Training in progress, epoch 1

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:37bcf89ac947900bb29c14fdab541d78ce1a97f36c429b65c1af82b74033be8a
+oid sha256:057364a5a2d45f4470bad361fa73b0f0e8bef6e789cbf4a074256afa49b207f8
 size 265503852

runs/Feb21_17-17-29_047c4dfc8464/events.out.tfevents.1708535850.047c4dfc8464.34.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52e1c7809b33f4fb14a3690762cd54702d540d4af203e23e04330f5b0e9e081a
+size 4789

runs/Feb21_17-18-42_047c4dfc8464/events.out.tfevents.1708535923.047c4dfc8464.34.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b06cea92472ffcf3e1d6731733125bf838d1ccafcede2d245a22b4657228e57
+size 4789

runs/Feb21_17-18-46_047c4dfc8464/events.out.tfevents.1708535927.047c4dfc8464.34.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a2183dd4c7f31b0c466146bd15b8a885fe85092fa562b4cda030e19b1496189
+size 4789

runs/Feb21_17-19-35_047c4dfc8464/events.out.tfevents.1708535975.047c4dfc8464.34.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:681b7e6024028becac174902afa64edd780318db8231704da9e743b19c46a5a2
+size 4789

runs/Feb21_17-19-37_047c4dfc8464/events.out.tfevents.1708535978.047c4dfc8464.34.4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1e04c513c0c216c53e4097ce71a242321c123ee75e09549f2d394c8c6595713
+size 4789

runs/Feb21_17-21-41_047c4dfc8464/events.out.tfevents.1708536102.047c4dfc8464.34.5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85c1a87abcf0f629666ad6d3e91e9d181585e31d23b7e6e6ed64c163b63d9c95
+size 4789

runs/Feb21_17-21-43_047c4dfc8464/events.out.tfevents.1708536104.047c4dfc8464.34.6

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:adcca51f495d09ad782cf1fbe5fb50ccb060ce8ef45f34ca4086abdf15b70cc9
+size 4789

runs/Feb21_17-23-09_047c4dfc8464/events.out.tfevents.1708536194.047c4dfc8464.34.7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82ad2c35a0be8aeccee2f3bdbf4f5678112774fe86a54f402bdf88b974eeaea2
+size 4789

runs/Feb21_17-23-16_047c4dfc8464/events.out.tfevents.1708536201.047c4dfc8464.34.8

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:013e43de58bfcc56530328b02899086ceee196a39830ea26357a75f71cab2ab7
+size 5297

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3234aa79348c5613964a96604cadf049e9ff147393933a104512083d38b43f7d
+oid sha256:4e117c2d84287a42d11fed7e4b4ca80d92527b15113fec7231f3b905c0d5ff88
 size 4728

wandb/debug.log

@@ -1,29 +1,90 @@
-2024-02-19 19:09:40,367 INFO    MainThread:26 [wandb_setup.py:_flush():76] Current SDK version is 0.16.2
-2024-02-19 19:09:40,367 INFO    MainThread:26 [wandb_setup.py:_flush():76] Configure stats pid to 26
-2024-02-19 19:09:40,367 INFO    MainThread:26 [wandb_setup.py:_flush():76] Loading settings from /root/.config/wandb/settings
-2024-02-19 19:09:40,367 INFO    MainThread:26 [wandb_setup.py:_flush():76] Loading settings from /kaggle/working/wandb/settings
-2024-02-19 19:09:40,367 INFO    MainThread:26 [wandb_setup.py:_flush():76] Loading settings from environment variables: {}
-2024-02-19 19:09:40,367 INFO    MainThread:26 [wandb_setup.py:_flush():76] Inferring run settings from compute environment: {'program': '<python with no main file>'}
-2024-02-19 19:09:40,367 INFO    MainThread:26 [wandb_setup.py:_flush():76] Applying login settings: {'api_key': '***REDACTED***'}
-2024-02-19 19:09:40,368 INFO    MainThread:26 [wandb_setup.py:_flush():76] Applying login settings: {'api_key': '***REDACTED***'}
-2024-02-19 19:09:40,368 INFO    MainThread:26 [wandb_init.py:_log_setup():526] Logging user logs to /kaggle/working/wandb/run-20240219_190940-ztflz27u/logs/debug.log
-2024-02-19 19:09:40,368 INFO    MainThread:26 [wandb_init.py:_log_setup():527] Logging internal logs to /kaggle/working/wandb/run-20240219_190940-ztflz27u/logs/debug-internal.log
-2024-02-19 19:09:40,368 INFO    MainThread:26 [wandb_init.py:init():566] calling init triggers
-2024-02-19 19:09:40,368 INFO    MainThread:26 [wandb_init.py:init():573] wandb.init called with sweep_config: {}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Current SDK version is 0.16.2
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Configure stats pid to 34
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Loading settings from /root/.config/wandb/settings
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Loading settings from /kaggle/working/wandb/settings
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Loading settings from environment variables: {}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Inferring run settings from compute environment: {'program': '<python with no main file>'}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Applying login settings: {'api_key': '***REDACTED***'}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Applying login settings: {'api_key': '***REDACTED***'}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_init.py:_log_setup():526] Logging user logs to /kaggle/working/wandb/run-20240221_171730-9wbsk0df/logs/debug.log
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_init.py:_log_setup():527] Logging internal logs to /kaggle/working/wandb/run-20240221_171730-9wbsk0df/logs/debug-internal.log
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:_jupyter_setup():472] configuring jupyter hooks <wandb.sdk.wandb_init._WandbInit object at 0x79ab9071fd60>
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:init():566] calling init triggers
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:init():573] wandb.init called with sweep_config: {}
 config: {}
-2024-02-19 19:09:40,368 INFO    MainThread:26 [wandb_init.py:init():616] starting backend
-2024-02-19 19:09:40,368 INFO    MainThread:26 [wandb_init.py:init():620] setting up manager
-2024-02-19 19:09:40,369 INFO    MainThread:26 [backend.py:_multiprocessing_setup():105] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
-2024-02-19 19:09:40,372 INFO    MainThread:26 [wandb_init.py:init():628] backend started and connected
-2024-02-19 19:09:40,375 INFO    MainThread:26 [wandb_init.py:init():720] updated telemetry
-2024-02-19 19:09:40,378 INFO    MainThread:26 [wandb_init.py:init():753] communicating run to backend with 90.0 second timeout
-2024-02-19 19:09:40,754 INFO    MainThread:26 [wandb_run.py:_on_init():2254] communicating current version
-2024-02-19 19:09:40,810 INFO    MainThread:26 [wandb_run.py:_on_init():2263] got version response upgrade_message: "wandb version 0.16.3 is available!  To upgrade, please run:\n $ pip install wandb --upgrade"
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:init():616] starting backend
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:init():620] setting up manager
+2024-02-21 17:17:30,949 INFO    MainThread:34 [backend.py:_multiprocessing_setup():105] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2024-02-21 17:17:30,951 INFO    MainThread:34 [wandb_init.py:init():628] backend started and connected
+2024-02-21 17:17:30,963 INFO    MainThread:34 [wandb_run.py:_label_probe_notebook():1294] probe notebook
+2024-02-21 17:17:31,264 INFO    MainThread:34 [wandb_init.py:init():720] updated telemetry
+2024-02-21 17:17:31,267 INFO    MainThread:34 [wandb_init.py:init():753] communicating run to backend with 90.0 second timeout
+2024-02-21 17:17:31,457 INFO    MainThread:34 [wandb_run.py:_on_init():2254] communicating current version
+2024-02-21 17:17:31,540 INFO    MainThread:34 [wandb_run.py:_on_init():2263] got version response upgrade_message: "wandb version 0.16.3 is available!  To upgrade, please run:\n $ pip install wandb --upgrade"
 
-2024-02-19 19:09:40,810 INFO    MainThread:26 [wandb_init.py:init():804] starting run threads in backend
-2024-02-19 19:10:10,843 INFO    MainThread:26 [wandb_run.py:_console_start():2233] atexit reg
-2024-02-19 19:10:10,844 INFO    MainThread:26 [wandb_run.py:_redirect():2088] redirect: wrap_raw
-2024-02-19 19:10:10,844 INFO    MainThread:26 [wandb_run.py:_redirect():2153] Wrapping output streams.
-2024-02-19 19:10:10,844 INFO    MainThread:26 [wandb_run.py:_redirect():2178] Redirects installed.
-2024-02-19 19:10:10,847 INFO    MainThread:26 [wandb_init.py:init():847] run started, returning control to user process
-2024-02-19 19:10:10,854 INFO    MainThread:26 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb19_19-09-39_ed7fe84f9951', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:17:31,540 INFO    MainThread:34 [wandb_init.py:init():804] starting run threads in backend
+2024-02-21 17:18:01,575 INFO    MainThread:34 [wandb_run.py:_console_start():2233] atexit reg
+2024-02-21 17:18:01,576 INFO    MainThread:34 [wandb_run.py:_redirect():2088] redirect: wrap_raw
+2024-02-21 17:18:01,576 INFO    MainThread:34 [wandb_run.py:_redirect():2153] Wrapping output streams.
+2024-02-21 17:18:01,576 INFO    MainThread:34 [wandb_run.py:_redirect():2178] Redirects installed.
+2024-02-21 17:18:01,577 INFO    MainThread:34 [wandb_init.py:init():847] run started, returning control to user process
+2024-02-21 17:18:01,585 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-17-29_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:18:06,099 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:06,099 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:18:37,525 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:18:37,529 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:37,529 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:18:40,523 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:18:40,528 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:40,528 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:18:42,574 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:18:43,613 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-18-42_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:18:44,545 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:44,545 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:18:46,421 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:18:47,676 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-18-46_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:18:48,053 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:48,054 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:17,565 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:17,567 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:17,567 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:22,718 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:22,723 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:22,724 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:29,543 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:29,547 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:29,547 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:33,814 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:33,815 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:33,815 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:35,014 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:35,936 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-19-35_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:19:36,340 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:36,340 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:37,545 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:38,762 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-19-37_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:19:39,145 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:39,145 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:21:19,590 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:21:19,595 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:21:19,595 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:21:35,970 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:21:35,973 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:21:35,974 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:21:41,190 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:21:42,149 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-21-41_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:21:42,563 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:21:42,563 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:21:43,654 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:21:44,868 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-21-43_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:21:47,045 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:21:47,046 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:23:07,591 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:23:07,593 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:23:07,594 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:23:09,895 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:23:14,718 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-23-09_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:23:15,139 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:23:15,139 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:23:16,280 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:23:21,586 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-23-16_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}

