w-ahmad commited on
Commit
aed94ee
Β·
verified Β·
1 Parent(s): 2dfc4fd

Auto upload 2026-08-07T22:11:20.527314 (part 2)

Browse files
wandb/debug-internal.log CHANGED
@@ -83,3 +83,43 @@
83
  {"time":"2026-08-07T22:06:03.437526634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
84
  {"time":"2026-08-07T22:06:18.191505048Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":42,"history_lines":1,"events_offset":72,"events_lines":2,"console_offset":66,"console_lines":12}
85
  {"time":"2026-08-07T22:06:18.464200347Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83
  {"time":"2026-08-07T22:06:03.437526634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
84
  {"time":"2026-08-07T22:06:18.191505048Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":42,"history_lines":1,"events_offset":72,"events_lines":2,"console_offset":66,"console_lines":12}
85
  {"time":"2026-08-07T22:06:18.464200347Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
86
+ {"time":"2026-08-07T22:06:33.193453422Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":43,"history_lines":1,"events_offset":74,"events_lines":2,"console_offset":76,"console_lines":1}
87
+ {"time":"2026-08-07T22:06:33.498831625Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
88
+ {"time":"2026-08-07T22:06:48.191104459Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":44,"history_lines":2,"events_offset":76,"events_lines":2,"console_offset":76,"console_lines":1}
89
+ {"time":"2026-08-07T22:06:48.459530923Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
90
+ {"time":"2026-08-07T22:07:03.191131801Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":46,"history_lines":1,"events_offset":78,"events_lines":2,"console_offset":76,"console_lines":1}
91
+ {"time":"2026-08-07T22:07:03.469479856Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
92
+ {"time":"2026-08-07T22:07:18.193514266Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":47,"history_lines":1,"events_offset":80,"events_lines":2,"console_offset":76,"console_lines":1}
93
+ {"time":"2026-08-07T22:07:18.479043758Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
94
+ {"time":"2026-08-07T22:07:33.190853404Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":48,"history_lines":1,"events_offset":82,"events_lines":2,"console_offset":76,"console_lines":1}
95
+ {"time":"2026-08-07T22:07:33.436649349Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
96
+ {"time":"2026-08-07T22:07:48.192527365Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":49,"history_lines":1,"events_offset":84,"events_lines":2,"console_offset":78,"console_lines":12}
97
+ {"time":"2026-08-07T22:07:48.506679473Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
98
+ {"time":"2026-08-07T22:08:03.192547352Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":50,"history_lines":1,"events_offset":86,"events_lines":2,"console_offset":88,"console_lines":1}
99
+ {"time":"2026-08-07T22:08:03.451276923Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
100
+ {"time":"2026-08-07T22:08:18.191893467Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":51,"history_lines":2,"events_offset":88,"events_lines":2,"console_offset":88,"console_lines":1}
101
+ {"time":"2026-08-07T22:08:18.419712526Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
102
+ {"time":"2026-08-07T22:08:33.190848976Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":53,"history_lines":1,"events_offset":90,"events_lines":2,"console_offset":88,"console_lines":1}
103
+ {"time":"2026-08-07T22:08:33.447595074Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
104
+ {"time":"2026-08-07T22:08:48.193871302Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":54,"history_lines":1,"events_offset":92,"events_lines":2,"console_offset":88,"console_lines":1}
105
+ {"time":"2026-08-07T22:08:48.456096357Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
106
+ {"time":"2026-08-07T22:09:03.190928744Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":55,"history_lines":1,"events_offset":94,"events_lines":2,"console_offset":88,"console_lines":1}
107
+ {"time":"2026-08-07T22:09:03.484898063Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
108
+ {"time":"2026-08-07T22:09:18.191470424Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":56,"history_lines":1,"events_offset":96,"events_lines":2,"console_offset":90,"console_lines":12}
109
+ {"time":"2026-08-07T22:09:18.463857634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
110
+ {"time":"2026-08-07T22:09:33.191222201Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":57,"history_lines":1,"events_offset":98,"events_lines":2,"console_offset":100,"console_lines":1}
111
+ {"time":"2026-08-07T22:09:33.473907457Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
112
+ {"time":"2026-08-07T22:09:48.190733549Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":58,"history_lines":2,"events_offset":100,"events_lines":2,"console_offset":100,"console_lines":1}
113
+ {"time":"2026-08-07T22:09:48.427605817Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
