per_second': 5.304, 'eval_steps_per_second': 0.663, 'epoch': 2.5} {'loss': 0.1415, 'learning_rate': 8.83022221559489e-05, 'epoch': 2.68} {'loss': 0.1405, 'learning_rate': 8.596699001693255e-05, 'epoch': 2.86} {'eval_loss': 0.30823495984077454, 'eval_wer': 0.1617044738500315, 'eval_runtime': 188.0037, 'eval_samples_per_second': 5.319, 'eval_steps_per_second': 0.665, 'epoch': 2.86} {'loss': 0.1271, 'learning_rate': 8.345653031794292e-05, 'epoch': 3.04} {'loss': 0.0639, 'learning_rate': 8.07830737662829e-05, 'epoch': 3.22} {'eval_loss': 0.3200896680355072, 'eval_wer': 0.1611925015752993, 'eval_runtime': 188.8387, 'eval_samples_per_second': 5.296, 'eval_steps_per_second': 0.662, 'epoch': 3.22} {'loss': 0.0644, 'learning_rate': 7.795964517353735e-05, 'epoch': 3.4} {'loss': 0.0673, 'learning_rate': 7.500000000000001e-05, 'epoch': 3.57} {'eval_loss': 0.3242465853691101, 'eval_wer': 0.1611925015752993, 'eval_runtime': 187.6919, 'eval_samples_per_second': 5.328, 'eval_steps_per_second': 0.666, 'epoch': 3.57} {'loss': 0.0692, 'learning_rate': 7.191855733945387e-05, 'epoch': 3.75} {'loss': 0.0688, 'learning_rate': 6.873032967079561e-05, 'epoch': 3.93} {'eval_loss': 0.3234550952911377, 'eval_wer': 0.15835696282293635, 'eval_runtime': 188.2056, 'eval_samples_per_second': 5.313, 'eval_steps_per_second': 0.664, 'epoch': 3.93} {'loss': 0.0424, 'learning_rate': 6.545084971874738e-05, 'epoch': 4.11} {'loss': 0.0227, 'learning_rate': 6.209609477998338e-05, 'epoch': 4.29} {'eval_loss': 0.3419637382030487, 'eval_wer': 0.15575771896660365, 'eval_runtime': 187.5691, 'eval_samples_per_second': 5.331, 'eval_steps_per_second': 0.666, 'epoch': 4.29} {'loss': 0.0227, 'learning_rate': 5.868240888334653e-05, 'epoch': 4.47} {'loss': 0.0232, 'learning_rate': 5.522642316338268e-05, 'epoch': 4.65} {'eval_loss': 0.343047559261322, 'eval_wer': 0.15248897290485192, 'eval_runtime': 187.938, 'eval_samples_per_second': 5.321, 'eval_steps_per_second': 0.665, 'epoch': 4.65} {'loss': 0.0235, 'learning_rate': 5.174497483512506e-05, 'epoch': 4.83} {'loss': 0.0229, 'learning_rate': 4.825502516487497e-05, 'epoch': 5.0} {'eval_loss': 0.3450343608856201, 'eval_wer': 0.15276465028355388, 'eval_runtime': 187.1866, 'eval_samples_per_second': 5.342, 'eval_steps_per_second': 0.668, 'epoch': 5.0} {'loss': 0.0071, 'learning_rate': 4.477357683661734e-05, 'epoch': 5.18} {'loss': 0.0064, 'learning_rate': 4.131759111665349e-05, 'epoch': 5.36} {'eval_loss': 0.36311542987823486, 'eval_wer': 0.1498109640831758, 'eval_runtime': 188.0361, 'eval_samples_per_second': 5.318, 'eval_steps_per_second': 0.665, 'epoch': 5.36} {'loss': 0.0062, 'learning_rate': 3.790390522001662e-05, 'epoch': 5.54} {'loss': 0.0059, 'learning_rate': 3.4549150281252636e-05, 'epoch': 5.72} {'eval_loss': 0.36524486541748047, 'eval_wer': 0.1482356647763075, 'eval_runtime': 188.0737, 'eval_samples_per_second': 5.317, 'eval_steps_per_second': 0.665, 'epoch': 5.72} {'loss': 0.0057, 'learning_rate': 3.12696703292044e-05, 'epoch': 5.9} {'loss': 0.0043, 'learning_rate': 2.8081442660546125e-05, 'epoch': 6.08} {'eval_loss': 0.3756215274333954, 'eval_wer': 0.14815689981096408, 'eval_runtime': 187.8897, 'eval_samples_per_second': 5.322, 'eval_steps_per_second': 0.665, 'epoch': 