|
Process rank: 0, device: cuda:0, n_gpu: 1distributed training: True, 16-bits training: True |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py:1070: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead. |
|
warnings.warn( |
|
[INFO|configuration_utils.py:737] 2024-01-08 11:04:51,653 >> loading configuration file configs/whisper_small_ner.json |
|
[INFO|configuration_utils.py:802] 2024-01-08 11:04:51,654 >> Model config WhisperConfig { |
|
"_name_or_path": "configs/whisper_small_ner.json", |
|
"activation_dropout": 0.0, |
|
"activation_function": "gelu", |
|
"adaptor_layernorm": true, |
|
"apply_spec_augment": false, |
|
"architectures": [ |
|
"WhisperForConditionalGeneration" |
|
], |
|
"attention_dropout": 0.0, |
|
"begin_suppress_tokens": [ |
|
220, |
|
50257 |
|
], |
|
"bos_token_id": 50257, |
|
"classifier_proj_size": 256, |
|
"d_model": 768, |
|
"decoder_attention_heads": 12, |
|
"decoder_ffn_dim": 3072, |
|
"decoder_layerdrop": 0.0, |
|
"decoder_layers": 12, |
|
"decoder_start_token_id": 50258, |
|
"dropout": 0.0, |
|
"encoder_attention_heads": 12, |
|
"encoder_ffn_dim": 3072, |
|
"encoder_layerdrop": 0.0, |
|
"encoder_layers": 12, |
|
"eos_token_id": 50257, |
|
"forced_decoder_ids": [ |
|
[ |
|
1, |
|
50259 |
|
], |
|
[ |
|
2, |
|
50359 |
|
], |
|
[ |
|
3, |
|
50363 |
|
] |
|
], |
|
"init_std": 0.02, |
|
"is_encoder_decoder": true, |
|
"mask_feature_length": 10, |
|
"mask_feature_min_masks": 0, |
|
"mask_feature_prob": 0.0, |
|
"mask_time_length": 10, |
|
"mask_time_min_masks": 2, |
|
"mask_time_prob": 0.05, |
|
"max_length": 448, |
|
"max_source_positions": 1500, |
|
"max_target_positions": 448, |
|
"median_filter_width": 7, |
|
"model_type": "whisper", |
|
"num_hidden_layers": 12, |
|
"num_mel_bins": 80, |
|
"pad_token_id": 50257, |
|
"scale_embedding": false, |
|
"slu_attention_heads": 12, |
|
"slu_dropout": 0.3, |
|
"slu_embed_dim": 768, |
|
"slu_focus": 1.0, |
|
"slu_input_from": "decoder", |
|
"slu_input_layers": -1, |
|
"slu_layers": 2, |
|
"slu_output_dim": 37, |
|
"slu_weight": 1.0, |
|
"suppress_tokens": [ |
|
1, |
|
2, |
|
7, |
|
8, |
|
9, |
|
10, |
|
14, |
|
25, |
|
26, |
|
27, |
|
28, |
|
29, |
|
31, |
|
58, |
|
59, |
|
60, |
|
61, |
|
62, |
|
63, |
|
90, |
|
91, |
|
92, |
|
93, |
|
359, |
|
503, |
|
522, |
|
542, |
|
873, |
|
893, |
|
902, |
|
918, |
|
922, |
|
931, |
|
1350, |
|
1853, |
|
1982, |
|
2460, |
|
2627, |
|
3246, |
|
3253, |
|
3268, |
|
3536, |
|
3846, |
|
3961, |
|
4183, |
|
4667, |
|
6585, |
|
6647, |
|
7273, |
|
9061, |
|
9383, |
|
10428, |
|
10929, |
|
11938, |
|
12033, |
|
12331, |
|
12562, |
|
13793, |
|
14157, |
|
14635, |
|
15265, |
|
15618, |
|
16553, |
|
16604, |
|
18362, |
|
18956, |
|
20075, |
|
21675, |
|
22520, |
|
26130, |
|
26161, |
|
26435, |
|
28279, |
|
29464, |
|
31650, |
|
32302, |
|
32470, |
|
36865, |
|
42863, |
|
47425, |
|
49870, |
|
50254, |
|
50258, |
|
50360, |
|
50361, |
|
50362 |
|
], |
|
"task": "token_classification", |
|
"torch_dtype": "float32", |
|
"transformers_version": "4.37.0.dev0", |
|
"use_cache": true, |
|
"use_crf": false, |
|
"use_weighted_layer_sum": false, |
|
"vocab_size": 51865 |
|
} |
|
|
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/transformers/models/auto/feature_extraction_auto.py:328: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead. |
|
warnings.warn( |
|
[INFO|feature_extraction_utils.py:537] 2024-01-08 11:04:51,810 >> loading configuration file preprocessor_config.json from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/preprocessor_config.json |
|
[INFO|feature_extraction_utils.py:579] 2024-01-08 11:04:51,814 >> Feature extractor WhisperFeatureExtractor { |
|
"chunk_length": 30, |
|
"feature_extractor_type": "WhisperFeatureExtractor", |
|
"feature_size": 80, |
|
"hop_length": 160, |
|
"n_fft": 400, |
|
"n_samples": 480000, |
|
"nb_max_frames": 3000, |
|
"padding_side": "right", |
|
"padding_value": 0.0, |
|
"processor_class": "WhisperProcessor", |
|
"return_attention_mask": false, |
|
"sampling_rate": 16000 |
|
} |
|
|
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/transformers/tokenization_utils_base.py:1899: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead. |
|
warnings.warn( |
|
[INFO|tokenization_utils_base.py:2026] 2024-01-08 11:04:51,980 >> loading file vocab.json from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/vocab.json |
|
[INFO|tokenization_utils_base.py:2026] 2024-01-08 11:04:51,980 >> loading file tokenizer.json from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/tokenizer.json |
|
[INFO|tokenization_utils_base.py:2026] 2024-01-08 11:04:51,980 >> loading file merges.txt from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/merges.txt |