114
+ {"time":"2026-08-07T22:10:03.190514166Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":60,"history_lines":1,"events_offset":102,"events_lines":2,"console_offset":100,"console_lines":1}
115
+ {"time":"2026-08-07T22:10:03.466281777Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
116
+ {"time":"2026-08-07T22:10:18.194146776Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":61,"history_lines":1,"events_offset":104,"events_lines":2,"console_offset":100,"console_lines":1}
117
+ {"time":"2026-08-07T22:10:18.482288854Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
118
+ {"time":"2026-08-07T22:10:33.190161734Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":62,"history_lines":1,"events_offset":106,"events_lines":2,"console_offset":100,"console_lines":1}
119
+ {"time":"2026-08-07T22:10:33.436720492Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
120
+ {"time":"2026-08-07T22:10:48.191398564Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":63,"history_lines":1,"events_offset":108,"events_lines":2,"console_offset":102,"console_lines":12}
121
+ {"time":"2026-08-07T22:10:48.472794309Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
122
+ {"time":"2026-08-07T22:11:03.190446716Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":64,"history_lines":1,"events_offset":110,"events_lines":2,"console_offset":112,"console_lines":1}
123
+ {"time":"2026-08-07T22:11:03.461724204Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
124
+ {"time":"2026-08-07T22:11:18.191588934Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":65,"history_lines":2,"events_offset":112,"events_lines":2,"console_offset":112,"console_lines":1}
125
+ {"time":"2026-08-07T22:11:18.454035207Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
wandb/run-20260807_215701-rojwnt3e/files/output.log CHANGED
@@ -74,5 +74,45 @@ Writing model shards: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ
74
  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
75
  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
76
  Writing model shards: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 1/1 [00:00<00:00, 82.06it/s]
77
- 64%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–Š | 636/1000 [09:16<04:14, 1.43it/s]
78
  {'loss': '2.952', 'grad_norm': '0.7812', 'learning_rate': '0.001', 'epoch': '0.3342', 'train/total_time_seconds': '197', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '544.3', 'train/estimated_remaining_minutes': '2.012'}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
74
  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
75
  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
76
  Writing model shards: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 1/1 [00:00<00:00, 82.06it/s]
77
+ 70%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–Ž | 700/1000 [10:19<03:30, 1.43it/s][transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From πŸ‘‰v4.50πŸ‘ˆ onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
78
  {'loss': '2.952', 'grad_norm': '0.7812', 'learning_rate': '0.001', 'epoch': '0.3342', 'train/total_time_seconds': '197', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '544.3', 'train/estimated_remaining_minutes': '2.012'}
79
+ {'loss': '2.913', 'grad_norm': '0.4492', 'learning_rate': '0.001', 'epoch': '0.345', 'train/total_time_seconds': '203.3', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '558.4', 'train/estimated_remaining_minutes': '1.906'}
80
+ - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
81
+ {'eval_loss': '2.905', 'eval_runtime': '9.152', 'eval_samples_per_second': '1041', 'eval_steps_per_second': '2.076', 'epoch': '0.3504', 'train/total_time_seconds': '206.5', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '574.6', 'train/estimated_remaining_minutes': '1.853'}
82
+ {'loss': '2.9', 'grad_norm': '0.9883', 'learning_rate': '0.001', 'epoch': '0.3558', 'train/total_time_seconds': '209.6', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '581.6', 'train/estimated_remaining_minutes': '1.8'}
83
+ {'loss': '2.878', 'grad_norm': '0.5664', 'learning_rate': '0.001', 'epoch': '0.3666', 'train/total_time_seconds': '215.9', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '595.6', 'train/estimated_remaining_minutes': '1.694'}
84
+ {'loss': '2.852', 'grad_norm': '0.9258', 'learning_rate': '0.001', 'epoch': '0.3774', 'train/total_time_seconds': '222.3', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '609.7', 'train/estimated_remaining_minutes': '1.588'}
85
+ {'eval_loss': '2.868', 'eval_runtime': '9.203', 'eval_samples_per_second': '1035', 'eval_steps_per_second': '2.065', 'epoch': '0.3774', 'train/total_time_seconds': '222.3', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '618.9', 'train/estimated_remaining_minutes': '1.588'}
86
+ - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
87
+ - If you are not the owner of the model architecture class, please contact the model code owner to update it.
