Auto upload zain 2026-08-20T23:24:20.784396 (part 3)

.gitattributes

@@ -57,3 +57,4 @@ zain/Activation/wandb/run-20260820_225644-vip8dia7/run-vip8dia7.wandb filter=lfs
 zain/Activation/wandb/run-20260820_230133-asswl5zl/run-asswl5zl.wandb filter=lfs diff=lfs merge=lfs -text
 zain/Activation/wandb/run-20260820_231106-yher6h2v/run-yher6h2v.wandb filter=lfs diff=lfs merge=lfs -text
 zain/Activation/wandb/run-20260820_231613-z5ghx5k4/run-z5ghx5k4.wandb filter=lfs diff=lfs merge=lfs -text
+zain/Activation/wandb/run-20260820_232141-62pb25xf/run-62pb25xf.wandb filter=lfs diff=lfs merge=lfs -text

zain/Activation/wandb/debug-internal.log

@@ -1,33 +1,29 @@
-{"time":"2026-08-20T23:16:13.99058953Z","level":"INFO","msg":"wandb-core"}
-{"time":"2026-08-20T23:16:13.990725942Z","level":"INFO","msg":"stream: starting","core version":"0.28.1"}
-{"time":"2026-08-20T23:16:14.255873976Z","level":"INFO","msg":"stream: created new stream","id":"z5ghx5k4"}
-{"time":"2026-08-20T23:16:14.261799372Z","level":"INFO","msg":"handler: started"}
-{"time":"2026-08-20T23:16:14.261923191Z","level":"INFO","msg":"stream: started"}
-{"time":"2026-08-20T23:16:14.2619377Z","level":"INFO","msg":"writer: started","stream_id":"z5ghx5k4"}
-{"time":"2026-08-20T23:16:14.261961669Z","level":"INFO","msg":"sender: started"}
-{"time":"2026-08-20T23:16:14.746596925Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1}
-{"time":"2026-08-20T23:16:14.928999337Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:16:29.746985888Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":0,"history_lines":8,"events_offset":0,"events_lines":1,"console_offset":0,"console_lines":14,"uploaded_len":2}
-{"time":"2026-08-20T23:16:29.921893424Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:16:44.746982989Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":8,"history_lines":8,"events_offset":1,"events_lines":2,"console_offset":11,"console_lines":1}
-{"time":"2026-08-20T23:16:45.10694945Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:16:59.747465601Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":16,"history_lines":7,"events_offset":3,"events_lines":2,"console_offset":14,"console_lines":25}
-{"time":"2026-08-20T23:16:59.97339168Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:17:14.747527213Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":23,"history_lines":6,"events_offset":5,"events_lines":2,"console_offset":33,"console_lines":1}
-{"time":"2026-08-20T23:17:14.954939843Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:17:29.747010635Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":29,"history_lines":6,"events_offset":7,"events_lines":2,"console_offset":39,"console_lines":23}
-{"time":"2026-08-20T23:17:30.256544005Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:17:44.746907448Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":35,"history_lines":7,"events_offset":9,"events_lines":2,"console_offset":55,"console_lines":1}
-{"time":"2026-08-20T23:17:45.121800138Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:17:59.746986087Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":42,"history_lines":9,"events_offset":11,"events_lines":2,"console_offset":61,"console_lines":31}
-{"time":"2026-08-20T23:18:00.082814389Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:18:14.747210463Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":51,"history_lines":8,"events_offset":13,"events_lines":2,"console_offset":88,"console_lines":1}
-{"time":"2026-08-20T23:18:15.168465067Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:18:29.747619692Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":59,"history_lines":6,"events_offset":15,"events_lines":2,"console_offset":92,"console_lines":25}
-{"time":"2026-08-20T23:18:30.25814208Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:18:44.747354408Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":65,"history_lines":6,"events_offset":17,"events_lines":2,"console_offset":110,"console_lines":1}
-{"time":"2026-08-20T23:18:45.176585339Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:18:59.747065177Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":71,"history_lines":6,"events_offset":19,"events_lines":2,"console_offset":116,"console_lines":23}
-{"time":"2026-08-20T23:18:59.964774028Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
-{"time":"2026-08-20T23:19:14.747180756Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":77,"history_lines":8,"events_offset":21,"events_lines":2,"console_offset":132,"console_lines":1}
-{"time":"2026-08-20T23:19:15.17955498Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:21:41.84229624Z","level":"INFO","msg":"wandb-core"}
+{"time":"2026-08-20T23:21:41.842431914Z","level":"INFO","msg":"stream: starting","core version":"0.28.1"}
+{"time":"2026-08-20T23:21:42.103652466Z","level":"INFO","msg":"stream: created new stream","id":"62pb25xf"}
+{"time":"2026-08-20T23:21:42.103749367Z","level":"INFO","msg":"handler: started"}
+{"time":"2026-08-20T23:21:42.103842743Z","level":"INFO","msg":"stream: started"}
+{"time":"2026-08-20T23:21:42.103855547Z","level":"INFO","msg":"writer: started","stream_id":"62pb25xf"}
+{"time":"2026-08-20T23:21:42.10388216Z","level":"INFO","msg":"sender: started"}
+{"time":"2026-08-20T23:21:42.612690788Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1}
+{"time":"2026-08-20T23:21:42.737898167Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:21:57.612950032Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":0,"history_lines":8,"events_offset":0,"events_lines":1,"console_offset":0,"console_lines":14,"uploaded_len":2}
+{"time":"2026-08-20T23:21:57.820912206Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:22:12.613688Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":8,"history_lines":9,"events_offset":1,"events_lines":2,"console_offset":11,"console_lines":1}
+{"time":"2026-08-20T23:22:13.235818035Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:22:27.612957229Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":17,"history_lines":6,"events_offset":3,"events_lines":2,"console_offset":14,"console_lines":26}
+{"time":"2026-08-20T23:22:28.0548937Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:22:42.613459262Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":23,"history_lines":6,"events_offset":5,"events_lines":2,"console_offset":33,"console_lines":1}
+{"time":"2026-08-20T23:22:43.163167398Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:22:57.613501341Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":29,"history_lines":6,"events_offset":7,"events_lines":2,"console_offset":39,"console_lines":23}
+{"time":"2026-08-20T23:22:57.787457711Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:23:12.614756489Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":35,"history_lines":9,"events_offset":9,"events_lines":2,"console_offset":55,"console_lines":1}
+{"time":"2026-08-20T23:23:12.910989461Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:23:27.613285577Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":44,"history_lines":8,"events_offset":11,"events_lines":2,"console_offset":61,"console_lines":32}
+{"time":"2026-08-20T23:23:27.704621158Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:23:42.61384811Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":52,"history_lines":7,"events_offset":13,"events_lines":2,"console_offset":88,"console_lines":1}
+{"time":"2026-08-20T23:23:43.160977529Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:23:57.613567795Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":59,"history_lines":6,"events_offset":15,"events_lines":2,"console_offset":93,"console_lines":24}
+{"time":"2026-08-20T23:23:57.908977513Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:24:12.613516868Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":65,"history_lines":6,"events_offset":17,"events_lines":2,"console_offset":110,"console_lines":1}
+{"time":"2026-08-20T23:24:12.938411406Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}