|
[INFO|tokenization_utils_base.py:2026] 2024-01-08 11:04:51,980 >> loading file normalizer.json from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/normalizer.json |
|
[INFO|tokenization_utils_base.py:2026] 2024-01-08 11:04:51,980 >> loading file added_tokens.json from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/added_tokens.json |
|
[INFO|tokenization_utils_base.py:2026] 2024-01-08 11:04:51,980 >> loading file special_tokens_map.json from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/special_tokens_map.json |
|
[INFO|tokenization_utils_base.py:2026] 2024-01-08 11:04:51,980 >> loading file tokenizer_config.json from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/tokenizer_config.json |
|
[WARNING|tokenization_utils_base.py:2140] 2024-01-08 11:04:51,980 >> The tokenizer class you load from this checkpoint is not the same type as the class this function is called from. It may result in unexpected tokenization. |
|
The tokenizer class you load from this checkpoint is 'WhisperTokenizer'. |
|
The class this function is called from is 'NERTokenizerEndToEndFast'. |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/transformers/modeling_utils.py:2790: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead. |
|
warnings.warn( |
|
[INFO|modeling_utils.py:2940] 2024-01-08 11:04:53,037 >> Overriding torch_dtype=None with `torch_dtype=torch.float16` due to requirements of `bitsandbytes` to enable model loading in 8-bit or 4-bit. Pass your own torch_dtype to specify the dtype of the remaining non-linear layers or pass torch_dtype=torch.float16 to remove this warning. |
|
[INFO|modeling_utils.py:2950] 2024-01-08 11:04:53,037 >> The device_map was not initialized. Setting device_map to {'':torch.cuda.current_device()}. If you want to use the model for inference, please set device_map ='auto' |
|
[INFO|modeling_utils.py:3376] 2024-01-08 11:04:53,143 >> loading weights file model.safetensors from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/model.safetensors |
|
[INFO|modeling_utils.py:1366] 2024-01-08 11:04:53,159 >> Instantiating WhisperForConditionalGeneration model under default dtype torch.float16. |
|
[INFO|configuration_utils.py:826] 2024-01-08 11:04:53,161 >> Generate config GenerationConfig { |
|
"begin_suppress_tokens": [ |
|
220, |
|
50257 |
|
], |
|
"bos_token_id": 50257, |
|
"decoder_start_token_id": 50258, |
|
"eos_token_id": 50257, |
|
"forced_decoder_ids": [ |
|
[ |
|
1, |
|
50259 |
|
], |
|
[ |
|
2, |
|
50359 |
|
], |
|
[ |
|
3, |
|
50363 |
|
] |
|
], |
|
"max_length": 448, |
|
"pad_token_id": 50257 |
|
} |
|
|
|
[INFO|modeling_utils.py:3513] 2024-01-08 11:04:53,182 >> Detected 8-bit loading: activating 8-bit loading for this model |
|
[INFO|modeling_utils.py:4227] 2024-01-08 11:04:54,125 >> All model checkpoint weights were used when initializing WhisperForConditionalGeneration. |
|
|
|
[INFO|modeling_utils.py:4235] 2024-01-08 11:04:54,126 >> All the weights of WhisperForConditionalGeneration were initialized from the model checkpoint at openai/whisper-small. |
|
If your task is similar to the task the model of the checkpoint was trained on, you can already use WhisperForConditionalGeneration for predictions without further training. |
|
[INFO|configuration_utils.py:781] 2024-01-08 11:04:54,237 >> loading configuration file generation_config.json from cache at /esat/audioslave/qmeeus/.cache/huggingface/hub/models--openai--whisper-small/snapshots/e34e8ae444c29815eca53e11383ea13b2e362eb0/generation_config.json |
|
[INFO|configuration_utils.py:826] 2024-01-08 11:04:54,237 >> Generate config GenerationConfig { |
|
"alignment_heads": [ |
|
[ |
|
5, |
|
3 |
|
], |
|
[ |
|
5, |
|
9 |
|
], |
|
[ |
|
8, |
|
0 |
|
], |
|
[ |
|
8, |
|
4 |
|
], |
|
[ |
|
8, |
|
7 |
|
], |
|
[ |
|
8, |
|
8 |
|
], |
|
[ |
|
9, |
|
0 |
|
], |
|
[ |
|
9, |
|
7 |
|
], |
|
[ |
|
9, |
|
9 |
|
], |
|
[ |
|
10, |
|
5 |
|
] |
|
], |
|
"begin_suppress_tokens": [ |
|
220, |
|
50257 |
|
], |
|
"bos_token_id": 50257, |
|
"decoder_start_token_id": 50258, |
|
"eos_token_id": 50257, |
|
"forced_decoder_ids": [ |
|
[ |
|
1, |
|
null |
|
], |
|
[ |
|
2, |
|
50359 |
|
] |
|
], |
|
"is_multilingual": true, |
|
"lang_to_id": { |
|
"<|af|>": 50327, |
|
"<|am|>": 50334, |
|
"<|ar|>": 50272, |
|
"<|as|>": 50350, |
|
"<|az|>": 50304, |
|
"<|ba|>": 50355, |
|
"<|be|>": 50330, |
|
"<|bg|>": 50292, |
|
"<|bn|>": 50302, |
|
"<|bo|>": 50347, |
|
"<|br|>": 50309, |
|
"<|bs|>": 50315, |
|
"<|ca|>": 50270, |
|
"<|cs|>": 50283, |
|
"<|cy|>": 50297, |
|
"<|da|>": 50285, |
|