wandb/run-20240221_171730-9wbsk0df/files/conda-environment.yaml

@@ -0,0 +1,1072 @@
+name: base
+channels:
+  - pytorch
+  - file:///tmp/conda
+  - rapidsai
+  - nvidia
+  - conda-forge
+  - defaults
+dependencies:
+  - _libgcc_mutex=0.1=conda_forge
+  - _openmp_mutex=4.5=2_gnu
+  - aiohttp=3.9.1=py310h2372a71_0
+  - aiosignal=1.3.1=pyhd8ed1ab_0
+  - annotated-types=0.6.0=pyhd8ed1ab_0
+  - anyio=4.2.0=pyhd8ed1ab_0
+  - archspec=0.2.2=pyhd8ed1ab_0
+  - argon2-cffi=23.1.0=pyhd8ed1ab_0
+  - argon2-cffi-bindings=21.2.0=py310h2372a71_4
+  - arrow=1.3.0=pyhd8ed1ab_0
+  - arrow-cpp=11.0.0=ha770c72_9_cpu
+  - asttokens=2.4.1=pyhd8ed1ab_0
+  - async-timeout=4.0.3=pyhd8ed1ab_0
+  - atk-1.0=2.38.0=hd4edc92_1
+  - attrs=23.2.0=pyh71513ae_0
+  - aws-c-auth=0.6.26=h987a71b_2
+  - aws-c-cal=0.5.21=h48707d8_2
+  - aws-c-common=0.8.14=h0b41bf4_0
+  - aws-c-compression=0.2.16=h03acc5a_5
+  - aws-c-event-stream=0.2.20=h00877a2_4
+  - aws-c-http=0.7.6=hf342b9f_0
+  - aws-c-io=0.13.19=h5b20300_3
+  - aws-c-mqtt=0.8.6=hc4349f7_12
+  - aws-c-s3=0.2.7=h909e904_1
+  - aws-c-sdkutils=0.1.9=h03acc5a_0
+  - aws-checksums=0.1.14=h03acc5a_5
+  - aws-crt-cpp=0.19.8=hf7fbfca_12
+  - aws-sdk-cpp=1.10.57=h17c43bd_8
+  - beautifulsoup4=4.12.2=pyha770c72_0
+  - bleach=6.1.0=pyhd8ed1ab_0
+  - bokeh=3.3.3=pyhd8ed1ab_0
+  - boltons=23.1.1=pyhd8ed1ab_0
+  - brotli=1.0.9=h166bdaf_9
+  - brotli-bin=1.0.9=h166bdaf_9
+  - brotli-python=1.0.9=py310hd8f1fbe_9
+  - brotlipy=0.7.0=py310h7f8727e_1002
+  - bzip2=1.0.8=h7b6447c_0
+  - c-ares=1.25.0=hd590300_0
+  - ca-certificates=2023.11.17=hbcca054_0
+  - cached-property=1.5.2=hd8ed1ab_1
+  - cached_property=1.5.2=pyha770c72_1
+  - cairo=1.18.0=h3faef2a_0
+  - cartopy=0.22.0=py310hcc13569_1
+  - catalogue=2.0.10=py310hff52083_0
+  - certifi=2023.11.17=pyhd8ed1ab_0
+  - cffi=1.16.0=py310h2fee648_0
+  - charset-normalizer=3.3.2=pyhd8ed1ab_0
+  - click=8.1.7=unix_pyh707e725_0
+  - cloudpathlib=0.16.0=pyhd8ed1ab_0
+  - colorama=0.4.6=pyhd8ed1ab_0
+  - comm=0.2.1=pyhd8ed1ab_0
+  - conda=23.7.4=py310hff52083_0
+  - conda-libmamba-solver=23.7.0=pyhd8ed1ab_0
+  - conda-package-handling=2.2.0=pyh38be061_0
+  - conda-package-streaming=0.9.0=pyhd8ed1ab_0
+  - confection=0.1.4=py310h17c5347_0
+  - contourpy=1.2.0=py310hd41b1e2_0
+  - cryptography=41.0.7=py310hb8475ec_1
+  - cuda-cccl=12.3.101=0
+  - cuda-cudart=12.3.101=0
+  - cuda-cudart-dev=12.3.101=0
+  - cuda-nvcc-dev_linux-64=12.1.105=ha770c72_0
+  - cuda-nvcc-impl=12.1.105=hd3aeb46_0
+  - cuda-nvcc-tools=12.1.105=hd3aeb46_0
+  - cuda-nvrtc=12.1.105=hd3aeb46_0
+  - cuda-profiler-api=12.3.101=0
+  - cuda-python=12.3.0=py310h7ca8cf1_0
+  - cuda-version=12.1=h1d6eff3_2
+  - cudf=23.08.00=cuda12_py310_230809_g8150d38e08_0
+  - cuml=23.08.00=cuda12_py310_230809_gd7162cdea_0
+  - cupy=13.0.0=py310h7aad9d2_0
+  - cupy-core=13.0.0=py310had4011e_0
+  - curl=8.5.0=hca28451_0
+  - cycler=0.12.1=pyhd8ed1ab_0
+  - cymem=2.0.8=py310hc6cd4ac_1
+  - cython-blis=0.7.10=py310h1f7b6fc_2
+  - cytoolz=0.12.2=py310h2372a71_1
+  - dask-cuda=23.08.00=py310_230809_gefbd6ca_0
+  - dask-cudf=23.08.00=cuda12_py310_230809_g8150d38e08_0
+  - debugpy=1.8.0=py310hc6cd4ac_1
+  - decorator=5.1.1=pyhd8ed1ab_0
+  - defusedxml=0.7.1=pyhd8ed1ab_0
+  - distro=1.9.0=pyhd8ed1ab_0
+  - dlenv-tf-2-15-gpu=1.0.20240111=py310ha20f8e0_0
+  - dlpack=0.5=h9c3ff4c_0
+  - entrypoints=0.4=pyhd8ed1ab_0
+  - exceptiongroup=1.2.0=pyhd8ed1ab_2
+  - executing=2.0.1=pyhd8ed1ab_0
+  - expat=2.5.0=hcb278e6_1
+  - fastrlock=0.8.2=py310hc6cd4ac_2
+  - fftw=3.3.10=nompi_hc118613_108
+  - fmt=9.1.0=h924138e_0
+  - font-ttf-dejavu-sans-mono=2.37=hab24e00_0
+  - font-ttf-inconsolata=3.000=h77eed37_0
+  - font-ttf-source-code-pro=2.038=h77eed37_0
+  - font-ttf-ubuntu=0.83=h77eed37_1
+  - fontconfig=2.14.2=h14ed4e7_0
+  - fonts-conda-ecosystem=1=0
+  - fonts-conda-forge=1=0
+  - fqdn=1.5.1=pyhd8ed1ab_0
+  - freetype=2.12.1=h267a509_2
+  - fribidi=1.0.10=h36c2ea0_0
+  - frozenlist=1.4.1=py310h2372a71_0
+  - fsspec=2023.12.2=pyhca7485f_0
+  - gdk-pixbuf=2.42.10=h829c605_4
+  - geos=3.11.1=h27087fc_0
+  - gettext=0.21.1=h27087fc_0
+  - gflags=2.2.2=he1b5a44_1004
+  - ghostscript=10.02.1=h59595ed_0
+  - giflib=5.2.1=h0b41bf4_3
+  - glog=0.6.0=h6f12383_0
+  - gmock=1.14.0=ha770c72_1
+  - gmp=6.3.0=h59595ed_0
+  - google-api-core-grpc=2.11.1=hd8ed1ab_0
+  - google-auth=2.26.1=pyhca7485f_0
+  - google-cloud-core=2.4.1=pyhd8ed1ab_0
+  - google-cloud-datastore=2.19.0=pyhd8ed1ab_0
+  - googleapis-common-protos=1.62.0=pyhd8ed1ab_0
+  - graphite2=1.3.13=h58526e2_1001
+  - graphviz=9.0.0=h78e8752_1
+  - grpc-cpp=1.51.1=h27aab58_3
+  - gtest=1.14.0=h00ab1b0_1
+  - gtk2=2.24.33=h7f000aa_3
+  - gts=0.7.6=h977cf35_4
+  - harfbuzz=8.3.0=h3d44ed6_0
+  - icu=73.2=h59595ed_0
+  - idna=3.6=pyhd8ed1ab_0
+  - imagemagick=7.1.1_26=pl5321h3fa1221_0
+  - importlib_metadata=7.0.1=hd8ed1ab_0
+  - importlib_resources=6.1.1=pyhd8ed1ab_0
+  - intel-openmp=2023.1.0=hdb19cb5_46306
+  - ipykernel=6.28.0=pyhd33586a_0
+  - ipython=8.20.0=pyh707e725_0
+  - ipython_genutils=0.2.0=py_1
+  - isoduration=20.11.0=pyhd8ed1ab_0
+  - jbig=2.1=h7f98852_2003
+  - jedi=0.19.1=pyhd8ed1ab_0
+  - jinja2=3.1.2=pyhd8ed1ab_1
+  - joblib=1.3.2=pyhd8ed1ab_0
+  - jsonpatch=1.33=pyhd8ed1ab_0
+  - jsonpointer=2.4=py310hff52083_3
+  - jsonschema=4.20.0=pyhd8ed1ab_0
+  - jsonschema-specifications=2023.12.1=pyhd8ed1ab_0
+  - jsonschema-with-format-nongpl=4.20.0=pyhd8ed1ab_0
+  - jupyter_client=8.6.0=pyhd8ed1ab_0
+  - jupyter_core=5.7.1=py310hff52083_0
+  - jupyter_events=0.9.0=pyhd8ed1ab_0
+  - jupyter_server_terminals=0.5.1=pyhd8ed1ab_0
+  - jupyterlab_pygments=0.3.0=pyhd8ed1ab_0
+  - keyutils=1.6.1=h166bdaf_0
+  - kiwisolver=1.4.5=py310hd41b1e2_1
+  - krb5=1.21.2=h659d440_0
+  - langcodes=3.3.0=pyhd8ed1ab_0
+  - lcms2=2.16=hb7c19ff_0
+  - ld_impl_linux-64=2.40=h41732ed_0
+  - lerc=4.0.0=h27087fc_0
+  - libabseil=20230125.0=cxx17_hcb278e6_1
+  - libarchive=3.6.2=h039dbb9_1
+  - libarrow=11.0.0=h33598ff_9_cpu
+  - libblas=3.9.0=21_linux64_openblas
+  - libbrotlicommon=1.0.9=h166bdaf_9
+  - libbrotlidec=1.0.9=h166bdaf_9
+  - libbrotlienc=1.0.9=h166bdaf_9
+  - libcblas=3.9.0=21_linux64_openblas
+  - libcrc32c=1.1.2=h9c3ff4c_0
+  - libcublas=12.1.3.1=hd3aeb46_0
+  - libcublas-dev=12.1.3.1=0
+  - libcudf=23.08.00=cuda12_230809_g8150d38e08_0
+  - libcufft=11.0.2.54=hd3aeb46_0
+  - libcufile=1.8.1.2=0
+  - libcufile-dev=1.8.1.2=0
+  - libcuml=23.08.00=cuda12_230809_gd7162cdea_0
+  - libcumlprims=23.08.00=cuda12_230809_g71c0a86_0
+  - libcurand=10.3.2.106=hd3aeb46_0
+  - libcurand-dev=10.3.2.106=0
+  - libcurl=8.5.0=hca28451_0
+  - libcusolver=11.4.5.107=hd3aeb46_0
+  - libcusolver-dev=11.4.5.107=0
+  - libcusparse=12.1.0.106=hd3aeb46_0
+  - libcusparse-dev=12.1.0.106=0
+  - libdeflate=1.19=hd590300_0
+  - libedit=3.1.20191231=he28a2e2_2
+  - libev=4.33=hd590300_2
+  - libevent=2.1.10=h28343ad_4
+  - libexpat=2.5.0=hcb278e6_1
+  - libffi=3.4.2=h7f98852_5
+  - libgcc-ng=13.2.0=h807b86a_3
+  - libgd=2.3.3=h119a65a_9
+  - libgfortran-ng=13.2.0=h69a702a_3
+  - libgfortran5=13.2.0=ha4646dd_3
+  - libglib=2.78.3=h783c2da_0
+  - libgomp=13.2.0=h807b86a_3
+  - libgoogle-cloud=2.8.0=h3c06191_0
+  - libgrpc=1.51.1=hcf146ea_3
+  - libhwloc=2.9.3=default_h554bfaf_1009
+  - libiconv=1.17=hd590300_2
+  - libjpeg-turbo=3.0.0=hd590300_1
+  - libkvikio=23.08.00=cuda12_230809_g51a9036_0
+  - liblapack=3.9.0=21_linux64_openblas
+  - libllvm14=14.0.6=hcd5def8_4
+  - libmamba=1.5.0=h658169a_0
+  - libmambapy=1.5.0=py310h8aae740_0
+  - libnghttp2=1.58.0=h47da74e_1
+  - libnsl=2.0.1=hd590300_0
+  - libnuma=2.0.16=h0b41bf4_1
+  - libnvjitlink=12.1.105=hd3aeb46_0
+  - libopenblas=0.3.26=pthreads_h413a1c8_0
+  - libpng=1.6.39=h753d276_0
+  - libprotobuf=3.21.12=hfc55251_2
+  - libraft=23.08.00=cuda12_230809_ge588d7b5_0
+  - libraft-headers=23.08.00=cuda12_230809_ge588d7b5_0
+  - libraft-headers-only=23.08.00=cuda12_230809_ge588d7b5_0
+  - librmm=23.08.00=cuda12_230809_gf3af0e8d_0
+  - librsvg=2.56.3=he3f83f7_1
+  - libsodium=1.0.18=h36c2ea0_1
+  - libsolv=0.7.27=hfc55251_0
+  - libsqlite=3.44.2=h2797004_0
+  - libssh2=1.11.0=h0841786_0
+  - libstdcxx-ng=13.2.0=h7e041cc_3
+  - libthrift=0.18.0=h5e4af38_0
+  - libtiff=4.6.0=ha9c0a0a_2
+  - libutf8proc=2.8.0=h166bdaf_0
+  - libuuid=2.38.1=h0b41bf4_0
+  - libuv=1.46.0=hd590300_0
+  - libwebp=1.3.2=h658648e_1
+  - libwebp-base=1.3.2=hd590300_0
+  - libxcb=1.15=h0b41bf4_0
+  - libxcrypt=4.4.36=hd590300_1
+  - libxml2=2.12.3=h232c23b_0
+  - libzlib=1.2.13=hd590300_5
+  - llvm-openmp=8.0.1=hc9558a2_0
+  - locket=1.0.0=pyhd8ed1ab_0
+  - lz4=4.3.3=py310h350c4a5_0
+  - lz4-c=1.9.4=hcb278e6_0
+  - lzo=2.10=h516909a_1000
+  - magma-cuda121=2.6.1=1
+  - mamba=1.5.0=py310h51d5547_0
+  - markdown-it-py=3.0.0=pyhd8ed1ab_0
+  - matplotlib-base=3.8.2=py310h62c0568_0
+  - matplotlib-inline=0.1.6=pyhd8ed1ab_0
+  - mdurl=0.1.2=pyhd8ed1ab_0
+  - menuinst=2.0.1=py310hff52083_0
+  - mkl=2023.1.0=h213fc3f_46344
+  - msgpack-python=1.0.7=py310hd41b1e2_0
+  - multidict=6.0.4=py310h2372a71_1
+  - munkres=1.1.4=pyh9f0ad1d_0
+  - murmurhash=1.0.10=py310hc6cd4ac_1
+  - nb_conda=2.2.1=unix_7
+  - nb_conda_kernels=2.3.1=pyhd8ed1ab_3
+  - nbclassic=1.0.0=pyhb4ecaf3_1
+  - nbconvert-pandoc=7.14.0=pyhd8ed1ab_0
+  - nbformat=5.9.2=pyhd8ed1ab_0
+  - nccl=2.19.4.1=h3a97aeb_0
+  - ncurses=6.4=h59595ed_2
+  - nest-asyncio=1.5.8=pyhd8ed1ab_0
+  - nodejs=20.9.0=hb753e55_0
+  - notebook-shim=0.2.3=pyhd8ed1ab_0
+  - numba=0.58.1=py310h7dc5dd1_0
+  - nvcomp=2.6.1=h10b603f_3
+  - nvtx=0.2.8=py310h2372a71_1
+  - openjpeg=2.5.0=h488ebb8_3
+  - openmp=8.0.1=0
+  - openssl=3.2.0=hd590300_1
+  - orc=1.8.2=hfdbbad2_2
+  - overrides=7.4.0=pyhd8ed1ab_0
+  - pandoc=3.1.3=h32600fe_0
+  - pandocfilters=1.5.0=pyhd8ed1ab_0
+  - pango=1.50.14=ha41ecd1_2
+  - parquet-cpp=1.5.1=2
+  - parso=0.8.3=pyhd8ed1ab_0
+  - partd=1.4.1=pyhd8ed1ab_0
+  - pathy=0.10.3=py310h06a4308_0
+  - pcre2=10.42=hcad00b1_0
+  - perl=5.32.1=7_hd590300_perl5
+  - pickleshare=0.7.5=py_1003
+  - pip=23.3.2=pyhd8ed1ab_0
+  - pixman=0.43.0=h59595ed_0
+  - pkg-config=0.29.2=h36c2ea0_1008
+  - pkgutil-resolve-name=1.3.10=pyhd8ed1ab_1
+  - platformdirs=4.1.0=pyhd8ed1ab_0
+  - pluggy=1.3.0=pyhd8ed1ab_0
+  - preshed=3.0.9=py310hc6cd4ac_1
+  - proj=9.3.1=h1d62c97_0
+  - prometheus_client=0.19.0=pyhd8ed1ab_0
+  - proto-plus=1.23.0=pyhd8ed1ab_0
+  - pthread-stubs=0.4=h36c2ea0_1001
+  - ptyprocess=0.7.0=pyhd3deb0d_0
+  - pure_eval=0.2.2=pyhd8ed1ab_0
+  - pyarrow=11.0.0=py310h633f555_9_cpu
+  - pyasn1=0.5.1=pyhd8ed1ab_0
+  - pyasn1-modules=0.3.0=pyhd8ed1ab_0
+  - pybind11-abi=4=hd8ed1ab_3
+  - pycosat=0.6.6=py310h2372a71_0
+  - pydantic=2.5.3=pyhd8ed1ab_0
+  - pydantic-core=2.14.6=py310hcb5633a_1
+  - pygments=2.17.2=pyhd8ed1ab_0
+  - pylibraft=23.08.00=cuda12_py310_230809_ge588d7b5_0
+  - pynvml=11.4.1=pyhd8ed1ab_0
+  - pyopenssl=23.3.0=pyhd8ed1ab_0
+  - pyparsing=3.1.1=pyhd8ed1ab_0
+  - pyproj=3.6.1=py310hd5c30f3_5