zain/Activation/wandb/debug.log

@@ -1,20 +1,20 @@
-2026-08-20 23:16:13,989 INFO    MainThread:58046 [wandb_init.py:setup_run_log_directory():729] Logging user logs to /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/wandb/run-20260820_231613-z5ghx5k4/logs/debug.log
-2026-08-20 23:16:13,989 INFO    MainThread:58046 [wandb_init.py:setup_run_log_directory():730] Logging internal logs to /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/wandb/run-20260820_231613-z5ghx5k4/logs/debug-internal.log
-2026-08-20 23:16:13,989 INFO    MainThread:58046 [wandb_init.py:init():772] calling init triggers
-2026-08-20 23:16:13,989 INFO    MainThread:58046 [wandb_init.py:init():777] wandb.init called with sweep_config: {}
+2026-08-20 23:21:41,840 INFO    MainThread:58046 [wandb_init.py:setup_run_log_directory():729] Logging user logs to /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/wandb/run-20260820_232141-62pb25xf/logs/debug.log
+2026-08-20 23:21:41,840 INFO    MainThread:58046 [wandb_init.py:setup_run_log_directory():730] Logging internal logs to /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/wandb/run-20260820_232141-62pb25xf/logs/debug-internal.log
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():772] calling init triggers
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():777] wandb.init called with sweep_config: {}
 config: {'_wandb': {}}
-2026-08-20 23:16:13,989 INFO    MainThread:58046 [wandb_init.py:init():820] starting backend
-2026-08-20 23:16:13,989 INFO    MainThread:58046 [wandb_init.py:init():826] Connected to an existing wandb-core service via WANDB_SERVICE
-2026-08-20 23:16:13,989 INFO    MainThread:58046 [wandb_init.py:init():835] sending inform_init request
-2026-08-20 23:16:14,262 INFO    MainThread:58046 [wandb_init.py:init():840] backend started and connected
-2026-08-20 23:16:14,263 INFO    MainThread:58046 [wandb_init.py:init():910] updated telemetry
-2026-08-20 23:16:14,270 INFO    MainThread:58046 [wandb_init.py:init():933] communicating run to backend with 90.0 second timeout
-2026-08-20 23:16:14,664 INFO    MainThread:58046 [wandb_init.py:init():978] starting run threads in backend
-2026-08-20 23:16:14,741 INFO    MainThread:58046 [wandb_run.py:_console_start():2621] atexit reg
-2026-08-20 23:16:14,741 INFO    MainThread:58046 [wandb_run.py:_redirect():2471] redirect: wrap_raw
-2026-08-20 23:16:14,741 INFO    MainThread:58046 [wandb_run.py:_redirect():2540] Wrapping output streams.
-2026-08-20 23:16:14,742 INFO    MainThread:58046 [wandb_run.py:_redirect():2563] Redirects installed.
-2026-08-20 23:16:14,742 INFO    MainThread:58046 [wandb_init.py:init():1016] run started, returning control to user process
-2026-08-20 23:16:14,743 INFO    MainThread:58046 [wandb_run.py:_config_callback():1346] config_cb None None {'transformers_version': '5.16.0.dev0', 'architectures': None, 'output_hidden_states': False, 'return_dict': True, 'dtype': None, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'problem_type': None, 'vocab_size': 4096, 'hidden_size': 128, 'intermediate_size': 256, 'num_hidden_layers': 9, 'num_attention_heads': 4, 'num_key_value_heads': 4, 'hidden_act': 'silu', 'max_position_embeddings': 512, 'initializer_range': 0.02, 'rms_norm_eps': 1e-06, 'use_cache': False, 'pad_token_id': 0, 'bos_token_id': 1, 'eos_token_id': 2, 'pretraining_tp': 1, 'tie_word_embeddings': True, 'rope_parameters': {'rope_theta': 10000.0, 'rope_type': 'default'}, 'attention_bias': False, 'attention_dropout': 0.0, 'mlp_bias': False, 'head_dim': 32, '_name_or_path': '', 'tokenizer_name': 'w-ahmad/tiny-stories-tokenizer', 'mlp_type': 'mlp', 'activation': 'quadratic', 'waleed_beta': 10.0, 'powlu_m': 3.0, 'model_type': 'tiny_llama', 'output_attentions': False, 'output_dir': 'out/mlp-quadratic-9L_run', 'per_device_train_batch_size': 32, 'num_train_epochs': 1, 'max_steps': 2500, 'learning_rate': 0.0003, 'lr_scheduler_type': 'constant_with_warmup', 'lr_scheduler_kwargs': None, 'warmup_steps': 500, 'optim': 'adamw_torch_fused', 'optim_args': None, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'optim_target_modules': None, 