"<|de|>": 50261, |
|
"<|el|>": 50281, |
|
"<|en|>": 50259, |
|
"<|es|>": 50262, |
|
"<|et|>": 50307, |
|
"<|eu|>": 50310, |
|
"<|fa|>": 50300, |
|
"<|fi|>": 50277, |
|
"<|fo|>": 50338, |
|
"<|fr|>": 50265, |
|
"<|gl|>": 50319, |
|
"<|gu|>": 50333, |
|
"<|haw|>": 50352, |
|
"<|ha|>": 50354, |
|
"<|he|>": 50279, |
|
"<|hi|>": 50276, |
|
"<|hr|>": 50291, |
|
"<|ht|>": 50339, |
|
"<|hu|>": 50286, |
|
"<|hy|>": 50312, |
|
"<|id|>": 50275, |
|
"<|is|>": 50311, |
|
"<|it|>": 50274, |
|
"<|ja|>": 50266, |
|
"<|jw|>": 50356, |
|
"<|ka|>": 50329, |
|
"<|kk|>": 50316, |
|
"<|km|>": 50323, |
|
"<|kn|>": 50306, |
|
"<|ko|>": 50264, |
|
"<|la|>": 50294, |
|
"<|lb|>": 50345, |
|
"<|ln|>": 50353, |
|
"<|lo|>": 50336, |
|
"<|lt|>": 50293, |
|
"<|lv|>": 50301, |
|
"<|mg|>": 50349, |
|
"<|mi|>": 50295, |
|
"<|mk|>": 50308, |
|
"<|ml|>": 50296, |
|
"<|mn|>": 50314, |
|
"<|mr|>": 50320, |
|
"<|ms|>": 50282, |
|
"<|mt|>": 50343, |
|
"<|my|>": 50346, |
|
"<|ne|>": 50313, |
|
"<|nl|>": 50271, |
|
"<|nn|>": 50342, |
|
"<|no|>": 50288, |
|
"<|oc|>": 50328, |
|
"<|pa|>": 50321, |
|
"<|pl|>": 50269, |
|
"<|ps|>": 50340, |
|
"<|pt|>": 50267, |
|
"<|ro|>": 50284, |
|
"<|ru|>": 50263, |
|
"<|sa|>": 50344, |
|
"<|sd|>": 50332, |
|
"<|si|>": 50322, |
|
"<|sk|>": 50298, |
|
"<|sl|>": 50305, |
|
"<|sn|>": 50324, |
|
"<|so|>": 50326, |
|
"<|sq|>": 50317, |
|
"<|sr|>": 50303, |
|
"<|su|>": 50357, |
|
"<|sv|>": 50273, |
|
"<|sw|>": 50318, |
|
"<|ta|>": 50287, |
|
"<|te|>": 50299, |
|
"<|tg|>": 50331, |
|
"<|th|>": 50289, |
|
"<|tk|>": 50341, |
|
"<|tl|>": 50348, |
|
"<|tr|>": 50268, |
|
"<|tt|>": 50351, |
|
"<|uk|>": 50280, |
|
"<|ur|>": 50290, |
|
"<|uz|>": 50337, |
|
"<|vi|>": 50278, |
|
"<|yi|>": 50335, |
|
"<|yo|>": 50325, |
|
"<|zh|>": 50260 |
|
}, |
|
"max_initial_timestamp_index": 1, |
|
"max_length": 448, |
|
"no_timestamps_token_id": 50363, |
|
"pad_token_id": 50257, |
|
"return_timestamps": false, |
|
"suppress_tokens": [ |
|
1, |
|
2, |
|
7, |
|
8, |
|
9, |
|
10, |
|
14, |
|
25, |
|
26, |
|
27, |
|
28, |
|
29, |
|
31, |
|
58, |
|
59, |
|
60, |
|
61, |
|
62, |
|
63, |
|
90, |
|
91, |
|
92, |
|
93, |
|
359, |
|
503, |
|
522, |
|
542, |
|
873, |
|
893, |
|
902, |
|
918, |
|
922, |
|
931, |
|
1350, |
|
1853, |
|
1982, |
|
2460, |
|
2627, |
|
3246, |
|
3253, |
|
3268, |
|
3536, |
|
3846, |
|
3961, |
|
4183, |
|
4667, |
|
6585, |
|
6647, |
|
7273, |
|
9061, |
|
9383, |
|
10428, |
|
10929, |
|
11938, |
|
12033, |
|
12331, |
|
12562, |
|
13793, |
|
14157, |
|
14635, |
|
15265, |
|
15618, |
|
16553, |
|
16604, |
|
18362, |
|
18956, |
|
20075, |
|
21675, |
|
22520, |
|
26130, |
|
26161, |
|
26435, |
|
28279, |
|
29464, |
|
31650, |
|
32302, |
|
32470, |
|
36865, |
|
42863, |
|
47425, |
|
49870, |
|
50254, |
|
50258, |
|
50358, |
|
50359, |
|
50360, |
|
50361, |
|
50362 |
|
], |
|
"task_to_id": { |
|
"transcribe": 50359, |
|
"translate": 50358 |
|
} |
|
} |
|
|
|
Resize embeddings to account for the new tokens |
|
[INFO|modeling_utils.py:1839] 2024-01-08 11:04:54,310 >> You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 51885. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https: |
|
PeftModel( |
|
(base_model): LoraModel( |
|
(model): WhisperForConditionalGeneration( |
|
(model): WhisperModel( |
|
(encoder): WhisperEncoder( |
|
(conv1): Conv1d(80, 768, kernel_size=(3,), stride=(1,), padding=(1,)) |
|
(conv2): Conv1d(768, 768, kernel_size=(3,), stride=(2,), padding=(1,)) |
|
(embed_positions): Embedding(1500, 768) |
|
(layers): ModuleList( |
|
(0-11): 12 x WhisperEncoderLayer( |
|
(self_attn): WhisperAttention( |
|
(k_proj): Linear8bitLt(in_features=768, out_features=768, bias=False) |
|
(v_proj): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
(q_proj): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
(out_proj): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
) |
|
(self_attn_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) |
|
(activation_fn): GELUActivation() |
|
(fc1): Linear8bitLt(in_features=768, out_features=3072, bias=True) |
|
(fc2): Linear8bitLt(in_features=3072, out_features=768, bias=True) |
|
(final_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) |
|
) |
|
) |
|
(layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) |
|
) |
|
(decoder): WhisperDecoder( |
|
(embed_tokens): ModulesToSaveWrapper( |
|
(original_module): Embedding(51885, 768) |
|
(modules_to_save): ModuleDict( |
|
(default): Embedding(51885, 768) |
|
) |
|
) |
|
(embed_positions): WhisperPositionalEmbedding(448, 768) |
|
(layers): ModuleList( |
|
(0-11): 12 x WhisperDecoderLayer( |
|
(self_attn): WhisperAttention( |
|