+  - pyshp=2.3.1=pyhd8ed1ab_0
+  - pysocks=1.7.1=py310h06a4308_0
+  - python=3.10.13=hd12c33a_1_cpython
+  - python-dateutil=2.8.2=pyhd8ed1ab_0
+  - python-fastjsonschema=2.19.1=pyhd8ed1ab_0
+  - python-json-logger=2.0.7=pyhd8ed1ab_0
+  - python_abi=3.10=4_cp310
+  - pytz=2023.3.post1=pyhd8ed1ab_0
+  - pyu2f=0.1.5=pyhd8ed1ab_0
+  - pyyaml=6.0.1=py310h2372a71_1
+  - raft-dask=23.08.00=cuda12_py310_230809_ge588d7b5_0
+  - rdma-core=28.9=h59595ed_1
+  - re2=2023.02.02=hcb278e6_0
+  - readline=8.2=h8228510_1
+  - referencing=0.32.1=pyhd8ed1ab_0
+  - reproc=14.2.4.post0=hd590300_1
+  - reproc-cpp=14.2.4.post0=h59595ed_1
+  - requests=2.31.0=pyhd8ed1ab_0
+  - rfc3339-validator=0.1.4=pyhd8ed1ab_0
+  - rfc3986-validator=0.1.1=pyh9f0ad1d_0
+  - rich=13.7.0=pyhd8ed1ab_0
+  - rmm=23.08.00=cuda12_py310_230809_gf3af0e8d_0
+  - rpds-py=0.16.2=py310hcb5633a_0
+  - rsa=4.9=pyhd8ed1ab_0
+  - ruamel.yaml=0.17.40=py310h2372a71_0
+  - ruamel.yaml.clib=0.2.7=py310h2372a71_2
+  - ruamel_yaml=0.15.100=py310h7f8727e_0
+  - s2n=1.3.41=h3358134_0
+  - send2trash=1.8.2=pyh41d4057_0
+  - setuptools=69.0.3=pyhd8ed1ab_0
+  - shapely=1.8.5=py310h5b266fc_2
+  - shellingham=1.5.4=pyhd8ed1ab_0
+  - smart_open=6.4.0=pyhd8ed1ab_0
+  - snappy=1.1.10=h9fff704_0
+  - sniffio=1.3.0=pyhd8ed1ab_0
+  - sortedcontainers=2.4.0=pyhd8ed1ab_0
+  - soupsieve=2.5=pyhd8ed1ab_1
+  - spacy=3.7.2=py310hcb52e73_0
+  - spacy-legacy=3.0.12=pyhd8ed1ab_0
+  - spacy-loggers=1.0.5=pyhd8ed1ab_0
+  - spdlog=1.11.0=h9b3ece8_1
+  - sqlite=3.38.2=hc218d9a_0
+  - srsly=2.4.8=py310hc6cd4ac_1
+  - stack_data=0.6.2=pyhd8ed1ab_0
+  - tblib=3.0.0=pyhd8ed1ab_0
+  - terminado=0.18.0=pyh0d859eb_0
+  - thinc=8.2.2=py310hcb52e73_0
+  - tinycss2=1.2.1=pyhd8ed1ab_0
+  - tk=8.6.13=noxft_h4845f30_101
+  - toolz=0.12.0=pyhd8ed1ab_0
+  - tornado=6.3.3=py310h2372a71_1
+  - tqdm=4.66.1=pyhd8ed1ab_0
+  - traitlets=5.9.0=pyhd8ed1ab_0
+  - treelite=3.2.0=py310h1be96d9_0
+  - truststore=0.8.0=pyhd8ed1ab_0
+  - typer=0.9.0=pyhd8ed1ab_0
+  - types-python-dateutil=2.8.19.20240106=pyhd8ed1ab_0
+  - typing-extensions=4.9.0=hd8ed1ab_0
+  - typing_extensions=4.9.0=pyha770c72_0
+  - typing_utils=0.1.0=pyhd8ed1ab_0
+  - ucx=1.14.1=h195a15c_5
+  - ucx-proc=1.0.0=gpu
+  - ucx-py=0.33.00=py310_230809_gea1eb8f_0
+  - unicodedata2=15.1.0=py310h2372a71_0
+  - uri-template=1.3.0=pyhd8ed1ab_0
+  - wasabi=1.1.2=py310hff52083_0
+  - wcwidth=0.2.13=pyhd8ed1ab_0
+  - weasel=0.3.4=pyhd8ed1ab_0
+  - webcolors=1.13=pyhd8ed1ab_0
+  - webencodings=0.5.1=pyhd8ed1ab_2
+  - websocket-client=1.7.0=pyhd8ed1ab_0
+  - wheel=0.42.0=pyhd8ed1ab_0
+  - xorg-kbproto=1.0.7=h7f98852_1002
+  - xorg-libice=1.1.1=hd590300_0
+  - xorg-libsm=1.2.4=h7391055_0
+  - xorg-libx11=1.8.7=h8ee46fc_0
+  - xorg-libxau=1.0.11=hd590300_0
+  - xorg-libxdmcp=1.1.3=h7f98852_0
+  - xorg-libxext=1.3.4=h0b41bf4_2
+  - xorg-libxrender=0.9.11=hd590300_0
+  - xorg-libxt=1.3.0=hd590300_1
+  - xorg-renderproto=0.11.1=h7f98852_1002
+  - xorg-xextproto=7.3.0=h0b41bf4_1003
+  - xorg-xproto=7.0.31=h7f98852_1007
+  - xyzservices=2023.10.1=pyhd8ed1ab_0
+  - xz=5.2.6=h166bdaf_0
+  - yaml=0.2.5=h7b6447c_0
+  - yaml-cpp=0.7.0=h59595ed_3
+  - zeromq=4.3.5=h59595ed_0
+  - zict=3.0.0=pyhd8ed1ab_0
+  - zipp=3.17.0=pyhd8ed1ab_0
+  - zlib=1.2.13=hd590300_5
+  - zstandard=0.22.0=py310h1275a96_0
+  - zstd=1.5.5=hfc55251_0
+  - pip:
+      - absl-py==1.4.0
+      - accelerate==0.26.1
+      - access==1.1.9
+      - affine==2.4.0
+      - aiobotocore==2.11.0
+      - aiofiles==22.1.0
+      - aiohttp-cors==0.7.0
+      - aioitertools==0.11.0
+      - aiorwlock==1.3.0
+      - aiosqlite==0.19.0
+      - albumentations==1.3.1
+      - alembic==1.13.1
+      - altair==5.2.0
+      - annoy==1.17.3
+      - apache-beam==2.46.0
+      - aplus==0.11.0
+      - appdirs==1.4.4
+      - array-record==0.5.0
+      - arviz==0.17.0
+      - astroid==3.0.2
+      - astropy==6.0.0
+      - astropy-iers-data==0.2024.1.22.0.30.30
+      - astunparse==1.6.3
+      - async-lru==2.0.4
+      - audioread==3.0.1
+      - autopep8==2.0.4
+      - babel==2.14.0
+      - backoff==2.2.1
+      - bayesian-optimization==1.4.3
+      - bayespy==0.5.26
+      - beatrix-jupyterlab==2023.128.151533
+      - bidict==0.22.1
+      - biopython==1.83
+      - blake3==0.2.1
+      - blessed==1.20.0
+      - blinker==1.7.0
+      - blosc2==2.4.0
+      - boruta==0.3
+      - boto3==1.26.100
+      - botocore==1.34.22
+      - bqplot==0.12.42
+      - branca==0.7.0
+      - brewer2mpl==1.4.1
+      - cachetools==4.2.4
+      - catalyst==22.4
+      - catboost==1.2.2
+      - category-encoders==2.6.3
+      - cesium==0.12.1
+      - chex==0.1.85
+      - cleverhans==4.0.0
+      - click-plugins==1.1.1
+      - cligj==0.7.2
+      - cloud-tpu-client==0.10
+      - cloud-tpu-profiler==2.4.0
+      - cloudpickle==2.2.1
+      - cmdstanpy==1.2.0
+      - cmudict==1.0.16
+      - colorcet==3.0.1
+      - colorful==0.5.6
+      - colorlog==6.8.0
+      - colorlover==0.3.0
+      - contextily==1.5.0
+      - convertdate==2.4.0
+      - crcmod==1.7
+      - cufflinks==0.17.3
+      - cvxcanon==0.1.2
+      - cysignals==1.11.4
+      - cython==3.0.8
+      - daal==2024.1.0
+      - daal4py==2024.1.0
+      - dacite==1.8.1
+      - dask==2024.1.0
+      - dataclasses-json==0.6.3
+      - dataproc-jupyter-plugin==0.1.66
+      - datasets==2.1.0
+      - datashader==0.16.0
+      - datatile==1.0.3
+      - db-dtypes==1.2.0
+      - deap==1.4.1
+      - deepdiff==6.7.1
+      - delorean==1.0.0
+      - deprecated==1.2.14
+      - deprecation==2.1.0
+      - descartes==1.1.0
+      - dill==0.3.7
+      - dipy==1.8.0
+      - distlib==0.3.8
+      - distributed==2024.1.0
+      - dm-tree==0.1.8
+      - docker==7.0.0
+      - docker-pycreds==0.4.0
+      - docopt==0.6.2
+      - docstring-parser==0.15
+      - docstring-to-markdown==0.13
+      - docutils==0.20.1
+      - earthengine-api==0.1.386
+      - easydict==1.11
+      - easyocr==1.7.1
+      - ecos==2.0.12
+      - eli5==0.13.0
+      - emoji==2.10.0
+      - en-core-web-lg==3.7.1
+      - en-core-web-sm==3.7.1
+      - ephem==4.1.5
+      - esda==2.5.1
+      - essentia==2.1b6.dev1110
+      - et-xmlfile==1.1.0
+      - etils==1.6.0
+      - evaluate==0.4.1
+      - explainable-ai-sdk==1.3.3
+      - farama-notifications==0.0.4
+      - fastai==2.7.13
+      - fastapi==0.108.0
+      - fastavro==1.9.3
+      - fastcore==1.5.29
+      - fastdownload==0.0.7
+      - fasteners==0.19
+      - fastprogress==1.0.3
+      - fasttext==0.9.2
+      - fbpca==1.0
+      - feather-format==0.4.1
+      - featuretools==1.28.0
+      - filelock==3.13.1
+      - fiona==1.9.5
+      - fitter==1.7.0
+      - flake8==7.0.0
+      - flashtext==2.7
+      - flask==3.0.1
+      - flatbuffers==23.5.26
+      - flax==0.7.5
+      - folium==0.15.1
+      - fonttools==4.47.0
+      - frozendict==2.4.0
+      - funcy==2.0
+      - fury==0.9.0
+      - future==0.18.3
+      - fuzzywuzzy==0.18.0
+      - gast==0.5.4
+      - gatspy==0.3
+      - gcsfs==2023.12.2.post1
+      - gensim==4.3.2
+      - geographiclib==2.0
+      - geohash==1.0
+      - geojson==3.1.0
+      - geopandas==0.14.2
+      - geoplot==0.5.1
+      - geopy==2.4.1
+      - geoviews==1.11.0
+      - ggplot==0.11.5
+      - giddy==2.3.5
+      - gitdb==4.0.11
+      - gitpython==3.1.41
+      - google-ai-generativelanguage==0.4.0
+      - google-api-core==2.15.0
+      - google-api-python-client==2.115.0
+      - google-apitools==0.5.31
+      - google-auth-httplib2==0.1.1
+      - google-auth-oauthlib==1.2.0
+      - google-cloud-aiplatform==0.6.0a1
+      - google-cloud-artifact-registry==1.10.0
+      - google-cloud-automl==1.0.1
+      - google-cloud-bigquery==2.34.4
+      - google-cloud-bigtable==1.7.3
+      - google-cloud-dlp==3.14.0
+      - google-cloud-jupyter-config==0.0.5
+      - google-cloud-language==2.12.0
+      - google-cloud-monitoring==2.18.0
+      - google-cloud-pubsub==2.19.0
+      - google-cloud-pubsublite==1.9.0
+      - google-cloud-recommendations-ai==0.7.1
+      - google-cloud-resource-manager==1.11.0
+      - google-cloud-spanner==3.40.1
+      - google-cloud-storage==1.44.0
+      - google-cloud-translate==3.12.1
+      - google-cloud-videointelligence==2.12.0
+      - google-cloud-vision==2.8.0
+      - google-crc32c==1.5.0
+      - google-generativeai==0.3.2
+      - google-pasta==0.2.0
+      - google-resumable-media==2.7.0
+      - gplearn==0.4.2
+      - gpustat==1.0.0
+      - gpxpy==1.6.2
+      - greenlet==3.0.3
+      - grpc-google-iam-v1==0.12.7
+      - grpcio==1.60.0
+      - grpcio-status==1.48.2
+      - gviz-api==1.10.0
+      - gym==0.26.2
+      - gym-notices==0.0.8
+      - gymnasium==0.29.0
+      - h11==0.14.0
+      - h2o==3.44.0.3
+      - h5netcdf==1.3.0
+      - h5py==3.10.0
+      - haversine==2.8.1
+      - hdfs==2.7.3
+      - hep-ml==0.7.2
+      - hijri-converter==2.3.1
+      - hmmlearn==0.3.0
+      - holidays==0.24
+      - holoviews==1.18.1
+      - hpsklearn==0.1.0
+      - html5lib==1.1
+      - htmlmin==0.1.12
+      - httplib2==0.21.0
+      - httptools==0.6.1
+      - huggingface-hub==0.20.3
+      - humanize==4.9.0
+      - hunspell==0.5.5
+      - husl==4.0.3
+      - hydra-slayer==0.5.0
+      - hyperopt==0.2.7
+      - hypertools==0.8.0
+      - igraph==0.11.3
+      - imagecodecs==2024.1.1
+      - imagehash==4.3.1
+      - imageio==2.33.1
+      - imbalanced-learn==0.11.0
+      - imgaug==0.4.0
+      - importlib-metadata==6.11.0
+      - inequality==1.0.1
+      - iniconfig==2.0.0
+      - ipydatawidgets==4.3.5
+      - ipyleaflet==0.18.2
+      - ipympl==0.7.0
+      - ipython-genutils==0.2.0
+      - ipython-sql==0.5.0
+      - ipyvolume==0.6.3
+      - ipyvue==1.10.1
+      - ipyvuetify==1.8.10
+      - ipywebrtc==0.6.0
+      - ipywidgets==7.7.1
+      - isort==5.13.2
+      - isoweek==1.3.3
+      - itsdangerous==2.1.2
+      - janome==0.5.0
+      - jaraco-classes==3.3.0
+      - jax==0.4.23
+      - jax-jumpy==1.0.0
+      - jaxlib==0.4.23.dev20240116
+      - jeepney==0.8.0
+      - jieba==0.42.1
+      - jmespath==1.0.1
+      - json5==0.9.14
+      - jupyter-client==7.4.9
+      - jupyter-console==6.6.3
+      - jupyter-http-over-ws==0.0.8
+      - jupyter-lsp==1.5.1
+      - jupyter-server==2.12.5
+      - jupyter-server-fileid==0.9.1
+      - jupyter-server-mathjax==0.2.6
+      - jupyter-server-proxy==4.1.0
+      - jupyter-server-ydoc==0.8.0
+      - jupyter-ydoc==0.2.5
+      - jupyterlab==4.0.11
+      - jupyterlab-git==0.44.0
+      - jupyterlab-lsp==5.0.2
+      - jupyterlab-server==2.25.2
+      - jupyterlab-widgets==3.0.9
+      - jupytext==1.16.0
+      - kaggle==1.6.3
+      - kaggle-environments==1.14.3
+      - kagglehub==0.1.6
+      - keras==2.15.0
+      - keras-core==0.1.7
+      - keras-cv==0.8.1
+      - keras-nlp==0.7.0
+      - keras-tuner==1.4.6