'gradient_accumulation_steps': 1, 'average_tokens_across_devices': True, 'max_grad_norm': 1.0, 'label_smoothing_factor': 0.0, 'bf16': True, 'fp16': False, 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'use_liger_kernel': False, 'liger_kernel_config': None, 'neftune_noise_alpha': None, 'torch_empty_cache_steps': None, 'auto_find_batch_size': False, 'logging_strategy': 'steps', 'logging_steps': 20, 'logging_first_step': False, 'log_on_each_node': True, 'logging_nan_inf_filter': True, 'include_num_input_tokens_seen': 'no', 'log_level': 'passive', 'log_level_replica': 'warning', 'disable_tqdm': False, 'report_to': ['wandb'], 'run_name': 'LM-mlp-quadratic-9L-2.0M-20260820-231613', 'project': 'huggingface', 'trackio_space_id': None, 'trackio_bucket_id': None, 'trackio_static_space_id': None, 'eval_strategy': 'steps', 'eval_steps': 100, 'eval_delay': 0, 'per_device_eval_batch_size': 1500, 'prediction_loss_only': False, 'eval_on_start': False, 'eval_do_concat_batches': True, 'eval_use_gather_object': False, 'eval_accumulation_steps': None, 'include_for_metrics': [], 'batch_eval_metrics': False, 'save_only_model': False, 'save_strategy': 'steps', 'save_steps': 100, 'save_on_each_node': False, 'save_total_limit': None, 'enable_jit_checkpoint': False, 'push_to_hub': False, 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': None, 'hub_model_id': 'w-ahmad/6L-mlp-quadratic-9L', 'hub_strategy': 'every_save', 'hub_always_push': False, 'hub_revision': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'restore_callback_states_from_checkpoint': False, 'full_determinism': False, 'seed': 42, 'data_seed': 42, 'use_cpu': False, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}, 'parallelism_config': None, 'dataloader_drop_last': False, 'dataloader_num_workers': 0, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'dataloader_prefetch_factor': None, 'dataloader_multiprocessing_context': None, 'dataloader_in_order': True, 'remove_unused_columns': False, 'label_names': None, 'train_sampling_strategy': 'random', 'length_column_name': 'length', 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'ddp_static_graph': None, 'ddp_backend': None, 'ddp_timeout': 1800, 'fsdp': None, 'fsdp_config': None, 'deepspeed': None, 'debug': [], 'skip_memory_metrics': True, 'do_train': False, 'do_eval': True, 'do_predict': False, 'resume_from_checkpoint': None, 'local_rank': -1}
-2026-08-20 23:16:14,745 INFO    MainThread:58046 [wandb_config.py:__setitem__():155] [no run ID] config set model/num_parameters = 2001280 - <bound method Run._config_callback of <wandb.sdk.wandb_run.Run object at 0x1526c0593f10>>
-2026-08-20 23:16:14,745 INFO    MainThread:58046 [wandb_run.py:_config_callback():1346] config_cb model/num_parameters 2001280 None
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():820] starting backend
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():826] Connected to an existing wandb-core service via WANDB_SERVICE
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():835] sending inform_init request
+2026-08-20 23:21:42,104 INFO    MainThread:58046 [wandb_init.py:init():840] backend started and connected
+2026-08-20 23:21:42,105 INFO    MainThread:58046 [wandb_init.py:init():910] updated telemetry
+2026-08-20 23:21:42,112 INFO    MainThread:58046 [wandb_init.py:init():933] communicating run to backend with 90.0 second timeout
+2026-08-20 23:21:42,530 INFO    MainThread:58046 [wandb_init.py:init():978] starting run threads in backend
+2026-08-20 23:21:42,607 INFO    MainThread:58046 [wandb_run.py:_console_start():2621] atexit reg
+2026-08-20 23:21:42,607 INFO    MainThread:58046 [wandb_run.py:_redirect():2471] redirect: wrap_raw
+2026-08-20 23:21:42,607 INFO    MainThread:58046 [wandb_run.py:_redirect():2540] Wrapping output streams.
+2026-08-20 23:21:42,607 INFO    MainThread:58046 [wandb_run.py:_redirect():2563] Redirects installed.
+2026-08-20 23:21:42,608 INFO    MainThread:58046 [wandb_init.py:init():1016] run started, returning control to user process