(k_proj): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=768, bias=False) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
(v_proj): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
(q_proj): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
(out_proj): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
) |
|
(activation_fn): GELUActivation() |
|
(self_attn_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) |
|
(encoder_attn): WhisperAttention( |
|
(k_proj): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=768, bias=False) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
(v_proj): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
(q_proj): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
(out_proj): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=768, bias=True) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
) |
|
(encoder_attn_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) |
|
(fc1): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=768, out_features=3072, bias=True) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=768, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=3072, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
(fc2): lora.Linear8bitLt( |
|
(base_layer): Linear8bitLt(in_features=3072, out_features=768, bias=True) |
|
(lora_dropout): ModuleDict( |
|
(default): Dropout(p=0.1, inplace=False) |
|
) |
|
(lora_A): ModuleDict( |
|
(default): Linear(in_features=3072, out_features=8, bias=False) |
|
) |
|
(lora_B): ModuleDict( |
|
(default): Linear(in_features=8, out_features=768, bias=False) |
|
) |
|
(lora_embedding_A): ParameterDict() |
|
(lora_embedding_B): ParameterDict() |
|
) |
|
(final_layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) |
|
) |
|
) |
|
(layer_norm): LayerNorm((768,), eps=1e-05, elementwise_affine=True) |
|
) |
|
) |
|
(proj_out): Linear(in_features=768, out_features=51885, bias=False) |
|
) |
|
) |
|
) |
|
trainable params: 41,764,608 || all params: 283,514,880 || trainable%: 14.731010943764222 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 11:04:55,231 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/preprocessor_config.json |
|
[INFO|tokenization_utils_base.py:2432] 2024-01-08 11:04:55,262 >> tokenizer config file saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tokenizer_config.json |
|
[INFO|tokenization_utils_base.py:2441] 2024-01-08 11:04:55,263 >> Special tokens file saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/special_tokens_map.json |
|
[INFO|configuration_utils.py:483] 2024-01-08 11:04:55,319 >> Configuration saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/config.json |
|
[INFO|image_processing_utils.py:373] 2024-01-08 11:04:55,320 >> loading configuration file /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/preprocessor_config.json |
|
[INFO|feature_extraction_utils.py:535] 2024-01-08 11:04:55,320 >> loading configuration file /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/preprocessor_config.json |
|
[INFO|feature_extraction_utils.py:579] 2024-01-08 11:04:55,320 >> Feature extractor WhisperFeatureExtractor { |
|
"chunk_length": 30, |
|
"feature_extractor_type": "WhisperFeatureExtractor", |
|
"feature_size": 80, |
|
"hop_length": 160, |
|
"n_fft": 400, |
|
"n_samples": 480000, |
|
"nb_max_frames": 3000, |
|
"padding_side": "right", |
|
"padding_value": 0.0, |
|
"processor_class": "WhisperProcessor", |
|
"return_attention_mask": false, |
|
"sampling_rate": 16000 |
|
} |
|
|
|
[INFO|tokenization_utils_base.py:2024] 2024-01-08 11:04:55,323 >> loading file vocab.json |
|
[INFO|tokenization_utils_base.py:2024] 2024-01-08 11:04:55,323 >> loading file tokenizer.json |
|
[INFO|tokenization_utils_base.py:2024] 2024-01-08 11:04:55,323 >> loading file merges.txt |
|
[INFO|tokenization_utils_base.py:2024] 2024-01-08 11:04:55,323 >> loading file normalizer.json |
|
[INFO|tokenization_utils_base.py:2024] 2024-01-08 11:04:55,323 >> loading file added_tokens.json |
|
[INFO|tokenization_utils_base.py:2024] 2024-01-08 11:04:55,323 >> loading file special_tokens_map.json |
|
[INFO|tokenization_utils_base.py:2024] 2024-01-08 11:04:55,323 >> loading file tokenizer_config.json |
|
[WARNING|tokenization_utils_base.py:2140] 2024-01-08 11:04:55,325 >> The tokenizer class you load from this checkpoint is not the same type as the class this function is called from. It may result in unexpected tokenization. |
|
The tokenizer class you load from this checkpoint is 'NERTokenizerEndToEnd'. |
|
The class this function is called from is 'WhisperTokenizer'. |
|
[WARNING|logging.py:314] 2024-01-08 11:04:55,408 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. |