+      - kernels-mixer==0.0.7
+      - keyring==24.3.0
+      - keyrings-google-artifactregistry-auth==1.1.2
+      - kfp==2.5.0
+      - kfp-pipeline-spec==0.2.2
+      - kfp-server-api==2.0.5
+      - kmapper==2.0.1
+      - kmodes==0.12.2
+      - korean-lunar-calendar==0.3.1
+      - kornia==0.7.1
+      - kt-legacy==1.0.5
+      - kubernetes==26.1.0
+      - langid==1.1.6
+      - lazy-loader==0.3
+      - learntools==0.3.4
+      - leven==1.0.4
+      - levenshtein==0.23.0
+      - libclang==16.0.6
+      - libpysal==4.9.2
+      - librosa==0.10.1
+      - lightgbm==4.2.0
+      - lightning-utilities==0.10.1
+      - lime==0.2.0.1
+      - line-profiler==4.1.2
+      - linkify-it-py==2.0.2
+      - llvmlite==0.41.1
+      - lml==0.1.0
+      - loguru==0.7.2
+      - lunarcalendar==0.0.9
+      - lxml==5.1.0
+      - mako==1.3.1
+      - mapclassify==2.6.1
+      - marisa-trie==1.1.0
+      - markdown==3.5.2
+      - markovify==0.9.4
+      - markupsafe==2.1.4
+      - marshmallow==3.20.2
+      - matplotlib==3.7.4
+      - matplotlib-venn==0.11.9
+      - mccabe==0.7.0
+      - mdit-py-plugins==0.4.0
+      - memory-profiler==0.61.0
+      - mercantile==1.2.1
+      - mgwr==2.2.1
+      - missingno==0.5.2
+      - mistune==0.8.4
+      - mizani==0.9.3
+      - ml-dtypes==0.2.0
+      - mlcrate==0.2.0
+      - mlens==0.2.3
+      - mlxtend==0.23.1
+      - mmh3==4.1.0
+      - mne==1.6.1
+      - mnist==0.2.2
+      - mock==5.1.0
+      - momepy==0.7.0
+      - more-itertools==10.2.0
+      - mpld3==0.5.10
+      - mpmath==1.3.0
+      - msgpack-numpy==0.4.8
+      - multimethod==1.10
+      - multipledispatch==1.0.0
+      - multiprocess==0.70.15
+      - mypy-extensions==1.0.0
+      - namex==0.0.7
+      - nbclient==0.5.13
+      - nbconvert==6.4.5
+      - nbdime==3.2.0
+      - ndindex==1.7
+      - networkx==3.2.1
+      - nibabel==5.2.0
+      - nilearn==0.10.2
+      - ninja==1.11.1.1
+      - nltk==3.2.4
+      - nose==1.3.7
+      - notebook==6.5.6
+      - notebook-executor==0.2
+      - numexpr==2.8.8
+      - numpy==1.24.4
+      - nvidia-ml-py==11.495.46
+      - oauth2client==4.1.3
+      - oauthlib==3.2.2
+      - objsize==0.6.1
+      - odfpy==1.4.1
+      - olefile==0.47
+      - onnx==1.15.0
+      - opencensus==0.11.4
+      - opencensus-context==0.1.3
+      - opencv-contrib-python==4.9.0.80
+      - opencv-python==4.9.0.80
+      - opencv-python-headless==4.9.0.80
+      - openpyxl==3.1.2
+      - openslide-python==1.3.1
+      - opentelemetry-api==1.22.0
+      - opentelemetry-exporter-otlp==1.22.0
+      - opentelemetry-exporter-otlp-proto-common==1.22.0
+      - opentelemetry-exporter-otlp-proto-grpc==1.22.0
+      - opentelemetry-exporter-otlp-proto-http==1.22.0
+      - opentelemetry-proto==1.22.0
+      - opentelemetry-sdk==1.22.0
+      - opentelemetry-semantic-conventions==0.43b0
+      - opt-einsum==3.3.0
+      - optax==0.1.8
+      - optuna==3.5.0
+      - orbax-checkpoint==0.4.4
+      - ordered-set==4.1.0
+      - orderedmultidict==1.0.1
+      - orjson==3.9.10
+      - ortools==9.4.1874
+      - osmnx==1.8.1
+      - packaging==21.3
+      - pandas==2.2.0
+      - pandas-datareader==0.10.0
+      - pandas-profiling==3.6.6
+      - pandas-summary==0.2.0
+      - pandasql==0.7.3
+      - panel==1.3.7
+      - papermill==2.5.0
+      - param==2.0.2
+      - path==16.9.0
+      - path-py==12.5.0
+      - pathos==0.3.1
+      - patsy==0.5.6
+      - pdf2image==1.17.0
+      - pettingzoo==1.24.0
+      - pexpect==4.9.0
+      - phik==0.12.4
+      - pillow==9.5.0
+      - plotly==5.18.0
+      - plotly-express==0.4.1
+      - plotnine==0.12.4
+      - pointpats==2.4.0
+      - polars==0.20.5
+      - polyglot==16.7.4
+      - pooch==1.8.0
+      - pox==0.3.3
+      - ppca==0.0.4
+      - ppft==1.7.6.7
+      - preprocessing==0.1.13
+      - prettytable==3.9.0
+      - progressbar2==4.3.2
+      - promise==2.3
+      - prompt-toolkit==3.0.43
+      - pronouncing==0.2.0
+      - prophet==1.1.1
+      - protobuf==3.20.3
+      - psutil==5.9.3
+      - pudb==2024.1
+      - pulp==2.8.0
+      - py-cpuinfo==9.0.0
+      - py-spy==0.3.14
+      - py4j==0.10.9.7
+      - pyaml==23.12.0
+      - pyarabic==0.6.15
+      - pyastronomy==0.20.0
+      - pybind11==2.11.1
+      - pyclipper==1.3.0.post5
+      - pycodestyle==2.11.1
+      - pycparser==2.21
+      - pycryptodome==3.20.0
+      - pyct==0.5.0
+      - pycuda==2024.1
+      - pydegensac==0.1.2
+      - pydicom==2.4.4
+      - pydocstyle==6.3.0
+      - pydot==1.4.2
+      - pydub==0.25.1
+      - pyemd==1.0.0
+      - pyerfa==2.0.1.1
+      - pyexcel-io==0.6.6
+      - pyexcel-ods==0.6.0
+      - pyfasttext==0.4.6
+      - pyflakes==3.2.0
+      - pygltflib==1.16.1
+      - pyjwt==2.8.0
+      - pykalman==0.9.5
+      - pyldavis==3.4.1
+      - pylint==3.0.3
+      - pymc3==3.11.4
+      - pymeeus==0.5.12
+      - pymongo==3.13.0
+      - pympler==1.0.1
+      - pynndescent==0.5.11
+      - pynvrtc==9.2
+      - pyocr==0.8.5
+      - pypdf==4.0.0
+      - pysal==23.7
+      - pytesseract==0.3.10
+      - pytest==7.4.4
+      - python-bidi==0.4.2
+      - python-dotenv==1.0.0
+      - python-graphviz==0.20.1
+      - python-levenshtein==0.23.0
+      - python-louvain==0.16
+      - python-lsp-jsonrpc==1.1.2
+      - python-lsp-server==1.10.0
+      - python-slugify==8.0.1
+      - python-utils==3.8.1
+      - pythreejs==2.4.2
+      - pytoolconfig==1.3.1
+      - pytools==2023.1.1
+      - pytorch-ignite==0.4.13
+      - pytorch-lightning==2.1.3
+      - pyupset==0.1.1.post7
+      - pyviz-comms==3.0.1
+      - pywavelets==1.5.0
+      - pyzmq==24.0.1
+      - qgrid==1.3.1
+      - qtconsole==5.5.1
+      - qtpy==2.4.1
+      - quantecon==0.7.1
+      - quantities==0.15.0
+      - qudida==0.0.4
+      - rapidfuzz==3.6.1
+      - rasterio==1.3.9
+      - rasterstats==0.19.0
+      - ray==2.9.0
+      - ray-cpp==2.9.0
+      - regex==2023.12.25
+      - requests-oauthlib==1.3.1
+      - requests-toolbelt==0.10.1
+      - responses==0.18.0
+      - retrying==1.3.4
+      - rgf-python==3.12.0
+      - rich-click==1.7.3
+      - rope==1.12.0
+      - rtree==1.2.0
+      - s2sphere==0.2.5
+      - s3fs==2023.12.2
+      - s3transfer==0.6.2
+      - safetensors==0.4.2
+      - scattertext==0.1.19
+      - scikit-image==0.22.0
+      - scikit-learn==1.2.2
+      - scikit-learn-intelex==2024.1.0
+      - scikit-multilearn==0.2.0
+      - scikit-optimize==0.9.0
+      - scikit-plot==0.3.7
+      - scikit-surprise==1.1.3
+      - scipy==1.11.4
+      - seaborn==0.12.2
+      - secretstorage==3.3.3
+      - segment-anything==1.0
+      - segregation==2.5
+      - semver==3.0.2
+      - sentencepiece==0.1.99
+      - sentry-sdk==1.39.2
+      - seqeval==1.2.2
+      - setproctitle==1.3.3
+      - setuptools-git==1.2
+      - setuptools-scm==8.0.4
+      - shap==0.44.0
+      - shimmy==1.3.0
+      - simpervisor==1.0.0
+      - simpleitk==2.3.1
+      - simplejson==3.19.2
+      - six==1.16.0
+      - sklearn-pandas==2.2.0
+      - slicer==0.0.7
+      - smhasher==0.150.1
+      - smmap==5.0.1
+      - snowballstemmer==2.2.0
+      - snuggs==1.4.7
+      - soundfile==0.12.1
+      - soxr==0.3.7
+      - spaghetti==1.7.4
+      - spectral==0.23.1
+      - spglm==1.1.0
+      - sphinx-rtd-theme==0.2.4
+      - spint==1.0.7
+      - splot==1.1.5.post1
+      - spopt==0.6.0
+      - spreg==1.4.2
+      - spvcm==0.3.0
+      - sqlalchemy==2.0.25
+      - sqlparse==0.4.4
+      - squarify==0.4.3
+      - stable-baselines3==2.1.0
+      - stack-data==0.6.3
+      - stanio==0.3.0
+      - starlette==0.32.0.post1
+      - statsmodels==0.14.1
+      - stemming==1.0.1
+      - stop-words==2018.7.23
+      - stopit==1.1.2
+      - stumpy==1.12.0
+      - sympy==1.12
+      - tables==3.9.2
+      - tabulate==0.9.0
+      - tangled-up-in-unicode==0.2.0
+      - tbb==2021.11.0
+      - tenacity==8.2.3
+      - tensorboard==2.15.1
+      - tensorboard-data-server==0.7.2
+      - tensorboard-plugin-profile==2.15.0
+      - tensorboardx==2.6.2.2
+      - tensorflow==2.15.0
+      - tensorflow-cloud==0.1.16
+      - tensorflow-datasets==4.9.4
+      - tensorflow-decision-forests==1.8.1
+      - tensorflow-estimator==2.15.0
+      - tensorflow-hub==0.15.0
+      - tensorflow-io==0.35.0
+      - tensorflow-io-gcs-filesystem==0.35.0
+      - tensorflow-metadata==0.14.0
+      - tensorflow-probability==0.23.0
+      - tensorflow-serving-api==2.14.1
+      - tensorflow-text==2.15.0
+      - tensorflow-transform==0.14.0
+      - tensorflowjs==4.16.0
+      - tensorpack==0.11
+      - tensorstore==0.1.45
+      - termcolor==2.4.0
+      - testpath==0.6.0
+      - text-unidecode==1.3
+      - textblob==0.17.1
+      - texttable==1.7.0
+      - theano==1.0.5
+      - theano-pymc==1.1.2
+      - threadpoolctl==3.2.0
+      - tifffile==2023.12.9
+      - timm==0.9.12
+      - tobler==0.11.2
+      - tokenizers==0.15.1
+      - toml==0.10.2
+      - tomli==2.0.1
+      - tomlkit==0.12.3
+      - torch==2.1.2
+      - torchaudio==2.1.2
+      - torchdata==0.7.1
+      - torchinfo==1.8.0
+      - torchmetrics==1.3.0.post0
+      - torchtext==0.16.2
+      - torchvision==0.16.2
+      - tpot==0.12.1
+      - traceml==1.0.8
+      - traittypes==0.2.1
+      - transformers==4.37.0
+      - treelite-runtime==3.2.0
+      - trueskill==0.4.5
+      - trx-python==0.2.9
+      - tsfresh==0.20.1
+      - typeguard==4.1.5
+      - typing-inspect==0.9.0
+      - tzdata==2023.4
+      - tzlocal==5.2
+      - uc-micro-py==1.0.2
+      - ujson==5.9.0
+      - umap-learn==0.5.5
+      - unidecode==1.3.8
+      - update-checker==0.18.0
+      - uritemplate==3.0.1
+      - urllib3==1.26.18
+      - urwid==2.4.6
+      - urwid-readline==0.13
+      - uvicorn==0.25.0
+      - uvloop==0.19.0
+      - vaex==4.17.0
+      - vaex-astro==0.9.3
+      - vaex-core==4.17.1
+      - vaex-hdf5==0.14.1
+      - vaex-jupyter==0.8.2
+      - vaex-ml==0.18.3
+      - vaex-server==0.9.0
+      - vaex-viz==0.5.4
+      - vec-noise==1.1.4
+      - vecstack==0.4.0
+      - virtualenv==20.21.0
+      - visions==0.7.5
+      - vowpalwabbit==9.9.0
+      - vtk==9.3.0
+      - wand==0.6.13
+      - wandb==0.16.2
+      - watchfiles==0.21.0
+      - wavio==0.0.8
+      - websockets==12.0
+      - werkzeug==3.0.1
+      - wfdb==4.1.2
+      - whatthepatch==1.0.5
+      - widgetsnbextension==3.6.6
+      - witwidget==1.8.1
+      - woodwork==0.27.0
+      - wordcloud==1.9.3
+      - wordsegment==1.3.1
+      - wrapt==1.14.1
+      - wurlitzer==3.0.3
+      - xarray==2024.1.0
+      - xarray-einstats==0.7.0
+      - xgboost==2.0.3
+      - xvfbwrapper==0.2.9
+      - xxhash==3.4.1
+      - y-py==0.6.2
+      - yapf==0.40.2
+      - yarl==1.9.4
+      - ydata-profiling==4.6.4
+      - yellowbrick==1.5
+      - ypy-websocket==0.8.4
+prefix: /opt/conda