88
+ Writing model shards: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 1/1 [00:00<00:00, 62.12it/s]
89
+ 80%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ– | 800/1000 [11:48<02:21, 1.41it/s][transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From πŸ‘‰v4.50πŸ‘ˆ onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
90
+ {'loss': '2.847', 'grad_norm': '0.4863', 'learning_rate': '0.001', 'epoch': '0.3881', 'train/total_time_seconds': '228.6', 'train/time_per_step_avg': '0.316', 'train/epoch_time_elapsed': '633', 'train/estimated_remaining_minutes': '1.482'}
91
+ {'loss': '2.815', 'grad_norm': '0.8242', 'learning_rate': '0.001', 'epoch': '0.3989', 'train/total_time_seconds': '234.9', 'train/time_per_step_avg': '0.3159', 'train/epoch_time_elapsed': '647', 'train/estimated_remaining_minutes': '1.375'}
92
+ - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
93
+ {'eval_loss': '2.813', 'eval_runtime': '9.119', 'eval_samples_per_second': '1045', 'eval_steps_per_second': '2.084', 'epoch': '0.4043', 'train/total_time_seconds': '238', 'train/time_per_step_avg': '0.3159', 'train/epoch_time_elapsed': '663.2', 'train/estimated_remaining_minutes': '1.322'}
94
+ {'loss': '2.807', 'grad_norm': '0.543', 'learning_rate': '0.001', 'epoch': '0.4097', 'train/total_time_seconds': '241.2', 'train/time_per_step_avg': '0.3159', 'train/epoch_time_elapsed': '670.3', 'train/estimated_remaining_minutes': '1.27'}
95
+ {'loss': '2.781', 'grad_norm': '0.6914', 'learning_rate': '0.001', 'epoch': '0.4205', 'train/total_time_seconds': '247.5', 'train/time_per_step_avg': '0.3159', 'train/epoch_time_elapsed': '684.5', 'train/estimated_remaining_minutes': '1.164'}
96
+ {'loss': '2.782', 'grad_norm': '0.7383', 'learning_rate': '0.001', 'epoch': '0.4313', 'train/total_time_seconds': '253.8', 'train/time_per_step_avg': '0.3158', 'train/epoch_time_elapsed': '698.7', 'train/estimated_remaining_minutes': '1.058'}
97
+ {'eval_loss': '2.778', 'eval_runtime': '9.235', 'eval_samples_per_second': '1032', 'eval_steps_per_second': '2.057', 'epoch': '0.4313', 'train/total_time_seconds': '253.8', 'train/time_per_step_avg': '0.3158', 'train/epoch_time_elapsed': '707.9', 'train/estimated_remaining_minutes': '1.058'}
98
+ - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
99
+ - If you are not the owner of the model architecture class, please contact the model code owner to update it.
100
+ Writing model shards: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 1/1 [00:00<00:00, 82.91it/s]
101
+ 90%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ | 900/1000 [13:18<01:12, 1.38it/s][transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From πŸ‘‰v4.50πŸ‘ˆ onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
102
+ {'loss': '2.755', 'grad_norm': '0.4492', 'learning_rate': '0.001', 'epoch': '0.442', 'train/total_time_seconds': '260.1', 'train/time_per_step_avg': '0.3157', 'train/epoch_time_elapsed': '722.1', 'train/estimated_remaining_minutes': '0.9518'}
103
+ {'loss': '2.739', 'grad_norm': '0.8945', 'learning_rate': '0.001', 'epoch': '0.4528', 'train/total_time_seconds': '266.5', 'train/time_per_step_avg': '0.3157', 'train/epoch_time_elapsed': '736.1', 'train/estimated_remaining_minutes': '0.8459'}
104
+ - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
105
+ {'eval_loss': '2.736', 'eval_runtime': '9.211', 'eval_samples_per_second': '1034', 'eval_steps_per_second': '2.063', 'epoch': '0.4582', 'train/total_time_seconds': '269.6', 'train/time_per_step_avg': '0.3157', 'train/epoch_time_elapsed': '752.3', 'train/estimated_remaining_minutes': '0.793'}
106
+ {'loss': '2.729', 'grad_norm': '0.5352', 'learning_rate': '0.001', 'epoch': '0.4636', 'train/total_time_seconds': '272.8', 'train/time_per_step_avg': '0.3158', 'train/epoch_time_elapsed': '759.5', 'train/estimated_remaining_minutes': '0.7401'}
107
+ {'loss': '2.704', 'grad_norm': '0.9883', 'learning_rate': '0.001', 'epoch': '0.4744', 'train/total_time_seconds': '279.1', 'train/time_per_step_avg': '0.3159', 'train/epoch_time_elapsed': '773.8', 'train/estimated_remaining_minutes': '0.6344'}
108
+ {'loss': '2.712', 'grad_norm': '0.7539', 'learning_rate': '0.001', 'epoch': '0.4852', 'train/total_time_seconds': '285.4', 'train/time_per_step_avg': '0.3161', 'train/epoch_time_elapsed': '788.2', 'train/estimated_remaining_minutes': '0.5286'}
109
+ {'eval_loss': '2.704', 'eval_runtime': '9.472', 'eval_samples_per_second': '1006', 'eval_steps_per_second': '2.006', 'epoch': '0.4852', 'train/total_time_seconds': '285.4', 'train/time_per_step_avg': '0.3161', 'train/epoch_time_elapsed': '797.7', 'train/estimated_remaining_minutes': '0.5286'}
110
+ - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
111
+ - If you are not the owner of the model architecture class, please contact the model code owner to update it.