|
[INFO|trainer.py:522] 2024-01-08 11:04:55,409 >> max_steps is given, it will override any value given in num_train_epochs |
|
[INFO|trainer.py:571] 2024-01-08 11:04:55,409 >> Using auto half precision backend |
|
[INFO|trainer.py:718] 2024-01-08 11:04:55,537 >> The following columns in the training set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:1712] 2024-01-08 11:04:55,580 >> ***** Running training ***** |
|
[INFO|trainer.py:1713] 2024-01-08 11:04:55,580 >> Num examples = 71,615 |
|
[INFO|trainer.py:1714] 2024-01-08 11:04:55,580 >> Num Epochs = 9 |
|
[INFO|trainer.py:1715] 2024-01-08 11:04:55,580 >> Instantaneous batch size per device = 32 |
|
[INFO|trainer.py:1718] 2024-01-08 11:04:55,580 >> Total train batch size (w. parallel, distributed & accumulation) = 128 |
|
[INFO|trainer.py:1719] 2024-01-08 11:04:55,580 >> Gradient Accumulation steps = 4 |
|
[INFO|trainer.py:1720] 2024-01-08 11:04:55,580 >> Total optimization steps = 5,000 |
|
[INFO|trainer.py:1721] 2024-01-08 11:04:55,582 >> Number of trainable parameters = 41,764,608 |
|
[INFO|integration_utils.py:722] 2024-01-08 11:04:55,585 >> Automatic Weights & Biases logging enabled, to disable set os.environ["WANDB_DISABLED"] = "true" |
|
wandb: Currently logged in as: qmeeus. Use `wandb login --relogin` to force relogin |
|
wandb: wandb version 0.16.1 is available! To upgrade, please run: |
|
wandb: $ pip install wandb --upgrade |
|
wandb: Tracking run with wandb version 0.15.12 |
|
wandb: Run data is saved locally in /esat/audioslave/qmeeus/repos/peft/examples/whisper_slu/wandb/run-20240108_110457-pwws6pgg |
|
wandb: Run `wandb offline` to turn off syncing. |
|
wandb: Syncing run happy-shape-146 |
|
wandb: ⭐️ View project at https://wandb.ai/qmeeus/WhisperForSpokenNER |
|
wandb: 🚀 View run at https://wandb.ai/qmeeus/WhisperForSpokenNER/runs/pwws6pgg |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[WARNING|logging.py:329] 2024-01-08 11:04:59,883 >> `use_cache = True` is incompatible with gradient checkpointing. Setting `use_cache = False`... |
|
[INFO|trainer.py:718] 2024-01-08 11:52:18,771 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 11:52:18,774 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 11:52:18,774 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 11:52:18,774 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 12:14:26,632 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-200 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 12:14:27,736 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-200/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 13:01:56,790 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 13:01:56,792 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 13:01:56,792 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 13:01:56,792 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 13:22:35,296 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-400 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 13:22:35,942 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-400/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 14:09:59,237 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 14:09:59,239 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 14:09:59,240 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 14:09:59,240 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 14:23:49,977 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-600 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 14:23:50,563 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-600/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 15:11:21,439 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 15:11:21,441 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 15:11:21,442 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 15:11:21,442 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 15:23:24,665 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-800 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 15:23:25,269 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-800/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 16:10:48,253 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 16:10:48,255 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 16:10:48,255 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 16:10:48,255 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 16:21:24,766 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1000 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 16:21:25,353 