wandb/run-20240221_171730-9wbsk0df/files/config.yaml

@@ -0,0 +1,656 @@
+wandb_version: 1
+
+_wandb:
+  desc: null
+  value:
+    python_version: 3.10.13
+    cli_version: 0.16.2
+    framework: huggingface
+    huggingface_version: 4.37.0
+    is_jupyter_run: true
+    is_kaggle_kernel: true
+    start_time: 1708535850.952207
+    t:
+      1:
+      - 1
+      - 2
+      - 3
+      - 5
+      - 11
+      - 12
+      - 49
+      - 51
+      - 53
+      - 55
+      - 71
+      - 100
+      - 105
+      2:
+      - 1
+      - 2
+      - 3
+      - 5
+      - 11
+      - 12
+      - 30
+      - 35
+      - 49
+      - 51
+      - 53
+      - 55
+      - 71
+      - 100
+      - 105
+      3:
+      - 7
+      - 23
+      4: 3.10.13
+      5: 0.16.2
+      6: 4.37.0
+      8:
+      - 1
+      - 2
+      - 5
+      9:
+        1: transformers_trainer
+      13: linux-x86_64
+    m:
+    - 1: train/global_step
+      6:
+      - 3
+vocab_size:
+  desc: null
+  value: 30522
+max_position_embeddings:
+  desc: null
+  value: 512
+sinusoidal_pos_embds:
+  desc: null
+  value: false
+n_layers:
+  desc: null
+  value: 6
+n_heads:
+  desc: null
+  value: 12
+dim:
+  desc: null
+  value: 768
+hidden_dim:
+  desc: null
+  value: 3072
+dropout:
+  desc: null
+  value: 0.1
+attention_dropout:
+  desc: null
+  value: 0.1
+activation:
+  desc: null
+  value: gelu
+initializer_range:
+  desc: null
+  value: 0.02
+qa_dropout:
+  desc: null
+  value: 0.1
+seq_classif_dropout:
+  desc: null
+  value: 0.2
+return_dict:
+  desc: null
+  value: true
+output_hidden_states:
+  desc: null
+  value: false
+output_attentions:
+  desc: null
+  value: false
+torchscript:
+  desc: null
+  value: false
+torch_dtype:
+  desc: null
+  value: null
+use_bfloat16:
+  desc: null
+  value: false
+tf_legacy_loss:
+  desc: null
+  value: false
+pruned_heads:
+  desc: null
+  value: {}
+tie_word_embeddings:
+  desc: null
+  value: true
+chunk_size_feed_forward:
+  desc: null
+  value: 0
+is_encoder_decoder:
+  desc: null
+  value: false
+is_decoder:
+  desc: null
+  value: false
+cross_attention_hidden_size:
+  desc: null
+  value: null
+add_cross_attention:
+  desc: null
+  value: false
+tie_encoder_decoder:
+  desc: null
+  value: false
+max_length:
+  desc: null
+  value: 20
+min_length:
+  desc: null
+  value: 0
+do_sample:
+  desc: null
+  value: false
+early_stopping:
+  desc: null
+  value: false
+num_beams:
+  desc: null
+  value: 1
+num_beam_groups:
+  desc: null
+  value: 1
+diversity_penalty:
+  desc: null
+  value: 0.0
+temperature:
+  desc: null
+  value: 1.0
+top_k:
+  desc: null
+  value: 50
+top_p:
+  desc: null
+  value: 1.0
+typical_p:
+  desc: null
+  value: 1.0
+repetition_penalty:
+  desc: null
+  value: 1.0
+length_penalty:
+  desc: null
+  value: 1.0
+no_repeat_ngram_size:
+  desc: null
+  value: 0
+encoder_no_repeat_ngram_size:
+  desc: null
+  value: 0
+bad_words_ids:
+  desc: null
+  value: null
+num_return_sequences:
+  desc: null
+  value: 1
+output_scores:
+  desc: null
+  value: false
+return_dict_in_generate:
+  desc: null
+  value: false
+forced_bos_token_id:
+  desc: null
+  value: null
+forced_eos_token_id:
+  desc: null
+  value: null
+remove_invalid_values:
+  desc: null
+  value: false
+exponential_decay_length_penalty:
+  desc: null
+  value: null
+suppress_tokens:
+  desc: null
+  value: null
+begin_suppress_tokens:
+  desc: null
+  value: null
+architectures:
+  desc: null
+  value:
+  - DistilBertForMaskedLM
+finetuning_task:
+  desc: null
+  value: null
+id2label:
+  desc: null
+  value:
+    '0': LABEL_0
+    '1': LABEL_1
+    '2': LABEL_2
+    '3': LABEL_3
+    '4': LABEL_4
+    '5': LABEL_5
+    '6': LABEL_6
+    '7': LABEL_7
+    '8': LABEL_8
+    '9': LABEL_9
+    '10': LABEL_10
+    '11': LABEL_11
+    '12': LABEL_12
+label2id:
+  desc: null
+  value:
+    LABEL_0: 0
+    LABEL_1: 1
+    LABEL_2: 2
+    LABEL_3: 3
+    LABEL_4: 4
+    LABEL_5: 5
+    LABEL_6: 6
+    LABEL_7: 7
+    LABEL_8: 8
+    LABEL_9: 9
+    LABEL_10: 10
+    LABEL_11: 11
+    LABEL_12: 12
+tokenizer_class:
+  desc: null
+  value: null
+prefix:
+  desc: null
+  value: null
+bos_token_id:
+  desc: null
+  value: null
+pad_token_id:
+  desc: null
+  value: 0
+eos_token_id:
+  desc: null
+  value: null
+sep_token_id:
+  desc: null
+  value: null
+decoder_start_token_id:
+  desc: null
+  value: null
+task_specific_params:
+  desc: null
+  value: null
+problem_type:
+  desc: null
+  value: null
+_name_or_path:
+  desc: null
+  value: distilbert-base-uncased
+transformers_version:
+  desc: null
+  value: 4.37.0
+model_type:
+  desc: null
+  value: distilbert
+tie_weights_:
+  desc: null
+  value: true
+output_dir:
+  desc: null
+  value: /kaggle/working
+overwrite_output_dir:
+  desc: null
+  value: false
+do_train:
+  desc: null
+  value: false
+do_eval:
+  desc: null
+  value: true
+do_predict:
+  desc: null
+  value: false
+evaluation_strategy:
+  desc: null
+  value: epoch
+prediction_loss_only:
+  desc: null
+  value: false
+per_device_train_batch_size:
+  desc: null
+  value: 32
+per_device_eval_batch_size:
+  desc: null
+  value: 8
+per_gpu_train_batch_size:
+  desc: null
+  value: null
+per_gpu_eval_batch_size:
+  desc: null
+  value: null
+gradient_accumulation_steps:
+  desc: null
+  value: 1
+eval_accumulation_steps:
+  desc: null
+  value: null
+eval_delay:
+  desc: null
+  value: 0
+learning_rate:
+  desc: null
+  value: 2.0e-05
+weight_decay:
+  desc: null
+  value: 0.01
+adam_beta1:
+  desc: null
+  value: 0.9
+adam_beta2:
+  desc: null
+  value: 0.999
+adam_epsilon:
+  desc: null
+  value: 1.0e-08
+max_grad_norm:
+  desc: null
+  value: 1.0
+num_train_epochs:
+  desc: null
+  value: 2
+max_steps:
+  desc: null
+  value: -1
+lr_scheduler_type:
+  desc: null
+  value: linear
+lr_scheduler_kwargs:
+  desc: null
+  value: {}
+warmup_ratio:
+  desc: null
+  value: 0.0
+warmup_steps:
+  desc: null
+  value: 0
+log_level:
+  desc: null
+  value: passive
+log_level_replica:
+  desc: null
+  value: warning
+log_on_each_node:
+  desc: null
+  value: true
+logging_dir:
+  desc: null
+  value: /kaggle/working/runs/Feb21_17-23-16_047c4dfc8464
+logging_strategy:
+  desc: null
+  value: steps
+logging_first_step:
+  desc: null
+  value: false
+logging_steps:
+  desc: null
+  value: 500
+logging_nan_inf_filter:
+  desc: null
+  value: true
+save_strategy:
+  desc: null
+  value: epoch
+save_steps:
+  desc: null
+  value: 500
+save_total_limit:
+  desc: null
+  value: null
+save_safetensors:
+  desc: null
+  value: true
+save_on_each_node:
+  desc: null
+  value: false
+save_only_model:
+  desc: null
+  value: false
+no_cuda:
+  desc: null
+  value: false
+use_cpu:
+  desc: null
+  value: false
+use_mps_device:
+  desc: null
+  value: false
+seed:
+  desc: null
+  value: 42
+data_seed:
+  desc: null
+  value: null
+jit_mode_eval:
+  desc: null
+  value: false
+use_ipex:
+  desc: null
+  value: false
+bf16:
+  desc: null
+  value: false
+fp16:
+  desc: null
+  value: false
+fp16_opt_level:
+  desc: null
+  value: O1
+half_precision_backend:
+  desc: null
+  value: auto
+bf16_full_eval:
+  desc: null
+  value: false
+fp16_full_eval:
+  desc: null
+  value: false
+tf32:
+  desc: null
+  value: null
+local_rank:
+  desc: null
+  value: 0
+ddp_backend:
+  desc: null
+  value: null
+tpu_num_cores:
+  desc: null
+  value: null
+tpu_metrics_debug:
+  desc: null
+  value: false
+debug:
+  desc: null
+  value: []
+dataloader_drop_last:
+  desc: null
+  value: false
+eval_steps:
+  desc: null
+  value: null
+dataloader_num_workers:
+  desc: null
+  value: 0
+past_index:
+  desc: null
+  value: -1
+run_name:
+  desc: null
+  value: /kaggle/working
+disable_tqdm:
+  desc: null
+  value: false
+remove_unused_columns:
+  desc: null
+  value: true
+label_names:
+  desc: null
+  value: null
+load_best_model_at_end:
+  desc: null
+  value: true
+metric_for_best_model:
+  desc: null
+  value: loss
+greater_is_better:
+  desc: null
+  value: false
+ignore_data_skip:
+  desc: null
+  value: false
+fsdp:
+  desc: null
+  value: []
+fsdp_min_num_params:
+  desc: null
+  value: 0
+fsdp_config:
+  desc: null
+  value:
+    min_num_params: 0
+    xla: false
+    xla_fsdp_grad_ckpt: false
+fsdp_transformer_layer_cls_to_wrap:
+  desc: null
+  value: null
+deepspeed:
+  desc: null
+  value: null
+label_smoothing_factor:
+  desc: null
+  value: 0.0
+optim:
+  desc: null
+  value: adamw_torch
+optim_args:
+  desc: null
+  value: null
+adafactor:
+  desc: null
+  value: false
+group_by_length:
+  desc: null
+  value: false
+length_column_name:
+  desc: null
+  value: length
+report_to:
+  desc: null
+  value:
+  - tensorboard
+  - wandb
+ddp_find_unused_parameters:
+  desc: null
+  value: null
+ddp_bucket_cap_mb:
+  desc: null
+  value: null
+ddp_broadcast_buffers:
+  desc: null
+  value: null
+dataloader_pin_memory:
+  desc: null
+  value: true
+dataloader_persistent_workers:
+  desc: null
+  value: false
+skip_memory_metrics:
+  desc: null
+  value: true
+use_legacy_prediction_loop:
+  desc: null
+  value: false
+push_to_hub:
+  desc: null
+  value: true
+resume_from_checkpoint:
+  desc: null
+  value: null
+hub_model_id:
+  desc: null
+  value: null
+hub_strategy:
+  desc: null
+  value: every_save
+hub_token:
+  desc: null
+  value: <HUB_TOKEN>
+hub_private_repo:
+  desc: null
+  value: false
+hub_always_push:
+  desc: null
+  value: false
+gradient_checkpointing:
+  desc: null
+  value: false
+gradient_checkpointing_kwargs:
+  desc: null
+  value: null
+include_inputs_for_metrics:
+  desc: null
+  value: false
+fp16_backend:
+  desc: null
+  value: auto
+push_to_hub_model_id:
+  desc: null
+  value: null
+push_to_hub_organization:
+  desc: null
+  value: null
+push_to_hub_token:
+  desc: null
+  value: <PUSH_TO_HUB_TOKEN>
+mp_parameters:
+  desc: null
+  value: ''
+auto_find_batch_size:
+  desc: null
+  value: false
+full_determinism:
+  desc: null
+  value: false
+torchdynamo:
+  desc: null
+  value: null
+ray_scope:
+  desc: null
+  value: last
+ddp_timeout:
+  desc: null
+  value: 1800
+torch_compile:
+  desc: null
+  value: false
+torch_compile_backend:
+  desc: null
+  value: null
+torch_compile_mode:
+  desc: null
+  value: null
+dispatch_batches:
+  desc: null
+  value: null
+split_batches:
+  desc: null
+  value: false
+include_tokens_per_second:
+  desc: null
+  value: false
+include_num_input_tokens_seen:
+  desc: null
+  value: false
+neftune_noise_alpha:
+  desc: null
+  value: null

wandb/run-20240221_171730-9wbsk0df/files/output.log

@@ -0,0 +1,4 @@
+2024-02-21 17:18:36,363	INFO util.py:124 -- Outdated packages:
+  ipywidgets==7.7.1 found, needs ipywidgets>=8
+Run `pip install -U ipywidgets`, then restart the notebook server for rich notebook output.
+/opt/conda/lib/python3.10/site-packages/seqeval/metrics/v1.py:57: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.