+2026-08-20 23:21:42,609 INFO    MainThread:58046 [wandb_run.py:_config_callback():1346] config_cb None None {'transformers_version': '5.16.0.dev0', 'architectures': None, 'output_hidden_states': False, 'return_dict': True, 'dtype': None, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'problem_type': None, 'vocab_size': 4096, 'hidden_size': 128, 'intermediate_size': 256, 'num_hidden_layers': 9, 'num_attention_heads': 4, 'num_key_value_heads': 4, 'hidden_act': 'silu', 'max_position_embeddings': 512, 'initializer_range': 0.02, 'rms_norm_eps': 1e-06, 'use_cache': False, 'pad_token_id': 0, 'bos_token_id': 1, 'eos_token_id': 2, 'pretraining_tp': 1, 'tie_word_embeddings': True, 'rope_parameters': {'rope_theta': 10000.0, 'rope_type': 'default'}, 'attention_bias': False, 'attention_dropout': 0.0, 'mlp_bias': False, 'head_dim': 32, '_name_or_path': '', 'tokenizer_name': 'w-ahmad/tiny-stories-tokenizer', 'mlp_type': 'mlp', 'activation': 'squared_relu', 'waleed_beta': 10.0, 'powlu_m': 3.0, 'model_type': 'tiny_llama', 'output_attentions': False, 'output_dir': 'out/mlp-squared_relu-9L_run', 'per_device_train_batch_size': 32, 'num_train_epochs': 1, 'max_steps': 2500, 'learning_rate': 0.0003, 'lr_scheduler_type': 'constant_with_warmup', 'lr_scheduler_kwargs': None, 'warmup_steps': 500, 'optim': 'adamw_torch_fused', 'optim_args': None, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'optim_target_modules': None, 'gradient_accumulation_steps': 1, 'average_tokens_across_devices': True, 'max_grad_norm': 1.0, 'label_smoothing_factor': 0.0, 'bf16': True, 'fp16': False, 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'use_liger_kernel': False, 'liger_kernel_config': None, 'neftune_noise_alpha': None, 'torch_empty_cache_steps': None, 'auto_find_batch_size': False, 'logging_strategy': 'steps', 'logging_steps': 20, 'logging_first_step': False, 'log_on_each_node': True, 'logging_nan_inf_filter': True, 'include_num_input_tokens_seen': 'no', 'log_level': 'passive', 'log_level_replica': 'warning', 'disable_tqdm': False, 'report_to': ['wandb'], 'run_name': 'LM-mlp-squared_relu-9L-2.0M-20260820-232141', 'project': 'huggingface', 'trackio_space_id': None, 'trackio_bucket_id': None, 'trackio_static_space_id': None, 'eval_strategy': 'steps', 'eval_steps': 100, 'eval_delay': 0, 'per_device_eval_batch_size': 1500, 'prediction_loss_only': False, 'eval_on_start': False, 'eval_do_concat_batches': True, 'eval_use_gather_object': False, 'eval_accumulation_steps': None, 'include_for_metrics': [], 'batch_eval_metrics': False, 'save_only_model': False, 'save_strategy': 'steps', 'save_steps': 100, 'save_on_each_node': False, 'save_total_limit': None, 'enable_jit_checkpoint': False, 'push_to_hub': False, 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': None, 'hub_model_id': 'w-ahmad/6L-mlp-squared_relu-9L', 'hub_strategy': 'every_save', 'hub_always_push': False, 'hub_revision': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'restore_callback_states_from_checkpoint': False, 'full_determinism': False, 'seed': 42, 'data_seed': 42, 'use_cpu': False, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}, 'parallelism_config': None, 'dataloader_drop_last': False, 'dataloader_num_workers': 0, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'dataloader_prefetch_factor': None, 'dataloader_multiprocessing_context': None, 'dataloader_in_order': True, 'remove_unused_columns': False, 'label_names': None, 'train_sampling_strategy': 'random', 'length_column_name': 'length', 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'ddp_static_graph': None, 'ddp_backend': None, 'ddp_timeout': 1800, 'fsdp': None, 'fsdp_config': None, 'deepspeed': None, 'debug': [], 'skip_memory_metrics': True, 'do_train': False, 'do_eval': True, 'do_predict': False, 'resume_from_checkpoint': None, 'local_rank': -1}
+2026-08-20 23:21:42,611 INFO    MainThread:58046 [wandb_config.py:__setitem__():155] [no run ID] config set model/num_parameters = 2001280 - <bound method Run._config_callback of <wandb.sdk.wandb_run.Run object at 0x1526c05aeb50>>
+2026-08-20 23:21:42,611 INFO    MainThread:58046 [wandb_run.py:_config_callback():1346] config_cb model/num_parameters 2001280 None