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1000/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 17:08:44,996 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 17:08:44,998 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 17:08:44,998 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 17:08:44,998 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 17:19:01,716 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1200 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 17:19:02,300 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1200/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 18:06:20,117 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 18:06:20,119 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 18:06:20,119 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 18:06:20,119 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 18:16:34,445 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1400 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 18:16:35,039 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1400/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 19:03:59,252 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 19:03:59,254 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 19:03:59,254 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 19:03:59,254 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 19:14:16,766 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1600 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 19:14:17,339 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1600/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 20:01:40,919 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 20:01:40,921 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 20:01:40,921 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 20:01:40,921 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 20:12:01,672 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1800 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 20:12:02,230 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-1800/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 20:59:23,021 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 20:59:23,023 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 20:59:23,023 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 20:59:23,023 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 21:09:48,958 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2000 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 21:09:49,754 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2000/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 21:57:16,746 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 21:57:16,748 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 21:57:16,748 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 21:57:16,748 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 22:07:51,020 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2200 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 22:07:51,610 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2200/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 22:55:16,722 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 22:55:16,724 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 22:55:16,725 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 22:55:16,725 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-08 23:05:32,086 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2400 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-08 23:05:32,656 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2400/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-08 23:52:52,778 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-08 23:52:52,780 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-08 23:52:52,780 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-08 23:52:52,780 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 00:03:21,033 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2600 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 00:03:21,665 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2600/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 00:50:41,209 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 00:50:41,212 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 00:50:41,212 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 00:50:41,212 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 01:00:54,242 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2800 