wandb/run-20240221_171730-9wbsk0df/files/requirements.txt

@@ -0,0 +1,851 @@
+absl-py==1.4.0
+accelerate==0.26.1
+access==1.1.9
+affine==2.4.0
+aiobotocore==2.11.0
+aiofiles==22.1.0
+aiohttp-cors==0.7.0
+aiohttp==3.9.1
+aioitertools==0.11.0
+aiorwlock==1.3.0
+aiosignal==1.3.1
+aiosqlite==0.19.0
+albumentations==1.3.1
+alembic==1.13.1
+altair==5.2.0
+annotated-types==0.6.0
+annoy==1.17.3
+anyio==4.2.0
+apache-beam==2.46.0
+aplus==0.11.0
+appdirs==1.4.4
+archspec==0.2.2
+argon2-cffi-bindings==21.2.0
+argon2-cffi==23.1.0
+array-record==0.5.0
+arrow==1.3.0
+arviz==0.17.0
+astroid==3.0.2
+astropy-iers-data==0.2024.1.22.0.30.30
+astropy==6.0.0
+asttokens==2.4.1
+astunparse==1.6.3
+async-lru==2.0.4
+async-timeout==4.0.3
+attrs==23.2.0
+audioread==3.0.1
+autopep8==2.0.4
+babel==2.14.0
+backoff==2.2.1
+bayesian-optimization==1.4.3
+bayespy==0.5.26
+beatrix-jupyterlab==2023.128.151533
+beautifulsoup4==4.12.2
+bidict==0.22.1
+biopython==1.83
+blake3==0.2.1
+bleach==6.1.0
+blessed==1.20.0
+blinker==1.7.0
+blis==0.7.10
+blosc2==2.4.0
+bokeh==3.3.3
+boltons==23.1.1
+boruta==0.3
+boto3==1.26.100
+botocore==1.34.22
+bq-helper==0.4.1
+bqplot==0.12.42
+branca==0.7.0
+brewer2mpl==1.4.1
+brotli==1.0.9
+brotlipy==0.7.0
+cached-property==1.5.2
+cachetools==4.2.4
+cartopy==0.22.0
+catalogue==2.0.10
+catalyst==22.4
+catboost==1.2.2
+category-encoders==2.6.3
+certifi==2023.11.17
+cesium==0.12.1
+cffi==1.16.0
+charset-normalizer==3.3.2
+chex==0.1.85
+cleverhans==4.0.0
+click-plugins==1.1.1
+click==8.1.7
+cligj==0.7.2
+cloud-tpu-client==0.10
+cloud-tpu-profiler==2.4.0
+cloudpathlib==0.16.0
+cloudpickle==2.2.1
+cmdstanpy==1.2.0
+cmudict==1.0.16
+colorama==0.4.6
+colorcet==3.0.1
+colorful==0.5.6
+colorlog==6.8.0
+colorlover==0.3.0
+comm==0.2.1
+conda-libmamba-solver==23.7.0
+conda-package-handling==2.2.0
+conda-package-streaming==0.9.0
+conda==23.7.4
+confection==0.1.4
+contextily==1.5.0
+contourpy==1.2.0
+convertdate==2.4.0
+crcmod==1.7
+cryptography==41.0.7
+cuda-python==12.3.0
+cudf==23.8.0
+cufflinks==0.17.3
+cuml==23.8.0
+cupy==13.0.0
+cvxcanon==0.1.2
+cycler==0.12.1
+cymem==2.0.8
+cysignals==1.11.4
+cython==3.0.8
+cytoolz==0.12.2
+daal4py==2024.1.0
+daal==2024.1.0
+dacite==1.8.1
+dask-cuda==23.8.0
+dask-cudf==23.8.0
+dask==2024.1.0
+dataclasses-json==0.6.3
+dataproc-jupyter-plugin==0.1.66
+datasets==2.1.0
+datashader==0.16.0
+datatile==1.0.3
+db-dtypes==1.2.0
+deap==1.4.1
+debugpy==1.8.0
+decorator==5.1.1
+deepdiff==6.7.1
+defusedxml==0.7.1
+delorean==1.0.0
+deprecated==1.2.14
+deprecation==2.1.0
+descartes==1.1.0
+dill==0.3.7
+dipy==1.8.0
+distlib==0.3.8
+distributed==2024.1.0
+distro==1.9.0
+dm-tree==0.1.8
+docker-pycreds==0.4.0
+docker==7.0.0
+docopt==0.6.2
+docstring-parser==0.15
+docstring-to-markdown==0.13
+docutils==0.20.1
+earthengine-api==0.1.386
+easydict==1.11
+easyocr==1.7.1
+ecos==2.0.12
+eli5==0.13.0
+emoji==2.10.0
+en-core-web-lg==3.7.1
+en-core-web-sm==3.7.1
+entrypoints==0.4
+ephem==4.1.5
+esda==2.5.1
+essentia==2.1b6.dev1110
+et-xmlfile==1.1.0
+etils==1.6.0
+evaluate==0.4.1
+exceptiongroup==1.2.0
+executing==2.0.1
+explainable-ai-sdk==1.3.3
+farama-notifications==0.0.4
+fastai==2.7.13
+fastapi==0.108.0
+fastavro==1.9.3
+fastcore==1.5.29
+fastdownload==0.0.7
+fasteners==0.19
+fastjsonschema==2.19.1
+fastprogress==1.0.3
+fastrlock==0.8.2
+fasttext==0.9.2
+fbpca==1.0
+feather-format==0.4.1
+featuretools==1.28.0
+filelock==3.13.1
+fiona==1.9.5
+fitter==1.7.0
+flake8==7.0.0
+flashtext==2.7
+flask==3.0.1
+flatbuffers==23.5.26
+flax==0.7.5
+folium==0.15.1
+fonttools==4.47.0
+fqdn==1.5.1
+frozendict==2.4.0
+frozenlist==1.4.1
+fsspec==2023.12.2
+funcy==2.0
+fury==0.9.0
+future==0.18.3
+fuzzywuzzy==0.18.0
+gast==0.5.4
+gatspy==0.3
+gcsfs==2023.12.2.post1
+gensim==4.3.2
+geographiclib==2.0
+geohash==1.0
+geojson==3.1.0
+geopandas==0.14.2
+geoplot==0.5.1
+geopy==2.4.1
+geoviews==1.11.0
+ggplot==0.11.5
+giddy==2.3.5
+gitdb==4.0.11
+gitpython==3.1.41
+google-ai-generativelanguage==0.4.0
+google-api-core==2.11.1
+google-api-python-client==2.115.0
+google-apitools==0.5.31
+google-auth-httplib2==0.1.1
+google-auth-oauthlib==1.2.0
+google-auth==2.26.1
+google-cloud-aiplatform==0.6.0a1
+google-cloud-artifact-registry==1.10.0
+google-cloud-automl==1.0.1
+google-cloud-bigquery==2.34.4
+google-cloud-bigtable==1.7.3
+google-cloud-core==2.4.1
+google-cloud-datastore==2.19.0
+google-cloud-dlp==3.14.0
+google-cloud-jupyter-config==0.0.5
+google-cloud-language==2.12.0
+google-cloud-monitoring==2.18.0
+google-cloud-pubsub==2.19.0
+google-cloud-pubsublite==1.9.0
+google-cloud-recommendations-ai==0.7.1
+google-cloud-resource-manager==1.11.0
+google-cloud-spanner==3.40.1
+google-cloud-storage==1.44.0
+google-cloud-translate==3.12.1
+google-cloud-videointelligence==2.12.0
+google-cloud-vision==2.8.0
+google-crc32c==1.5.0
+google-generativeai==0.3.2
+google-pasta==0.2.0
+google-resumable-media==2.7.0
+googleapis-common-protos==1.62.0
+gplearn==0.4.2
+gpustat==1.0.0
+gpxpy==1.6.2
+graphviz==0.20.1
+greenlet==3.0.3
+grpc-google-iam-v1==0.12.7
+grpcio-status==1.48.1
+grpcio==1.51.1
+gviz-api==1.10.0
+gym-notices==0.0.8
+gym==0.26.2
+gymnasium==0.29.0
+h11==0.14.0
+h2o==3.44.0.3
+h5netcdf==1.3.0
+h5py==3.10.0
+haversine==2.8.1
+hdfs==2.7.3
+hep-ml==0.7.2
+hijri-converter==2.3.1
+hmmlearn==0.3.0
+holidays==0.24
+holoviews==1.18.1
+hpsklearn==0.1.0
+html5lib==1.1
+htmlmin==0.1.12
+httplib2==0.21.0
+httptools==0.6.1
+huggingface-hub==0.20.3
+humanize==4.9.0
+hunspell==0.5.5
+husl==4.0.3
+hydra-slayer==0.5.0
+hyperopt==0.2.7
+hypertools==0.8.0
+idna==3.6
+igraph==0.11.3
+imagecodecs==2024.1.1
+imagehash==4.3.1
+imageio==2.33.1
+imbalanced-learn==0.11.0
+imgaug==0.4.0
+importlib-metadata==6.11.0
+importlib-resources==6.1.1
+inequality==1.0.1
+iniconfig==2.0.0
+ipydatawidgets==4.3.5
+ipykernel==6.28.0
+ipyleaflet==0.18.2
+ipympl==0.7.0
+ipython-genutils==0.2.0
+ipython-sql==0.5.0
+ipython==8.20.0
+ipyvolume==0.6.3
+ipyvue==1.10.1
+ipyvuetify==1.8.10
+ipywebrtc==0.6.0
+ipywidgets==7.7.1
+isoduration==20.11.0
+isort==5.13.2
+isoweek==1.3.3
+itsdangerous==2.1.2
+janome==0.5.0
+jaraco.classes==3.3.0
+jax-jumpy==1.0.0
+jax==0.4.23
+jaxlib==0.4.23.dev20240116
+jedi==0.19.1
+jeepney==0.8.0
+jieba==0.42.1
+jinja2==3.1.2
+jmespath==1.0.1
+joblib==1.3.2
+json5==0.9.14
+jsonpatch==1.33
+jsonpointer==2.4
+jsonschema-specifications==2023.12.1
+jsonschema==4.20.0
+jupyter-client==7.4.9
+jupyter-console==6.6.3
+jupyter-core==5.7.1
+jupyter-events==0.9.0
+jupyter-http-over-ws==0.0.8
+jupyter-lsp==1.5.1
+jupyter-server-fileid==0.9.1
+jupyter-server-mathjax==0.2.6
+jupyter-server-proxy==4.1.0
+jupyter-server-terminals==0.5.1
+jupyter-server-ydoc==0.8.0
+jupyter-server==2.12.5
+jupyter-ydoc==0.2.5
+jupyterlab-git==0.44.0
+jupyterlab-lsp==5.0.2
+jupyterlab-pygments==0.3.0
+jupyterlab-server==2.25.2
+jupyterlab-widgets==3.0.9
+jupyterlab==4.0.11
+jupytext==1.16.0
+kaggle-environments==1.14.3
+kaggle==1.6.3
+kagglehub==0.1.6
+keras-core==0.1.7
+keras-cv==0.8.1
+keras-nlp==0.7.0
+keras-tuner==1.4.6
+keras==2.15.0
+kernels-mixer==0.0.7
+keyring==24.3.0
+keyrings.google-artifactregistry-auth==1.1.2
+kfp-pipeline-spec==0.2.2
+kfp-server-api==2.0.5
+kfp==2.5.0
+kiwisolver==1.4.5
+kmapper==2.0.1
+kmodes==0.12.2
+korean-lunar-calendar==0.3.1
+kornia==0.7.1
+kt-legacy==1.0.5
+kubernetes==26.1.0
+langcodes==3.3.0
+langid==1.1.6
+lazy-loader==0.3
+learntools==0.3.4
+leven==1.0.4
+levenshtein==0.23.0
+libclang==16.0.6
+libmambapy==1.5.0
+libpysal==4.9.2
+librosa==0.10.1
+lightgbm==4.2.0
+lightning-utilities==0.10.1
+lime==0.2.0.1
+line-profiler==4.1.2
+linkify-it-py==2.0.2
+llvmlite==0.41.1
+lml==0.1.0
+locket==1.0.0
+loguru==0.7.2
+lunarcalendar==0.0.9
+lxml==5.1.0
+lz4==4.3.3
+mako==1.3.1
+mamba==1.5.0
+mapclassify==2.6.1
+marisa-trie==1.1.0
+markdown-it-py==3.0.0
+markdown==3.5.2
+markovify==0.9.4
+markupsafe==2.1.3
+marshmallow==3.20.2
+matplotlib-inline==0.1.6
+matplotlib-venn==0.11.9
+matplotlib==3.7.4
+mccabe==0.7.0
+mdit-py-plugins==0.4.0
+mdurl==0.1.2
+memory-profiler==0.61.0
+menuinst==2.0.1
+mercantile==1.2.1
+mgwr==2.2.1
+missingno==0.5.2
+mistune==0.8.4
+mizani==0.9.3
+ml-dtypes==0.2.0
+mlcrate==0.2.0
+mlens==0.2.3
+mlxtend==0.23.1
+mmh3==4.1.0
+mne==1.6.1
+mnist==0.2.2
+mock==5.1.0
+momepy==0.7.0
+more-itertools==10.2.0
+mpld3==0.5.10
+mpmath==1.3.0
+msgpack-numpy==0.4.8
+msgpack==1.0.7
+multidict==6.0.4
+multimethod==1.10
+multipledispatch==1.0.0
+multiprocess==0.70.15
+munkres==1.1.4
+murmurhash==1.0.10
+mypy-extensions==1.0.0
+namex==0.0.7
+nb-conda-kernels==2.3.1
+nb-conda==2.2.1
+nbclassic==1.0.0
+nbclient==0.5.13
+nbconvert==6.4.5
+nbdime==3.2.0
+nbformat==5.9.2
+ndindex==1.7
+nest-asyncio==1.5.8
+networkx==3.2.1
+nibabel==5.2.0
+nilearn==0.10.2
+ninja==1.11.1.1
+nltk==3.2.4
+nose==1.3.7
+notebook-executor==0.2
+notebook-shim==0.2.3
+notebook==6.5.4
+numba==0.58.1
+numexpr==2.8.8
+numpy==1.24.4
+nvidia-ml-py==11.495.46
+nvtx==0.2.8
+oauth2client==4.1.3
+oauthlib==3.2.2
+objsize==0.6.1
+odfpy==1.4.1
+olefile==0.47
+onnx==1.15.0
+opencensus-context==0.1.3
+opencensus==0.11.4
+opencv-contrib-python==4.9.0.80
+opencv-python-headless==4.9.0.80
+opencv-python==4.9.0.80
+openpyxl==3.1.2
+openslide-python==1.3.1
+opentelemetry-api==1.22.0
+opentelemetry-exporter-otlp-proto-common==1.22.0
+opentelemetry-exporter-otlp-proto-grpc==1.22.0
+opentelemetry-exporter-otlp-proto-http==1.22.0
+opentelemetry-exporter-otlp==1.22.0
+opentelemetry-proto==1.22.0
+opentelemetry-sdk==1.22.0
+opentelemetry-semantic-conventions==0.43b0
+opt-einsum==3.3.0
+optax==0.1.8
+optuna==3.5.0
+orbax-checkpoint==0.4.4
+ordered-set==4.1.0
+orderedmultidict==1.0.1
+orjson==3.9.10
+ortools==9.4.1874
+osmnx==1.8.1
+overrides==7.4.0
+packaging==21.3
+pandas-datareader==0.10.0
+pandas-profiling==3.6.6
+pandas-summary==0.2.0
+pandas==2.1.4
+pandasql==0.7.3
+pandocfilters==1.5.0
+panel==1.3.7
+papermill==2.5.0
+param==2.0.2
+parso==0.8.3
+partd==1.4.1
+path.py==12.5.0
+path==16.9.0
+pathos==0.3.1
+pathy==0.10.3
+patsy==0.5.6
+pdf2image==1.17.0
+pettingzoo==1.24.0
+pexpect==4.8.0
+phik==0.12.4
+pickleshare==0.7.5
+pillow==9.5.0
+pip==23.3.2
+pkgutil-resolve-name==1.3.10
+platformdirs==4.1.0
+plotly-express==0.4.1
+plotly==5.18.0
+plotnine==0.12.4
+pluggy==1.3.0
+pointpats==2.4.0
+polars==0.20.5
+polyglot==16.7.4
+pooch==1.8.0
+pox==0.3.3
+ppca==0.0.4
+ppft==1.7.6.7
+preprocessing==0.1.13
+preshed==3.0.9
+prettytable==3.9.0
+progressbar2==4.3.2
+prometheus-client==0.19.0
+promise==2.3
+prompt-toolkit==3.0.42
+pronouncing==0.2.0
+prophet==1.1.1
+proto-plus==1.23.0
+protobuf==3.20.3
+psutil==5.9.3
+ptyprocess==0.7.0
+pudb==2024.1
+pulp==2.8.0
+pure-eval==0.2.2
+py-cpuinfo==9.0.0
+py-spy==0.3.14
+py4j==0.10.9.7
+pyaml==23.12.0
+pyarabic==0.6.15
+pyarrow==11.0.0
+pyasn1-modules==0.3.0
+pyasn1==0.5.1
+pyastronomy==0.20.0
+pybind11==2.11.1
+pyclipper==1.3.0.post5
+pycodestyle==2.11.1
+pycosat==0.6.6
+pycparser==2.21
+pycryptodome==3.20.0
+pyct==0.5.0
+pycuda==2024.1
+pydantic-core==2.14.6
+pydantic==2.5.3
+pydegensac==0.1.2
+pydicom==2.4.4
+pydocstyle==6.3.0
+pydot==1.4.2
+pydub==0.25.1
+pyemd==1.0.0
+pyerfa==2.0.1.1
+pyexcel-io==0.6.6
+pyexcel-ods==0.6.0
+pyfasttext==0.4.6
+pyflakes==3.2.0
+pygltflib==1.16.1
+pygments==2.17.2
+pyjwt==2.8.0
+pykalman==0.9.5
+pyldavis==3.4.1
+pylibraft==23.8.0
+pylint==3.0.3
+pymc3==3.11.4
+pymeeus==0.5.12
+pymongo==3.13.0
+pympler==1.0.1
+pynndescent==0.5.11
+pynvml==11.4.1
+pynvrtc==9.2
+pyocr==0.8.5
+pyopenssl==23.3.0
+pyparsing==3.1.1
+pypdf==4.0.0
+pyproj==3.6.1
+pysal==23.7
+pyshp==2.3.1
+pysocks==1.7.1
+pytesseract==0.3.10
+pytest==7.4.4
+python-bidi==0.4.2
+python-dateutil==2.8.2
+python-dotenv==1.0.0
+python-json-logger==2.0.7
+python-levenshtein==0.23.0
+python-louvain==0.16
+python-lsp-jsonrpc==1.1.2
+python-lsp-server==1.10.0
+python-slugify==8.0.1
+python-utils==3.8.1
+pythreejs==2.4.2
+pytoolconfig==1.3.1
+pytools==2023.1.1
+pytorch-ignite==0.4.13
+pytorch-lightning==2.1.3
+pytz==2023.3.post1
+pyu2f==0.1.5
+pyupset==0.1.1.post7
+pyviz-comms==3.0.1
+pywavelets==1.5.0
+pyyaml==6.0.1
+pyzmq==24.0.1
+qgrid==1.3.1
+qtconsole==5.5.1
+qtpy==2.4.1
+quantecon==0.7.1
+quantities==0.15.0
+qudida==0.0.4
+raft-dask==23.8.0
+rapidfuzz==3.6.1
+rasterio==1.3.9
+rasterstats==0.19.0
+ray-cpp==2.9.0
+ray==2.9.0
+referencing==0.32.1
+regex==2023.12.25
+requests-oauthlib==1.3.1
+requests-toolbelt==0.10.1
+requests==2.31.0
+responses==0.18.0
+retrying==1.3.3
+rfc3339-validator==0.1.4
+rfc3986-validator==0.1.1
+rgf-python==3.12.0
+rich-click==1.7.3
+rich==13.7.0
+rmm==23.8.0
+rope==1.12.0
+rpds-py==0.16.2
+rsa==4.9
+rtree==1.2.0
+ruamel-yaml-conda==0.15.100
+ruamel.yaml.clib==0.2.7
+ruamel.yaml==0.17.40
+s2sphere==0.2.5
+s3fs==2023.12.2
+s3transfer==0.6.2
+safetensors==0.4.2
+scattertext==0.1.19
+scikit-image==0.22.0
+scikit-learn-intelex==2024.1.0
+scikit-learn==1.2.2
+scikit-multilearn==0.2.0
+scikit-optimize==0.9.0
+scikit-plot==0.3.7
+scikit-surprise==1.1.3
+scipy==1.11.4
+seaborn==0.12.2
+secretstorage==3.3.3
+segment-anything==1.0
+segregation==2.5
+semver==3.0.2
+send2trash==1.8.2
+sentencepiece==0.1.99
+sentry-sdk==1.39.2
+seqeval==1.2.2
+setproctitle==1.3.3
+setuptools-git==1.2
+setuptools-scm==8.0.4
+setuptools==69.0.3
+shap==0.44.0
+shapely==1.8.5.post1
+shellingham==1.5.4
+shimmy==1.3.0
+simpervisor==1.0.0
+simpleitk==2.3.1
+simplejson==3.19.2
+six==1.16.0
+sklearn-pandas==2.2.0
+slicer==0.0.7
+smart-open==6.4.0
+smhasher==0.150.1
+smmap==5.0.1
+sniffio==1.3.0
+snowballstemmer==2.2.0
+snuggs==1.4.7
+sortedcontainers==2.4.0
+soundfile==0.12.1
+soupsieve==2.5
+soxr==0.3.7
+spacy-legacy==3.0.12
+spacy-loggers==1.0.5
+spacy==3.7.2
+spaghetti==1.7.4
+spectral==0.23.1
+spglm==1.1.0
+sphinx-rtd-theme==0.2.4
+spint==1.0.7
+splot==1.1.5.post1
+spopt==0.6.0
+spreg==1.4.2
+spvcm==0.3.0
+sqlalchemy==2.0.25
+sqlparse==0.4.4
+squarify==0.4.3
+srsly==2.4.8
+stable-baselines3==2.1.0
+stack-data==0.6.2
+stanio==0.3.0
+starlette==0.32.0.post1
+statsmodels==0.14.1
+stemming==1.0.1
+stop-words==2018.7.23
+stopit==1.1.2
+stumpy==1.12.0
+sympy==1.12
+tables==3.9.2
+tabulate==0.9.0
+tangled-up-in-unicode==0.2.0
+tbb==2021.11.0
+tblib==3.0.0
+tenacity==8.2.3
+tensorboard-data-server==0.7.2
+tensorboard-plugin-profile==2.15.0
+tensorboard==2.15.1
+tensorboardx==2.6.2.2
+tensorflow-cloud==0.1.16
+tensorflow-datasets==4.9.4
+tensorflow-decision-forests==1.8.1
+tensorflow-estimator==2.15.0
+tensorflow-hub==0.15.0
+tensorflow-io-gcs-filesystem==0.35.0
+tensorflow-io==0.35.0
+tensorflow-metadata==0.14.0
+tensorflow-probability==0.23.0
+tensorflow-serving-api==2.14.1
+tensorflow-text==2.15.0
+tensorflow-transform==0.14.0
+tensorflow==2.15.0
+tensorflowjs==4.16.0
+tensorpack==0.11
+tensorstore==0.1.45
+termcolor==2.4.0
+terminado==0.18.0
+testpath==0.6.0
+text-unidecode==1.3
+textblob==0.17.1
+texttable==1.7.0
+theano-pymc==1.1.2
+theano==1.0.5
+thinc==8.2.2
+threadpoolctl==3.2.0
+tifffile==2023.12.9
+timm==0.9.12
+tinycss2==1.2.1
+tobler==0.11.2
+tokenizers==0.15.1
+toml==0.10.2
+tomli==2.0.1
+tomlkit==0.12.3
+toolz==0.12.0
+torch==2.1.2
+torchaudio==2.1.2
+torchdata==0.7.1
+torchinfo==1.8.0
+torchmetrics==1.3.0.post0
+torchtext==0.16.2
+torchvision==0.16.2
+tornado==6.3.3
+tpot==0.12.1
+tqdm==4.66.1
+traceml==1.0.8
+traitlets==5.9.0
+traittypes==0.2.1
+transformers==4.37.0
+treelite-runtime==3.2.0
+treelite==3.2.0
+trueskill==0.4.5
+truststore==0.8.0
+trx-python==0.2.9
+tsfresh==0.20.1
+typeguard==4.1.5
+typer==0.9.0
+types-python-dateutil==2.8.19.20240106
+typing-extensions==4.9.0
+typing-inspect==0.9.0
+typing-utils==0.1.0
+tzdata==2023.4
+tzlocal==5.2
+uc-micro-py==1.0.2
+ucx-py==0.33.0
+ujson==5.9.0
+umap-learn==0.5.5
+unicodedata2==15.1.0
+unidecode==1.3.8
+update-checker==0.18.0
+uri-template==1.3.0
+uritemplate==3.0.1
+urllib3==1.26.18
+urwid-readline==0.13
+urwid==2.4.6
+uvicorn==0.25.0
+uvloop==0.19.0
+vaex-astro==0.9.3
+vaex-core==4.17.1
+vaex-hdf5==0.14.1
+vaex-jupyter==0.8.2
+vaex-ml==0.18.3
+vaex-server==0.9.0
+vaex-viz==0.5.4
+vaex==4.17.0
+vec-noise==1.1.4
+vecstack==0.4.0
+virtualenv==20.21.0
+visions==0.7.5
+vowpalwabbit==9.9.0
+vtk==9.3.0
+wand==0.6.13
+wandb==0.16.2
+wasabi==1.1.2
+watchfiles==0.21.0
+wavio==0.0.8
+wcwidth==0.2.13
+weasel==0.3.4
+webcolors==1.13
+webencodings==0.5.1
+websocket-client==1.7.0
+websockets==12.0
+werkzeug==3.0.1
+wfdb==4.1.2
+whatthepatch==1.0.5
+wheel==0.42.0
+widgetsnbextension==3.6.6
+witwidget==1.8.1
+woodwork==0.27.0
+wordcloud==1.9.3
+wordsegment==1.3.1
+wrapt==1.14.1
+wurlitzer==3.0.3
+xarray-einstats==0.7.0
+xarray==2024.1.0
+xgboost==2.0.3
+xvfbwrapper==0.2.9
+xxhash==3.4.1
+xyzservices==2023.10.1
+y-py==0.6.2
+yapf==0.40.2
+yarl==1.9.3
+ydata-profiling==4.6.4
+yellowbrick==1.5
+ypy-websocket==0.8.4
+zict==3.0.0
+zipp==3.17.0
+zstandard==0.22.0