6.08} {'loss': 0.0023, 'learning_rate': 2.500000000000001e-05, 'epoch': 6.26} {'loss': 0.0021, 'learning_rate': 2.2040354826462668e-05, 'epoch': 6.43} {'eval_loss': 0.3798007071018219, 'eval_wer': 0.1476843100189036, 'eval_runtime': 187.5918, 'eval_samples_per_second': 5.331, 'eval_steps_per_second': 0.666, 'epoch': 6.43} {'loss': 0.0021, 'learning_rate': 1.9216926233717085e-05, 'epoch': 6.61} {'loss': 0.002, 'learning_rate': 1.6543469682057106e-05, 'epoch': 6.79} {'eval_loss': 0.3823593854904175, 'eval_wer': 0.14839319470699433, 'eval_runtime': 187.1697, 'eval_samples_per_second': 5.343, 'eval_steps_per_second': 0.668, 'epoch': 6.79} {'loss': 0.0019, 'learning_rate': 1.4033009983067452e-05, 'epoch': 6.97} {'loss': 0.0014, 'learning_rate': 1.1697777844051105e-05, 'epoch': 7.15} {'eval_loss': 0.3876166343688965, 'eval_wer': 0.1471329552614997, 'eval_runtime': 186.567, 'eval_samples_per_second': 5.36, 'eval_steps_per_second': 0.67, 'epoch': 7.15} {'loss': 0.0013, 'learning_rate': 9.549150281252633e-06, 'epoch': 7.33} {'loss': 0.0013, 'learning_rate': 7.597595192178702e-06, 'epoch': 7.51} {'eval_loss': 0.3900451064109802, 'eval_wer': 0.14729048519218652, 'eval_runtime': 187.3398, 'eval_samples_per_second': 5.338, 'eval_steps_per_second': 0.667, 'epoch': 7.51} {'loss': 0.0013, 'learning_rate': 5.852620357053651e-06, 'epoch': 7.69} {'loss': 0.0013, 'learning_rate': 4.322727117869951e-06, 'epoch': 7.86} {'eval_loss': 0.39165398478507996, 'eval_wer': 0.14610901071203528, 'eval_runtime': 187.1352, 'eval_samples_per_second': 5.344, 'eval_steps_per_second': 0.668, 'epoch': 7.86} {'loss': 0.0013, 'learning_rate': 3.0153689607045845e-06, 'epoch': 8.04} {'loss': 0.0012, 'learning_rate': 1.9369152030840556e-06, 'epoch': 8.22} {'eval_loss': 0.39290326833724976, 'eval_wer': 0.1462271581600504, 'eval_runtime': 187.7766, 'eval_samples_per_second': 5.325, 'eval_steps_per_second': 0.666, 'epoch': 8.22} {'loss': 0.0011, 'learning_rate': 1.0926199633097157e-06, 'epoch': 8.4} {'loss': 0.0012, 'learning_rate': 4.865965629214819e-07, 'epoch': 8.58} {'eval_loss': 0.3932369351387024, 'eval_wer': 0.14654221802142406, 'eval_runtime': 187.8702, 'eval_samples_per_second': 5.323, 'eval_steps_per_second': 0.665, 'epoch': 8.58} {'loss': 0.0012, 'learning_rate': 1.2179748700879012e-07, 'epoch': 8.76} {'loss': 0.0012, 'learning_rate': 0.0, 'epoch': 8.94} {'eval_loss': 0.3932703733444214, 'eval_wer': 0.14642407057340895, 'eval_runtime': 186.9555, 'eval_samples_per_second': 5.349, 'eval_steps_per_second': 0.669, 'epoch': 8.94} {'train_runtime': 20915.2691, 'train_samples_per_second': 30.6, 'train_steps_per_second': 0.239, 'train_loss': 0.1056724779933691, 'epoch': 8.94} ***** train metrics ***** epoch = 8.94 train_loss = 0.1057 train_runtime = 5:48:35.26 train_samples_per_second = 30.6 train_steps_per_second = 0.239 01/05/2024 04:41:23 - INFO - __main__ - *** Evaluate *** {'eval_loss': 0.3932703733444214, 'eval_wer': 0.14642407057340895, 'eval_runtime': 186.6982, 'eval_samples_per_second': 5.356, 'eval_steps_per_second': 0.67, 'epoch': 8.94} ***** eval metrics ***** epoch = 8.94 eval_loss = 0.3933 eval_runtime = 0:03:06.69 eval_samples = 1000 eval_samples_per_second = 5.356 eval_steps_per_second = 0.67 eval_wer = 0.1464