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 01:00:54,841 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-2800/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 01:48:21,353 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 01:48:21,355 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 01:48:21,355 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 01:48:21,355 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 01:58:50,224 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3000 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 01:58:50,832 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3000/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 02:46:15,949 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 02:46:15,952 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 02:46:15,953 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 02:46:15,953 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 02:56:43,336 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3200 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 02:56:43,939 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3200/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 03:43:53,333 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 03:43:53,335 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 03:43:53,335 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 03:43:53,335 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 03:53:50,834 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3400 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 03:53:51,418 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3400/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 04:40:53,943 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 04:40:53,945 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 04:40:53,945 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 04:40:53,945 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 04:50:57,328 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3600 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 04:50:57,938 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3600/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 05:38:00,576 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 05:38:00,578 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 05:38:00,578 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 05:38:00,578 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 05:47:48,535 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3800 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 05:47:49,128 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-3800/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 06:34:48,377 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 06:34:48,379 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 06:34:48,379 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 06:34:48,379 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 06:44:39,057 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4000 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 06:44:39,679 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4000/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 07:31:41,732 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 07:31:41,734 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 07:31:41,735 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 07:31:41,735 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 07:41:28,691 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4200 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 07:41:29,297 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4200/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 08:28:32,720 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 08:28:32,722 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 08:28:32,723 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 08:28:32,723 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 08:38:23,039 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4400 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 08:38:23,606 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4400/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 09:25:25,423 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 09:25:25,425 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 09:25:25,426 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 09:25:25,426 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 09:35:16,052 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4600 