wandb/run-20240221_171730-9wbsk0df/files/wandb-metadata.json

@@ -0,0 +1,66 @@
+{
+    "os": "Linux-5.15.133+-x86_64-with-glibc2.31",
+    "python": "3.10.13",
+    "heartbeatAt": "2024-02-21T17:17:31.575117",
+    "startedAt": "2024-02-21T17:17:30.944813",
+    "docker": null,
+    "cuda": null,
+    "args": [],
+    "state": "running",
+    "program": "kaggle.ipynb",
+    "codePathLocal": null,
+    "root": "/kaggle/working",
+    "host": "047c4dfc8464",
+    "username": "root",
+    "executable": "/opt/conda/bin/python3.10",
+    "cpu_count": 2,
+    "cpu_count_logical": 4,
+    "cpu_freq": {
+        "current": 2000.14,
+        "min": 0.0,
+        "max": 0.0
+    },
+    "cpu_freq_per_core": [
+        {
+            "current": 2000.14,
+            "min": 0.0,
+            "max": 0.0
+        },
+        {
+            "current": 2000.14,
+            "min": 0.0,
+            "max": 0.0
+        },
+        {
+            "current": 2000.14,
+            "min": 0.0,
+            "max": 0.0
+        },
+        {
+            "current": 2000.14,
+            "min": 0.0,
+            "max": 0.0
+        }
+    ],
+    "disk": {
+        "/": {
+            "total": 8062.387607574463,
+            "used": 5432.354721069336
+        }
+    },
+    "gpu": "Tesla T4",
+    "gpu_count": 2,
+    "gpu_devices": [
+        {
+            "name": "Tesla T4",
+            "memory_total": 16106127360
+        },
+        {
+            "name": "Tesla T4",
+            "memory_total": 16106127360
+        }
+    ],
+    "memory": {
+        "total": 31.357559204101562
+    }
+}

wandb/run-20240221_171730-9wbsk0df/files/wandb-summary.json

@@ -0,0 +1 @@
+{"eval/loss": 0.42691269516944885, "eval/precision": 0.022452124147039402, "eval/recall": 0.4396551724137931, "eval/f1": 0.04272251308900524, "eval/accuracy": 0.9917921579142576, "eval/f5": 0.0591704534467756, "eval/runtime": 32.8688, "eval/samples_per_second": 41.438, "eval/steps_per_second": 2.616, "train/epoch": 1.0, "train/global_step": 86, "_timestamp": 1708536373.207186, "_runtime": 522.2549788951874, "_step": 0}

wandb/run-20240221_171730-9wbsk0df/logs/debug-internal.log

@@ -0,0 +1,661 @@
+2024-02-21 17:17:30,951 INFO    StreamThr :235 [internal.py:wandb_internal():86] W&B internal server running at pid: 235, started at: 2024-02-21 17:17:30.951077
+2024-02-21 17:17:30,955 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status
+2024-02-21 17:17:31,265 INFO    WriterThread:235 [datastore.py:open_for_write():87] open: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/run-9wbsk0df.wandb
+2024-02-21 17:17:31,265 DEBUG   SenderThread:235 [sender.py:send():382] send: header
+2024-02-21 17:17:31,268 DEBUG   SenderThread:235 [sender.py:send():382] send: run
+2024-02-21 17:17:31,450 INFO    SenderThread:235 [dir_watcher.py:__init__():211] watching files in: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files
+2024-02-21 17:17:31,450 INFO    SenderThread:235 [sender.py:_start_run_threads():1136] run started: 9wbsk0df with start time 1708535850.952207
+2024-02-21 17:17:31,458 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: check_version
+2024-02-21 17:17:31,458 DEBUG   SenderThread:235 [sender.py:send_request():409] send_request: check_version
+2024-02-21 17:17:31,547 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: run_start
+2024-02-21 17:17:31,559 DEBUG   HandlerThread:235 [system_info.py:__init__():27] System info init
+2024-02-21 17:17:31,559 DEBUG   HandlerThread:235 [system_info.py:__init__():42] System info init done
+2024-02-21 17:17:31,559 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:17:31,559 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:17:31,560 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:17:31,561 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:17:31,561 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:17:31,562 INFO    HandlerThread:235 [system_monitor.py:probe():214] Collecting system info
+2024-02-21 17:17:31,562 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:17:31,563 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:17:31,575 DEBUG   HandlerThread:235 [system_info.py:probe():151] Probing system
+2024-02-21 17:17:31,578 DEBUG   HandlerThread:235 [gitlib.py:_init_repo():56] git repository is invalid
+2024-02-21 17:17:31,578 DEBUG   HandlerThread:235 [system_info.py:probe():199] Probing system done
+2024-02-21 17:17:31,578 DEBUG   HandlerThread:235 [system_monitor.py:probe():223] {'os': 'Linux-5.15.133+-x86_64-with-glibc2.31', 'python': '3.10.13', 'heartbeatAt': '2024-02-21T17:17:31.575117', 'startedAt': '2024-02-21T17:17:30.944813', 'docker': None, 'cuda': None, 'args': (), 'state': 'running', 'program': 'kaggle.ipynb', 'codePathLocal': None, 'root': '/kaggle/working', 'host': '047c4dfc8464', 'username': 'root', 'executable': '/opt/conda/bin/python3.10', 'cpu_count': 2, 'cpu_count_logical': 4, 'cpu_freq': {'current': 2000.14, 'min': 0.0, 'max': 0.0}, 'cpu_freq_per_core': [{'current': 2000.14, 'min': 0.0, 'max': 0.0}, {'current': 2000.14, 'min': 0.0, 'max': 0.0}, {'current': 2000.14, 'min': 0.0, 'max': 0.0}, {'current': 2000.14, 'min': 0.0, 'max': 0.0}], 'disk': {'/': {'total': 8062.387607574463, 'used': 5432.354721069336}}, 'gpu': 'Tesla T4', 'gpu_count': 2, 'gpu_devices': [{'name': 'Tesla T4', 'memory_total': 16106127360}, {'name': 'Tesla T4', 'memory_total': 16106127360}], 'memory': {'total': 31.357559204101562}}
+2024-02-21 17:17:31,578 INFO    HandlerThread:235 [system_monitor.py:probe():224] Finished collecting system info
+2024-02-21 17:17:31,578 INFO    HandlerThread:235 [system_monitor.py:probe():227] Publishing system info
+2024-02-21 17:17:31,578 DEBUG   HandlerThread:235 [system_info.py:_save_conda():208] Saving list of conda packages installed into the current environment
+2024-02-21 17:17:32,452 INFO    Thread-12 :235 [dir_watcher.py:_on_file_created():271] file/dir created: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/conda-environment.yaml
+2024-02-21 17:18:26,461 DEBUG   HandlerThread:235 [system_info.py:_save_conda():220] Saving conda packages done
+2024-02-21 17:18:26,461 INFO    HandlerThread:235 [system_monitor.py:probe():229] Finished publishing system info
+2024-02-21 17:18:26,467 INFO    Thread-12 :235 [dir_watcher.py:_on_file_modified():288] file/dir modified: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/conda-environment.yaml
+2024-02-21 17:18:26,468 INFO    Thread-12 :235 [dir_watcher.py:_on_file_created():271] file/dir created: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/wandb-metadata.json
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,471 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,472 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,472 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,472 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: python_packages
+2024-02-21 17:18:26,472 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: stop_status
+2024-02-21 17:18:26,473 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: internal_messages
+2024-02-21 17:18:26,473 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:18:26,473 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:18:26,473 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:18:26,474 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:18:26,474 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:18:26,474 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:18:26,474 DEBUG   SenderThread:235 [sender.py:send_request():409] send_request: python_packages
+2024-02-21 17:18:26,474 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:18:26,477 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:18:26,477 DEBUG   SenderThread:235 [sender.py:send_request():409] send_request: stop_status
+2024-02-21 17:18:26,490 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:18:26,490 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:18:26,490 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:18:26,490 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,490 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,490 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,490 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,490 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,491 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,491 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,491 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:26,554 DEBUG   SenderThread:235 [sender.py:send():382] send: telemetry
+2024-02-21 17:18:26,555 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:18:26,555 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:26,556 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:18:26,556 DEBUG   SenderThread:235 [sender.py:send():382] send: telemetry
+2024-02-21 17:18:26,556 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:18:26,557 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:18:26,557 DEBUG   SenderThread:235 [sender.py:send():382] send: telemetry
+2024-02-21 17:18:26,557 DEBUG   SenderThread:235 [sender.py:send():382] send: files
+2024-02-21 17:18:26,557 INFO    SenderThread:235 [sender.py:_save_file():1403] saving file wandb-metadata.json with policy now
+2024-02-21 17:18:26,557 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:18:26,766 INFO    wandb-upload_0:235 [upload_job.py:push():131] Uploaded file /tmp/tmpgge_bdx8wandb/lv5ylz55-wandb-metadata.json
+2024-02-21 17:18:27,468 INFO    Thread-12 :235 [dir_watcher.py:_on_file_created():271] file/dir created: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/requirements.txt
+2024-02-21 17:18:27,590 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:31,563 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:32,470 INFO    Thread-12 :235 [dir_watcher.py:_on_file_modified():288] file/dir modified: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/config.yaml
+2024-02-21 17:18:32,708 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:36,379 DEBUG   SenderThread:235 [sender.py:send():382] send: telemetry
+2024-02-21 17:18:36,471 INFO    Thread-12 :235 [dir_watcher.py:_on_file_created():271] file/dir created: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/output.log
+2024-02-21 17:18:37,380 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:37,485 DEBUG   SenderThread:235 [sender.py:send():382] send: telemetry
+2024-02-21 17:18:37,526 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:18:37,526 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:18:37,526 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:18:37,526 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:18:37,528 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:18:37,529 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:18:37,530 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:18:37,530 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:18:37,530 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:18:37,531 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:18:37,531 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:18:37,531 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:18:37,531 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:18:37,531 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:18:37,533 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:18:37,534 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:18:37,534 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:18:37,725 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:38,472 INFO    Thread-12 :235 [dir_watcher.py:_on_file_modified():288] file/dir modified: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/output.log
+2024-02-21 17:18:40,523 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:18:40,523 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:18:40,523 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:18:40,524 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:18:40,524 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:18:40,525 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:18:40,526 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:18:40,528 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:18:40,528 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:18:40,529 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:18:40,529 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:18:40,530 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:18:40,530 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:18:40,530 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:18:40,531 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:18:40,531 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:18:40,531 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:18:40,553 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:18:40,553 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:18:40,553 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:18:40,553 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:18:42,554 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:42,575 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:18:42,575 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:18:42,575 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:18:42,575 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:18:42,576 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:18:42,577 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:18:42,580 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:18:42,580 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:18:42,581 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:18:43,598 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:43,617 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:18:43,619 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:18:43,619 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:18:43,619 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:18:44,545 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:18:44,546 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:18:44,546 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:18:44,546 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:18:44,547 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:18:44,547 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:18:44,547 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:18:44,547 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:18:44,559 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:18:44,559 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:18:44,560 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:18:44,560 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:18:46,421 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:18:46,422 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:18:46,422 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:18:46,422 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:18:46,423 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:18:46,423 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:18:46,424 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:18:46,425 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:18:46,426 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:18:47,561 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:47,681 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:18:47,682 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:18:47,682 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:18:47,682 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:18:48,054 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:18:48,054 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:18:48,054 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:18:48,055 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:18:48,055 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:18:48,055 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:18:48,056 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:18:48,056 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:18:48,068 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:18:48,068 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:18:48,068 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:18:48,068 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:18:49,314 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:49,315 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:53,070 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:54,315 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:18:58,071 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:18:59,316 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:03,077 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:03,480 INFO    Thread-12 :235 [dir_watcher.py:_on_file_modified():288] file/dir modified: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/config.yaml
+2024-02-21 17:19:04,317 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:08,212 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:09,318 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:13,213 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:14,324 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:17,566 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:19:17,566 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:19:17,566 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:19:17,566 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:19:17,567 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:19:17,568 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:19:17,568 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:19:17,568 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:19:17,568 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:19:17,568 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:19:17,569 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:19:17,569 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:19:17,570 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:19:17,572 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:19:17,573 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:19:18,573 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:19,325 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:22,720 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:19:22,720 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:19:22,720 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:19:22,720 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:19:22,721 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:19:22,722 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:19:22,723 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:19:22,723 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:19:22,725 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:19:22,725 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:19:22,725 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:19:22,725 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:19:22,726 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:19:22,726 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:19:22,726 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:19:22,727 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:19:22,729 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:19:22,747 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:19:22,748 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:19:22,748 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:19:22,748 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:19:23,749 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:24,326 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:28,750 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:29,327 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:29,544 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:19:29,544 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:19:29,544 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:19:29,544 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:19:29,545 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:19:29,545 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:19:29,546 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:19:29,547 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:19:29,547 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:19:29,548 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:19:29,548 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:19:29,549 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:19:29,549 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:19:29,549 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:19:29,549 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:19:29,550 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:19:29,553 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:19:29,572 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:19:29,572 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:19:29,572 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:19:29,573 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:19:33,815 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:19:33,815 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:19:33,815 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:19:33,815 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:19:33,816 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:19:33,816 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:19:33,817 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:19:33,818 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:19:33,818 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:19:33,818 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:19:33,818 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:19:33,818 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:19:33,818 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:19:33,820 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:19:33,822 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:19:33,843 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:19:33,844 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:19:33,845 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:34,328 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:35,014 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:19:35,014 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:19:35,014 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:19:35,015 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:19:35,015 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:19:35,016 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:19:35,020 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:19:35,020 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:19:35,021 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:19:35,941 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:19:35,942 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:19:35,942 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:19:35,942 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:19:36,341 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:19:36,341 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:19:36,341 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:19:36,342 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:19:36,342 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:19:36,342 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:19:36,343 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:19:36,343 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:19:36,354 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:19:36,354 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:19:36,355 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:19:36,355 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:19:37,546 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:19:37,546 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:19:37,546 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:19:37,547 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:19:37,547 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:19:37,548 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:19:37,549 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:19:37,550 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:19:37,551 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:19:38,771 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:19:38,772 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:19:38,772 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:19:38,772 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:19:39,146 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:19:39,146 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:19:39,146 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:19:39,146 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:19:39,146 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:19:39,146 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:19:39,148 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:19:39,148 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:19:39,158 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:19:39,158 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:19:39,158 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:19:39,159 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:19:39,159 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:39,420 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:44,160 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:44,420 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:49,161 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:49,421 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:54,162 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:54,422 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:19:59,163 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:19:59,423 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:04,169 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:04,424 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:04,501 INFO    Thread-12 :235 [dir_watcher.py:_on_file_modified():288] file/dir modified: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/config.yaml
+2024-02-21 17:20:09,300 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:09,425 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:14,301 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:14,426 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:19,302 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:19,427 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:24,303 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:24,428 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:29,304 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:29,429 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:34,305 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:34,430 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:39,306 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:39,431 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:44,307 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:44,432 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:49,308 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:49,433 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:54,309 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:54,435 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:20:59,310 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:20:59,436 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:04,311 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:04,437 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:09,312 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:09,438 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:14,313 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:14,439 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:19,314 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:19,440 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:19,591 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:21:19,591 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:21:19,591 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:21:19,591 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:21:19,592 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:21:19,593 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:21:19,595 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:21:19,596 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:21:19,596 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:21:19,596 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:21:19,597 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:21:19,598 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:21:19,598 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:21:19,598 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:21:19,598 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:21:19,601 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:21:19,619 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:21:19,620 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:21:19,620 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:21:24,441 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:24,622 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:29,442 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:29,622 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:34,443 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:34,623 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:35,971 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:21:35,971 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:21:35,971 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:21:35,971 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:21:35,972 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:21:35,972 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:21:35,973 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:21:35,974 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:21:35,975 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:21:35,975 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:21:35,975 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:21:35,976 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:21:35,976 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:21:35,976 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:21:35,977 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:21:35,977 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:21:35,980 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:21:35,997 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:21:35,997 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:21:35,997 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:21:35,997 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:21:39,444 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:39,998 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:41,191 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:21:41,191 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:21:41,191 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:21:41,191 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:21:41,192 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:21:41,192 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:21:41,193 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:21:41,193 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:21:41,195 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:21:42,153 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:21:42,154 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:21:42,155 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:21:42,155 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:21:42,563 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:21:42,563 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:21:42,563 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:21:42,564 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:21:42,564 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:21:42,565 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:21:42,565 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:21:42,565 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:21:42,577 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:21:42,577 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:21:42,577 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:21:42,578 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:21:43,655 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:21:43,655 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:21:43,655 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:21:43,655 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:21:43,656 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:21:43,656 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:21:43,658 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:21:43,658 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:21:43,659 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:21:44,520 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:44,873 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:21:44,874 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:21:44,874 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:21:44,874 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:21:45,874 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:47,046 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:21:47,046 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:21:47,046 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:21:47,047 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:21:47,047 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:21:47,047 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:21:47,048 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:21:47,048 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:21:47,059 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:21:47,059 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:21:47,059 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:21:47,060 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:21:49,521 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:51,061 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:54,522 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:21:56,062 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:21:59,523 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:01,063 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:04,524 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:06,069 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:06,543 INFO    Thread-12 :235 [dir_watcher.py:_on_file_modified():288] file/dir modified: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/config.yaml
+2024-02-21 17:22:09,525 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:11,178 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:14,526 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:16,179 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:19,527 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:21,180 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:24,528 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:26,181 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:29,529 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:31,182 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:34,530 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:36,183 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:39,531 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:41,184 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:44,532 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:46,185 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:49,534 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:51,186 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:54,535 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:22:56,187 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:22:59,536 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:01,188 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:04,537 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:06,188 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:07,592 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:23:07,592 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:23:07,592 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:23:07,592 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:23:07,593 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:23:07,593 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:23:07,594 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:23:07,596 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:23:07,596 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:23:07,596 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:23:07,596 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:23:07,597 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:23:07,598 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:23:07,598 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:23:07,598 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:23:07,600 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:23:07,601 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:23:07,619 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:23:07,619 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:23:07,619 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:23:07,619 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:23:09,538 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:09,896 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:23:09,896 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:23:09,896 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:23:09,896 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:23:09,897 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:23:09,898 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:23:09,898 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:23:09,899 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:23:09,901 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:23:11,620 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:14,690 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:14,722 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:23:14,724 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:23:14,724 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:23:14,724 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:23:15,140 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: pause
+2024-02-21 17:23:15,140 INFO    HandlerThread:235 [handler.py:handle_request_pause():708] stopping system metrics thread
+2024-02-21 17:23:15,140 INFO    HandlerThread:235 [system_monitor.py:finish():203] Stopping system monitor
+2024-02-21 17:23:15,140 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:23:15,141 DEBUG   SystemMonitor:235 [system_monitor.py:_start():179] Finished system metrics aggregation loop
+2024-02-21 17:23:15,141 DEBUG   SystemMonitor:235 [system_monitor.py:_start():183] Publishing last batch of metrics
+2024-02-21 17:23:15,142 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined cpu monitor
+2024-02-21 17:23:15,142 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined disk monitor
+2024-02-21 17:23:15,153 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined gpu monitor
+2024-02-21 17:23:15,154 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined memory monitor
+2024-02-21 17:23:15,154 INFO    HandlerThread:235 [interfaces.py:finish():202] Joined network monitor
+2024-02-21 17:23:15,154 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:23:16,281 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: resume
+2024-02-21 17:23:16,281 INFO    HandlerThread:235 [handler.py:handle_request_resume():699] starting system metrics thread
+2024-02-21 17:23:16,281 INFO    HandlerThread:235 [system_monitor.py:start():194] Starting system monitor
+2024-02-21 17:23:16,281 INFO    SystemMonitor:235 [system_monitor.py:_start():158] Starting system asset monitoring threads
+2024-02-21 17:23:16,281 INFO    SystemMonitor:235 [interfaces.py:start():190] Started cpu monitoring
+2024-02-21 17:23:16,282 INFO    SystemMonitor:235 [interfaces.py:start():190] Started disk monitoring
+2024-02-21 17:23:16,284 INFO    SystemMonitor:235 [interfaces.py:start():190] Started gpu monitoring
+2024-02-21 17:23:16,285 INFO    SystemMonitor:235 [interfaces.py:start():190] Started memory monitoring
+2024-02-21 17:23:16,285 INFO    SystemMonitor:235 [interfaces.py:start():190] Started network monitoring
+2024-02-21 17:23:17,155 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:19,972 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:21,590 DEBUG   SenderThread:235 [sender.py:send():382] send: config
+2024-02-21 17:23:21,591 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:23:21,591 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:23:21,591 WARNING SenderThread:235 [sender.py:send_metric():1354] Seen metric with glob (shouldn't happen)
+2024-02-21 17:23:22,592 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:25,241 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:27,592 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:30,242 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:32,593 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:35,243 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:37,599 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:38,576 INFO    Thread-12 :235 [dir_watcher.py:_on_file_modified():288] file/dir modified: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/config.yaml
+2024-02-21 17:23:40,246 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:42,743 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:45,247 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:47,744 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:50,248 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:52,745 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:23:55,249 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:23:57,746 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:00,250 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:02,747 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:05,256 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:07,747 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:10,257 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:12,748 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:15,261 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:16,286 DEBUG   SystemMonitor:235 [system_monitor.py:_start():172] Starting system metrics aggregation loop
+2024-02-21 17:24:16,293 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:24:18,293 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:20,262 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:23,294 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:25,263 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:28,295 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:30,264 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:33,296 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:35,265 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:38,297 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:40,266 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:43,297 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:45,267 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:46,288 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:24:49,289 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:50,268 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:54,290 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:24:55,270 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:24:59,291 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:00,270 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:04,291 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:05,281 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:09,292 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:10,282 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:14,293 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:15,283 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:16,289 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:25:20,284 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:20,291 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:25,285 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:25,292 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:30,293 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:30,293 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:35,293 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:35,304 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:40,294 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:40,305 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:45,295 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:45,306 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:46,289 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:25:50,307 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:51,291 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:25:55,318 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:25:56,292 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:26:00,380 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:26:01,292 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:26:05,416 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:26:06,378 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:26:07,628 INFO    Thread-12 :235 [dir_watcher.py:_on_file_modified():288] file/dir modified: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/output.log
+2024-02-21 17:26:10,453 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:26:11,379 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:26:13,207 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: partial_history
+2024-02-21 17:26:13,212 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,212 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,212 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,212 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,212 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,212 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,212 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,213 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,213 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,213 DEBUG   SenderThread:235 [sender.py:send():382] send: metric
+2024-02-21 17:26:13,213 DEBUG   SenderThread:235 [sender.py:send():382] send: history
+2024-02-21 17:26:13,213 DEBUG   SenderThread:235 [sender.py:send_request():409] send_request: summary_record
+2024-02-21 17:26:13,214 INFO    SenderThread:235 [sender.py:_save_file():1403] saving file wandb-summary.json with policy end
+2024-02-21 17:26:13,630 INFO    Thread-12 :235 [dir_watcher.py:_on_file_created():271] file/dir created: /kaggle/working/wandb/run-20240221_171730-9wbsk0df/files/wandb-summary.json
+2024-02-21 17:26:15,610 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:26:16,290 DEBUG   SenderThread:235 [sender.py:send():382] send: stats
+2024-02-21 17:26:17,291 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report
+2024-02-21 17:26:20,611 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: keepalive
+2024-02-21 17:26:22,292 DEBUG   HandlerThread:235 [handler.py:handle_request():146] handle_request: status_report