zain/Activation/wandb/run-20260820_224446-wyqgp22a/logs/debug-core.log

@@ -63,3 +63,10 @@
 {"time":"2026-08-20T23:16:13.990443666Z","level":"INFO","msg":"handleInformInit: received","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:14.261935343Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:20.112222253Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:40.403706695Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.099860028Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.101376929Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.10192683Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.842145532Z","level":"INFO","msg":"handleInformInit: received","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:42.103852099Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:47.827855381Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"xrlr57kv5zv0"}

zain/Activation/wandb/run-20260820_224528-m7e0j3tz/logs/debug-core.log

@@ -63,3 +63,10 @@
 {"time":"2026-08-20T23:16:13.990443666Z","level":"INFO","msg":"handleInformInit: received","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:14.261935343Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:20.112222253Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:40.403706695Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.099860028Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.101376929Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.10192683Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.842145532Z","level":"INFO","msg":"handleInformInit: received","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:42.103852099Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:47.827855381Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"xrlr57kv5zv0"}

zain/Activation/wandb/run-20260820_225014-5ntg3lev/logs/debug-core.log

@@ -63,3 +63,10 @@
 {"time":"2026-08-20T23:16:13.990443666Z","level":"INFO","msg":"handleInformInit: received","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:14.261935343Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:20.112222253Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:40.403706695Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.099860028Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.101376929Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.10192683Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.842145532Z","level":"INFO","msg":"handleInformInit: received","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:42.103852099Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:47.827855381Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"xrlr57kv5zv0"}

zain/Activation/wandb/run-20260820_225644-vip8dia7/logs/debug-core.log

@@ -63,3 +63,10 @@
 {"time":"2026-08-20T23:16:13.990443666Z","level":"INFO","msg":"handleInformInit: received","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:14.261935343Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:20.112222253Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:40.403706695Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.099860028Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.101376929Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.10192683Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.842145532Z","level":"INFO","msg":"handleInformInit: received","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:42.103852099Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:47.827855381Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"xrlr57kv5zv0"}

zain/Activation/wandb/run-20260820_230133-asswl5zl/logs/debug-core.log

@@ -63,3 +63,10 @@
 {"time":"2026-08-20T23:16:13.990443666Z","level":"INFO","msg":"handleInformInit: received","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:14.261935343Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:20.112222253Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:40.403706695Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.099860028Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.101376929Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.10192683Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.842145532Z","level":"INFO","msg":"handleInformInit: received","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:42.103852099Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:47.827855381Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"xrlr57kv5zv0"}

zain/Activation/wandb/run-20260820_231106-yher6h2v/logs/debug-core.log

@@ -63,3 +63,10 @@
 {"time":"2026-08-20T23:16:13.990443666Z","level":"INFO","msg":"handleInformInit: received","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:14.261935343Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:20.112222253Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:40.403706695Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.099860028Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.101376929Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.10192683Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.842145532Z","level":"INFO","msg":"handleInformInit: received","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:42.103852099Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:47.827855381Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"xrlr57kv5zv0"}