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 09:35:16,635 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4600/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 10:22:49,334 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 10:22:49,336 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 10:22:49,336 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 10:22:49,336 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 10:32:42,825 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4800 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 10:32:43,402 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-4800/preprocessor_config.json |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants. |
|
warnings.warn( |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
[INFO|trainer.py:718] 2024-01-09 11:20:29,411 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 11:20:29,413 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 11:20:29,413 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 11:20:29,413 >> Batch size = 16 |
|
[INFO|trainer.py:2895] 2024-01-09 11:30:34,161 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-5000 |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 11:30:34,756 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/tmp-checkpoint-5000/preprocessor_config.json |
|
[INFO|trainer.py:1953] 2024-01-09 11:30:35,705 >> |
|
|
|
Training completed. Do not forget to share your model on huggingface.co/models =) |
|
|
|
|
|
[INFO|trainer.py:2895] 2024-01-09 11:30:35,710 >> Saving model checkpoint to /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora |
|
[INFO|feature_extraction_utils.py:425] 2024-01-09 11:30:36,330 >> Feature extractor saved in /esat/audioslave/qmeeus/exp/whisper_slu/e2e/whisper-small-spoken-ner-lora/preprocessor_config.json |
|
[INFO|trainer.py:718] 2024-01-09 11:30:36,338 >> The following columns in the evaluation set don't have a corresponding argument in `PeftModel.forward` and have been ignored: input_length. If input_length are not expected by `PeftModel.forward`, you can safely ignore this message. |
|
[INFO|trainer.py:3199] 2024-01-09 11:30:36,341 >> ***** Running Evaluation ***** |
|
[INFO|trainer.py:3201] 2024-01-09 11:30:36,341 >> Num examples = 1000 |
|
[INFO|trainer.py:3204] 2024-01-09 11:30:36,341 >> Batch size = 16 |
|
/users/spraak/qmeeus/micromamba/envs/torch-cu121/lib/python3.10/site-packages/bitsandbytes/autograd/_functions.py:322: UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization |
|
warnings.warn(f"MatMul8bitLt: inputs will be cast from {A.dtype} to float16 during quantization") |
|
wandb: Waiting for W&B process to finish... (success). |
|
wandb: |
|
wandb: Run history: |
|
wandb: eval/loss █▄▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ |
|
wandb: eval/runtime █▇▃▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ |
|
wandb: eval/samples_per_second ▁▁▄▆▇▇▇▇▇▇▇▇▇▇▇▇██████████ |
|
wandb: eval/steps_per_second ▁▁▄▆▇▇█▇▇▇▇▇▇█▇▇██████████ |
|
wandb: eval/wer █▃▁▁▁▁▁▁▁▁▂▁▁▁▂▂▂▂▂▂▂▂▂▂▂▂ |
|
wandb: train/epoch ▁▁▁▁▂▂▂▂▂▃▃▃▃▃▄▄▄▄▄▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇████ |
|
wandb: train/global_step ▁▁▁▁▂▂▂▂▂▃▃▃▃▃▄▄▄▄▄▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇████ |
|
wandb: train/learning_rate ▂▄▅▇██████▇▇▇▇▇▆▆▆▆▅▅▅▄▄▄▃▃▃▃▂▂▂▂▂▁▁▁▁▁▁ |
|
wandb: train/loss █▆▄▃▂▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ |
|
wandb: train/total_flos ▁ |
|
wandb: train/train_loss ▁ |
|
wandb: train/train_runtime ▁ |
|
wandb: train/train_samples_per_second ▁ |
|
wandb: train/train_steps_per_second ▁ |
|
wandb: |
|
wandb: Run summary: |
|
wandb: eval/loss 0.33808 |
|
wandb: eval/runtime 601.5847 |
|
wandb: eval/samples_per_second 1.662 |
|
wandb: eval/steps_per_second 0.105 |
|
wandb: eval/wer 0.38886 |
|
wandb: train/epoch 8.94 |
|
wandb: train/global_step 5000 |
|
wandb: train/learning_rate 0.0 |
|
wandb: train/loss 0.3079 |
|
wandb: train/total_flos 1.8645896628338688e+20 |
|
wandb: train/train_loss 0.49281 |
|
wandb: train/train_runtime 87940.1228 |
|
wandb: train/train_samples_per_second 7.278 |
|
wandb: train/train_steps_per_second 0.057 |
|
wandb: |
|
wandb: 🚀 View run happy-shape-146 at: https://wandb.ai/qmeeus/WhisperForSpokenNER/runs/pwws6pgg |
|
wandb: ️⚡ View job at https://wandb.ai/qmeeus/WhisperForSpokenNER/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjEyNzk2NDY0Mw==/version_details/v2 |
|
wandb: Synced 5 W&B file(s), 0 media file(s), 2 artifact file(s) and 0 other file(s) |
|
wandb: Find logs at: ./wandb/run-20240108_110457-pwws6pgg/logs |
|
|