wandb/run-20240221_171730-9wbsk0df/logs/debug.log

@@ -0,0 +1,90 @@
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Current SDK version is 0.16.2
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Configure stats pid to 34
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Loading settings from /root/.config/wandb/settings
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Loading settings from /kaggle/working/wandb/settings
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Loading settings from environment variables: {}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Inferring run settings from compute environment: {'program': '<python with no main file>'}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Applying login settings: {'api_key': '***REDACTED***'}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_setup.py:_flush():76] Applying login settings: {'api_key': '***REDACTED***'}
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_init.py:_log_setup():526] Logging user logs to /kaggle/working/wandb/run-20240221_171730-9wbsk0df/logs/debug.log
+2024-02-21 17:17:30,946 INFO    MainThread:34 [wandb_init.py:_log_setup():527] Logging internal logs to /kaggle/working/wandb/run-20240221_171730-9wbsk0df/logs/debug-internal.log
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:_jupyter_setup():472] configuring jupyter hooks <wandb.sdk.wandb_init._WandbInit object at 0x79ab9071fd60>
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:init():566] calling init triggers
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:init():573] wandb.init called with sweep_config: {}
+config: {}
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:init():616] starting backend
+2024-02-21 17:17:30,947 INFO    MainThread:34 [wandb_init.py:init():620] setting up manager
+2024-02-21 17:17:30,949 INFO    MainThread:34 [backend.py:_multiprocessing_setup():105] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2024-02-21 17:17:30,951 INFO    MainThread:34 [wandb_init.py:init():628] backend started and connected
+2024-02-21 17:17:30,963 INFO    MainThread:34 [wandb_run.py:_label_probe_notebook():1294] probe notebook
+2024-02-21 17:17:31,264 INFO    MainThread:34 [wandb_init.py:init():720] updated telemetry
+2024-02-21 17:17:31,267 INFO    MainThread:34 [wandb_init.py:init():753] communicating run to backend with 90.0 second timeout
+2024-02-21 17:17:31,457 INFO    MainThread:34 [wandb_run.py:_on_init():2254] communicating current version
+2024-02-21 17:17:31,540 INFO    MainThread:34 [wandb_run.py:_on_init():2263] got version response upgrade_message: "wandb version 0.16.3 is available!  To upgrade, please run:\n $ pip install wandb --upgrade"
+
+2024-02-21 17:17:31,540 INFO    MainThread:34 [wandb_init.py:init():804] starting run threads in backend
+2024-02-21 17:18:01,575 INFO    MainThread:34 [wandb_run.py:_console_start():2233] atexit reg
+2024-02-21 17:18:01,576 INFO    MainThread:34 [wandb_run.py:_redirect():2088] redirect: wrap_raw
+2024-02-21 17:18:01,576 INFO    MainThread:34 [wandb_run.py:_redirect():2153] Wrapping output streams.
+2024-02-21 17:18:01,576 INFO    MainThread:34 [wandb_run.py:_redirect():2178] Redirects installed.
+2024-02-21 17:18:01,577 INFO    MainThread:34 [wandb_init.py:init():847] run started, returning control to user process
+2024-02-21 17:18:01,585 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-17-29_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:18:06,099 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:06,099 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:18:37,525 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:18:37,529 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:37,529 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:18:40,523 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:18:40,528 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:40,528 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:18:42,574 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:18:43,613 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-18-42_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:18:44,545 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:44,545 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:18:46,421 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:18:47,676 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-18-46_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:18:48,053 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:18:48,054 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:17,565 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:17,567 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:17,567 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:22,718 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:22,723 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:22,724 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:29,543 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:29,547 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:29,547 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:33,814 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:33,815 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:33,815 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:35,014 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:35,936 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-19-35_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:19:36,340 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:36,340 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:19:37,545 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:19:38,762 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-19-37_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:19:39,145 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:19:39,145 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:21:19,590 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:21:19,595 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:21:19,595 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:21:35,970 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:21:35,973 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:21:35,974 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:21:41,190 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:21:42,149 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-21-41_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:21:42,563 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:21:42,563 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:21:43,654 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:21:44,868 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-21-43_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:21:47,045 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:21:47,046 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:23:07,591 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:23:07,593 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:23:07,594 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:23:09,895 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:23:14,718 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-23-09_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}
+2024-02-21 17:23:15,139 INFO    MainThread:34 [jupyter.py:save_ipynb():373] not saving jupyter notebook
+2024-02-21 17:23:15,139 INFO    MainThread:34 [wandb_init.py:_pause_backend():437] pausing backend
+2024-02-21 17:23:16,280 INFO    MainThread:34 [wandb_init.py:_resume_backend():442] resuming backend
+2024-02-21 17:23:21,586 INFO    MainThread:34 [wandb_run.py:_config_callback():1342] config_cb None None {'vocab_size': 30522, 'max_position_embeddings': 512, 'sinusoidal_pos_embds': False, 'n_layers': 6, 'n_heads': 12, 'dim': 768, 'hidden_dim': 3072, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation': 'gelu', 'initializer_range': 0.02, 'qa_dropout': 0.1, 'seq_classif_dropout': 0.2, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'tf_legacy_loss': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'typical_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'exponential_decay_length_penalty': None, 'suppress_tokens': None, 'begin_suppress_tokens': None, 'architectures': ['DistilBertForMaskedLM'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2', 3: 'LABEL_3', 4: 'LABEL_4', 5: 'LABEL_5', 6: 'LABEL_6', 7: 'LABEL_7', 8: 'LABEL_8', 9: 'LABEL_9', 10: 'LABEL_10', 11: 'LABEL_11', 12: 'LABEL_12'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2, 'LABEL_3': 3, 'LABEL_4': 4, 'LABEL_5': 5, 'LABEL_6': 6, 'LABEL_7': 7, 'LABEL_8': 8, 'LABEL_9': 9, 'LABEL_10': 10, 'LABEL_11': 11, 'LABEL_12': 12}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 0, 'eos_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'distilbert-base-uncased', 'transformers_version': '4.37.0', 'model_type': 'distilbert', 'tie_weights_': True, 'output_dir': '/kaggle/working', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'epoch', 'prediction_loss_only': False, 'per_device_train_batch_size': 32, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 1, 'eval_accumulation_steps': None, 'eval_delay': 0, 'learning_rate': 2e-05, 'weight_decay': 0.01, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 2, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.0, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': '/kaggle/working/runs/Feb21_17-23-16_047c4dfc8464', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 500, 'logging_nan_inf_filter': True, 'save_strategy': 'epoch', 'save_steps': 500, 'save_total_limit': None, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': False, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': '/kaggle/working', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': True, 'metric_for_best_model': 'loss', 'greater_is_better': False, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False}, 'fsdp_transformer_layer_cls_to_wrap': None, 'deepspeed': None, 'label_smoothing_factor': 0.0, 'optim': 'adamw_torch', 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['tensorboard', 'wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': False, 'hub_always_push': False, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'dispatch_batches': None, 'split_batches': False, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None}