112
+ Writing model shards: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 1/1 [00:00<00:00, 86.80it/s]
113
+ 97%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–‰ | 974/1000 [14:20<00:18, 1.40it/s]
114
+ {'loss': '2.677', 'grad_norm': '0.4883', 'learning_rate': '0.001', 'epoch': '0.496', 'train/total_time_seconds': '291.8', 'train/time_per_step_avg': '0.3162', 'train/epoch_time_elapsed': '812.1', 'train/estimated_remaining_minutes': '0.4228'}
115
+ {'loss': '2.672', 'grad_norm': '0.9922', 'learning_rate': '0.001', 'epoch': '0.5067', 'train/total_time_seconds': '298.1', 'train/time_per_step_avg': '0.3163', 'train/epoch_time_elapsed': '826.4', 'train/estimated_remaining_minutes': '0.3171'}
116
+
117
+ {'eval_loss': '2.671', 'eval_runtime': '9.52', 'eval_samples_per_second': '1001', 'eval_steps_per_second': '1.996', 'epoch': '0.5121', 'train/total_time_seconds': '301.3', 'train/time_per_step_avg': '0.3163', 'train/epoch_time_elapsed': '843.1', 'train/estimated_remaining_minutes': '0.2643'}
118
+ {'loss': '2.673', 'grad_norm': '0.4824', 'learning_rate': '0.001', 'epoch': '0.5175', 'train/total_time_seconds': '304.4', 'train/time_per_step_avg': '0.3163', 'train/epoch_time_elapsed': '850.2', 'train/estimated_remaining_minutes': '0.2114'}
wandb/run-20260807_215701-rojwnt3e/logs/debug-internal.log CHANGED
@@ -83,3 +83,43 @@
83
  {"time":"2026-08-07T22:06:03.437526634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
84
  {"time":"2026-08-07T22:06:18.191505048Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":42,"history_lines":1,"events_offset":72,"events_lines":2,"console_offset":66,"console_lines":12}
85
  {"time":"2026-08-07T22:06:18.464200347Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83
  {"time":"2026-08-07T22:06:03.437526634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
84
  {"time":"2026-08-07T22:06:18.191505048Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":42,"history_lines":1,"events_offset":72,"events_lines":2,"console_offset":66,"console_lines":12}
85
  {"time":"2026-08-07T22:06:18.464200347Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
86
+ {"time":"2026-08-07T22:06:33.193453422Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":43,"history_lines":1,"events_offset":74,"events_lines":2,"console_offset":76,"console_lines":1}
87
+ {"time":"2026-08-07T22:06:33.498831625Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
88
+ {"time":"2026-08-07T22:06:48.191104459Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":44,"history_lines":2,"events_offset":76,"events_lines":2,"console_offset":76,"console_lines":1}
89
+ {"time":"2026-08-07T22:06:48.459530923Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
90
+ {"time":"2026-08-07T22:07:03.191131801Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":46,"history_lines":1,"events_offset":78,"events_lines":2,"console_offset":76,"console_lines":1}
91
+ {"time":"2026-08-07T22:07:03.469479856Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
92
+ {"time":"2026-08-07T22:07:18.193514266Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":47,"history_lines":1,"events_offset":80,"events_lines":2,"console_offset":76,"console_lines":1}
93
+ {"time":"2026-08-07T22:07:18.479043758Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
94
+ {"time":"2026-08-07T22:07:33.190853404Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":48,"history_lines":1,"events_offset":82,"events_lines":2,"console_offset":76,"console_lines":1}
95
+ {"time":"2026-08-07T22:07:33.436649349Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
96
+ {"time":"2026-08-07T22:07:48.192527365Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":49,"history_lines":1,"events_offset":84,"events_lines":2,"console_offset":78,"console_lines":12}
97
+ {"time":"2026-08-07T22:07:48.506679473Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
98
+ {"time":"2026-08-07T22:08:03.192547352Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":50,"history_lines":1,"events_offset":86,"events_lines":2,"console_offset":88,"console_lines":1}
99
+ {"time":"2026-08-07T22:08:03.451276923Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