zain/Activation/wandb/run-20260820_231613-z5ghx5k4/files/config.yaml

@@ -0,0 +1,437 @@
+_name_or_path:
+    value: ""
+_wandb:
+    value:
+        cli_version: 0.28.1
+        e:
+            p3hjdvywf27uoyhms00l2z3ut26knnfu:
+                args:
+                    - --config
+                    - /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/configs/baseline100L.yaml
+                    - --variants
+                    - mlp-squared_relu-6L
+                    - mlp-quadratic-9L
+                    - mlp-squared_relu-9L
+                codePath: sweep.py
+                codePathLocal: sweep.py
+                cpu_count: 112
+                cpu_count_logical: 224
+                cudaVersion: "12.4"
+                disk:
+                    /:
+                        total: "1560765693952"
+                        used: "707203641344"
+                email: deepnevro@gmail.com
+                executable: /mnt/data/zainulabideen/zain-exp/notebooks/my_env/bin/python
+                git:
+                    commit: 444fe7f656e69d3ec9e58941464da02fec8c6047
+                    remote: https://github.com/w-ahmad1a10/Activation.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-39c684a5-fde6-83d7-1663-0859795881ae
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-68012e5a-38b6-b643-0ca6-62fb66720bf3
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-132944c4-b689-2b5f-89a4-d730401677ab
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-2df386cc-6d26-d0e2-7a2d-a057b0d95864
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-bfa16575-1d94-1aa2-4537-2c93433f42ef
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-bc6c3e3c-9b90-09ca-c034-774961847c54
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-00a441e1-7c95-e7d6-4c35-43d6b291aea9
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-1c4d29a2-4647-6fce-d8fc-0c5ecfbbd6ea
+                host: deeplens-k3s-node1
+                memory:
+                    total: "2164089937920"
+                os: Linux-5.15.0-126-generic-x86_64-with-glibc2.35
+                program: /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/sweep.py
+                python: CPython 3.11.15
+                root: /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation
+                startedAt: "2026-08-20T23:16:13.988163Z"
+                writerId: p3hjdvywf27uoyhms00l2z3ut26knnfu
+        m:
+            - "1": train/global_step
+              "6":
+                - 3
+              "7": []
+            - "2": '*'
+              "5": 1
+              "6":
+                - 1
+              "7": []
+        python_version: 3.11.15
+        t:
+            "1":
+                - 1
+                - 5
+                - 11
+                - 41
+                - 49
+                - 51
+                - 53
+                - 71
+            "2":
+                - 1
+                - 5
+                - 11
+                - 41
+                - 49
+                - 51
+                - 53
+                - 71
+            "3":
+                - 2
+                - 7
+                - 13
+                - 19
+                - 62
+                - 66
+            "4": 3.11.15
+            "5": 0.28.1
+            "6": 5.16.0.dev0
+            "9":
+                "1": transformers_trainer
+            "12": 0.28.1
+            "13": linux-x86_64
+accelerator_config:
+    value:
+        dispatch_batches: null
+        even_batches: true
+        gradient_accumulation_kwargs: null
+        non_blocking: false
+        split_batches: false
+        use_seedable_sampler: true
+activation:
+    value: quadratic
+adam_beta1:
+    value: 0.9
+adam_beta2:
+    value: 0.999
+adam_epsilon:
+    value: 1e-08
+architectures:
+    value: null
+attention_bias:
+    value: false
+attention_dropout:
+    value: 0
+auto_find_batch_size:
+    value: false
+average_tokens_across_devices:
+    value: true
+batch_eval_metrics:
+    value: false
+bf16:
+    value: true
+bf16_full_eval:
+    value: false
+bos_token_id:
+    value: 1
+chunk_size_feed_forward:
+    value: 0
+data_seed:
+    value: 42
+dataloader_drop_last:
+    value: false
+dataloader_in_order:
+    value: true
+dataloader_multiprocessing_context:
+    value: null
+dataloader_num_workers:
+    value: 0
+dataloader_persistent_workers:
+    value: false
+dataloader_pin_memory:
+    value: true
+dataloader_prefetch_factor:
+    value: null
+ddp_backend:
+    value: null
+ddp_broadcast_buffers:
+    value: null
+ddp_bucket_cap_mb:
+    value: null
+ddp_find_unused_parameters:
+    value: null
+ddp_static_graph:
+    value: null
+ddp_timeout:
+    value: 1800
+debug:
+    value: []
+deepspeed:
+    value: null
+disable_tqdm:
+    value: false
+do_eval:
+    value: true
+do_predict:
+    value: false
+do_train:
+    value: false
+dtype:
+    value: null
+enable_jit_checkpoint:
+    value: false
+eos_token_id:
+    value: 2
+eval_accumulation_steps:
+    value: null
+eval_delay:
+    value: 0
+eval_do_concat_batches:
+    value: true
+eval_on_start:
+    value: false
+eval_steps:
+    value: 100
+eval_strategy:
+    value: steps
+eval_use_gather_object:
+    value: false
+fp16:
+    value: false
+fp16_full_eval:
+    value: false
+fsdp:
+    value: null
+fsdp_config:
+    value: null
+full_determinism:
+    value: false
+gradient_accumulation_steps:
+    value: 1
+gradient_checkpointing:
+    value: false
+gradient_checkpointing_kwargs:
+    value: null
+greater_is_better:
+    value: null
+head_dim:
+    value: 32
+hidden_act:
+    value: silu
+hidden_size:
+    value: 128
+hub_always_push:
+    value: false
+hub_model_id:
+    value: w-ahmad/6L-mlp-quadratic-9L
+hub_private_repo:
+    value: null
+hub_revision:
+    value: null
+hub_strategy:
+    value: every_save
+hub_token:
+    value: <HUB_TOKEN>
+id2label:
+    value:
+        "0": LABEL_0
+        "1": LABEL_1
+ignore_data_skip:
+    value: false
+include_for_metrics:
+    value: []
+include_num_input_tokens_seen:
+    value: "no"
+initializer_range:
+    value: 0.02
+intermediate_size:
+    value: 256
+is_encoder_decoder:
+    value: false
+label_names:
+    value: null
+label_smoothing_factor:
+    value: 0
+label2id:
+    value:
+        LABEL_0: 0
+        LABEL_1: 1
+learning_rate:
+    value: 0.0003
+length_column_name:
+    value: length
+liger_kernel_config:
+    value: null
+load_best_model_at_end:
+    value: false
+local_rank:
+    value: -1
+log_level:
+    value: passive
+log_level_replica:
+    value: warning
+log_on_each_node:
+    value: true
+logging_first_step:
+    value: false
+logging_nan_inf_filter:
+    value: true
+logging_steps:
+    value: 20
+logging_strategy:
+    value: steps
+lr_scheduler_kwargs:
+    value: null
+lr_scheduler_type:
+    value: constant_with_warmup
+max_grad_norm:
+    value: 1
+max_position_embeddings:
+    value: 512
+max_steps:
+    value: 2500
+metric_for_best_model:
+    value: null
+mlp_bias:
+    value: false
+mlp_type:
+    value: mlp
+model/num_parameters:
+    value: 2001280
+model_type:
+    value: tiny_llama
+neftune_noise_alpha:
+    value: null
+num_attention_heads:
+    value: 4
+num_hidden_layers:
+    value: 9
+num_key_value_heads:
+    value: 4
+num_train_epochs:
+    value: 1
+optim:
+    value: adamw_torch_fused
+optim_args:
+    value: null
+optim_target_modules:
+    value: null
+output_attentions:
+    value: false
+output_dir:
+    value: out/mlp-quadratic-9L_run
+output_hidden_states:
+    value: false
+pad_token_id:
+    value: 0
+parallelism_config:
+    value: null
+per_device_eval_batch_size:
+    value: 1500
+per_device_train_batch_size:
+    value: 32
+powlu_m:
+    value: 3
+prediction_loss_only:
+    value: false
+pretraining_tp:
+    value: 1
+problem_type:
+    value: null
+project:
+    value: huggingface
+push_to_hub:
+    value: false
+remove_unused_columns:
+    value: false
+report_to:
+    value:
+        - wandb
+restore_callback_states_from_checkpoint:
+    value: false
+resume_from_checkpoint:
+    value: null
+return_dict:
+    value: true
+rms_norm_eps:
+    value: 1e-06
+rope_parameters:
+    value:
+        rope_theta: 10000
+        rope_type: default
+run_name:
+    value: LM-mlp-quadratic-9L-2.0M-20260820-231613
+save_on_each_node:
+    value: false
+save_only_model:
+    value: false
+save_steps:
+    value: 100
+save_strategy:
+    value: steps
+save_total_limit:
+    value: null
+seed:
+    value: 42
+skip_memory_metrics:
+    value: true
+tf32:
+    value: null
+tie_word_embeddings:
+    value: true
+tokenizer_name:
+    value: w-ahmad/tiny-stories-tokenizer
+torch_compile:
+    value: false
+torch_compile_backend:
+    value: null
+torch_compile_mode:
+    value: null
+torch_empty_cache_steps:
+    value: null
+trackio_bucket_id:
+    value: null
+trackio_space_id:
+    value: null
+trackio_static_space_id:
+    value: null
+train_sampling_strategy:
+    value: random
+transformers_version:
+    value: 5.16.0.dev0
+use_cache:
+    value: false
+use_cpu:
+    value: false
+use_liger_kernel:
+    value: false
+vocab_size:
+    value: 4096
+waleed_beta:
+    value: 10
+warmup_steps:
+    value: 500
+weight_decay:
+    value: 0