100
+ {"time":"2026-08-07T22:08:18.191893467Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":51,"history_lines":2,"events_offset":88,"events_lines":2,"console_offset":88,"console_lines":1}
101
+ {"time":"2026-08-07T22:08:18.419712526Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
102
+ {"time":"2026-08-07T22:08:33.190848976Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":53,"history_lines":1,"events_offset":90,"events_lines":2,"console_offset":88,"console_lines":1}
103
+ {"time":"2026-08-07T22:08:33.447595074Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
104
+ {"time":"2026-08-07T22:08:48.193871302Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":54,"history_lines":1,"events_offset":92,"events_lines":2,"console_offset":88,"console_lines":1}
105
+ {"time":"2026-08-07T22:08:48.456096357Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
106
+ {"time":"2026-08-07T22:09:03.190928744Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":55,"history_lines":1,"events_offset":94,"events_lines":2,"console_offset":88,"console_lines":1}
107
+ {"time":"2026-08-07T22:09:03.484898063Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
108
+ {"time":"2026-08-07T22:09:18.191470424Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":56,"history_lines":1,"events_offset":96,"events_lines":2,"console_offset":90,"console_lines":12}
109
+ {"time":"2026-08-07T22:09:18.463857634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
110
+ {"time":"2026-08-07T22:09:33.191222201Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":57,"history_lines":1,"events_offset":98,"events_lines":2,"console_offset":100,"console_lines":1}
111
+ {"time":"2026-08-07T22:09:33.473907457Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
112
+ {"time":"2026-08-07T22:09:48.190733549Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":58,"history_lines":2,"events_offset":100,"events_lines":2,"console_offset":100,"console_lines":1}
113
+ {"time":"2026-08-07T22:09:48.427605817Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
114
+ {"time":"2026-08-07T22:10:03.190514166Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":60,"history_lines":1,"events_offset":102,"events_lines":2,"console_offset":100,"console_lines":1}
115
+ {"time":"2026-08-07T22:10:03.466281777Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
116
+ {"time":"2026-08-07T22:10:18.194146776Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":61,"history_lines":1,"events_offset":104,"events_lines":2,"console_offset":100,"console_lines":1}
117
+ {"time":"2026-08-07T22:10:18.482288854Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
118
+ {"time":"2026-08-07T22:10:33.190161734Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":62,"history_lines":1,"events_offset":106,"events_lines":2,"console_offset":100,"console_lines":1}
119
+ {"time":"2026-08-07T22:10:33.436720492Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
120
+ {"time":"2026-08-07T22:10:48.191398564Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":63,"history_lines":1,"events_offset":108,"events_lines":2,"console_offset":102,"console_lines":12}
121
+ {"time":"2026-08-07T22:10:48.472794309Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
122
+ {"time":"2026-08-07T22:11:03.190446716Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":64,"history_lines":1,"events_offset":110,"events_lines":2,"console_offset":112,"console_lines":1}
123
+ {"time":"2026-08-07T22:11:03.461724204Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
124
+ {"time":"2026-08-07T22:11:18.191588934Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":65,"history_lines":2,"events_offset":112,"events_lines":2,"console_offset":112,"console_lines":1}
125
+ {"time":"2026-08-07T22:11:18.454035207Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
wandb/run-20260807_215701-rojwnt3e/run-rojwnt3e.wandb CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a3442a24464c5d1c70615e0b690569e4b946cc8d1cd9f1f579d0c326cb11cbbb
3
- size 1835008
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebe71d2e699a73411f784457c3cd36b415b9200c66b1c7672a1763b52f6900f1
3
+ size 2129920