zain/Activation/wandb/run-20260820_231613-z5ghx5k4/files/output.log

@@ -152,9 +152,133 @@ Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 167
   - 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).
   - If you are not the owner of the model architecture class, please contact the model code owner to update it.
 Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 64.66it/s]
- 60%|██████    | 1500/2500 [03:03<00:46, 21.40it/s], ?it/s]
+ 60%|██████    | 1500/2500 [03:11<00:46, 21.4[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
 {'loss': '3.311', 'grad_norm': '1.297', 'learning_rate': '0.0003', 'epoch': '0.04786', 'train/total_time_seconds': '31.39', 'train/time_per_step_avg': '0.02087', 'train/epoch_time_elapsed': '179.4', 'train/estimated_remaining_minutes': '0.3979'}
 {'loss': '3.296', 'grad_norm': '1.43', 'learning_rate': '0.0003', 'epoch': '0.04853', 'train/total_time_seconds': '31.85', 'train/time_per_step_avg': '0.02131', 'train/epoch_time_elapsed': '180.3', 'train/estimated_remaining_minutes': '0.3908'}
 {'loss': '3.273', 'grad_norm': '1.172', 'learning_rate': '0.0003', 'epoch': '0.04921', 'train/total_time_seconds': '32.28', 'train/time_per_step_avg': '0.02128', 'train/epoch_time_elapsed': '181.2', 'train/estimated_remaining_minutes': '0.3832'}
 {'loss': '3.298', 'grad_norm': '1.453', 'learning_rate': '0.0003', 'epoch': '0.04988', 'train/total_time_seconds': '32.7', 'train/time_per_step_avg': '0.02138', 'train/epoch_time_elapsed': '182.1', 'train/estimated_remaining_minutes': '0.3756'}
 {'loss': '3.293', 'grad_norm': '1.477', 'learning_rate': '0.0003', 'epoch': '0.05056', 'train/total_time_seconds': '33.12', 'train/time_per_step_avg': '0.02152', 'train/epoch_time_elapsed': '183.1', 'train/estimated_remaining_minutes': '0.368'}
+  - 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
+{'eval_loss': '3.288', 'eval_runtime': '8.043', 'eval_samples_per_second': '1185', 'eval_steps_per_second': '0.87', 'epoch': '0.05056', 'train/total_time_seconds': '33.12', 'train/time_per_step_avg': '0.02152', 'train/epoch_time_elapsed': '191.1', 'train/estimated_remaining_minutes': '0.368'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 143.74it/s]
+ 64%|██████▍   | 1600/2500 [03:23<00:39, 22.6[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.282', 'grad_norm': '1.25', 'learning_rate': '0.0003', 'epoch': '0.05123', 'train/total_time_seconds': '33.53', 'train/time_per_step_avg': '0.02142', 'train/epoch_time_elapsed': '192.1', 'train/estimated_remaining_minutes': '0.3603'}
+{'loss': '3.288', 'grad_norm': '1.312', 'learning_rate': '0.0003', 'epoch': '0.0519', 'train/total_time_seconds': '33.94', 'train/time_per_step_avg': '0.02088', 'train/epoch_time_elapsed': '193', 'train/estimated_remaining_minutes': '0.3526'}
+{'loss': '3.29', 'grad_norm': '1.344', 'learning_rate': '0.0003', 'epoch': '0.05258', 'train/total_time_seconds': '34.37', 'train/time_per_step_avg': '0.02097', 'train/epoch_time_elapsed': '193.9', 'train/estimated_remaining_minutes': '0.3452'}
+{'loss': '3.269', 'grad_norm': '1.484', 'learning_rate': '0.0003', 'epoch': '0.05325', 'train/total_time_seconds': '34.77', 'train/time_per_step_avg': '0.02078', 'train/epoch_time_elapsed': '194.8', 'train/estimated_remaining_minutes': '0.3375'}
+{'loss': '3.26', 'grad_norm': '1.266', 'learning_rate': '0.0003', 'epoch': '0.05393', 'train/total_time_seconds': '35.17', 'train/time_per_step_avg': '0.02051', 'train/epoch_time_elapsed': '195.7', 'train/estimated_remaining_minutes': '0.3297'}
+  - 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
+{'eval_loss': '3.256', 'eval_runtime': '8.032', 'eval_samples_per_second': '1186', 'eval_steps_per_second': '0.872', 'epoch': '0.05393', 'train/total_time_seconds': '35.17', 'train/time_per_step_avg': '0.02051', 'train/epoch_time_elapsed': '203.7', 'train/estimated_remaining_minutes': '0.3297'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 175.49it/s]
+ 68%|██████▊   | 1700/2500 [03:36<00:36, 22.1[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.28', 'grad_norm': '1.18', 'learning_rate': '0.0003', 'epoch': '0.0546', 'train/total_time_seconds': '35.57', 'train/time_per_step_avg': '0.02039', 'train/epoch_time_elapsed': '204.6', 'train/estimated_remaining_minutes': '0.3221'}
+{'loss': '3.256', 'grad_norm': '1.398', 'learning_rate': '0.0003', 'epoch': '0.05527', 'train/total_time_seconds': '35.98', 'train/time_per_step_avg': '0.02037', 'train/epoch_time_elapsed': '205.5', 'train/estimated_remaining_minutes': '0.3145'}
+{'loss': '3.264', 'grad_norm': '1.25', 'learning_rate': '0.0003', 'epoch': '0.05595', 'train/total_time_seconds': '36.39', 'train/time_per_step_avg': '0.02013', 'train/epoch_time_elapsed': '206.4', 'train/estimated_remaining_minutes': '0.3069'}
+{'loss': '3.24', 'grad_norm': '1.508', 'learning_rate': '0.0003', 'epoch': '0.05662', 'train/total_time_seconds': '36.79', 'train/time_per_step_avg': '0.02019', 'train/epoch_time_elapsed': '207.3', 'train/estimated_remaining_minutes': '0.2993'}
+{'loss': '3.227', 'grad_norm': '1.484', 'learning_rate': '0.0003', 'epoch': '0.0573', 'train/total_time_seconds': '37.2', 'train/time_per_step_avg': '0.02026', 'train/epoch_time_elapsed': '208.2', 'train/estimated_remaining_minutes': '0.2918'}
+  - 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
+{'eval_loss': '3.232', 'eval_runtime': '8.35', 'eval_samples_per_second': '1141', 'eval_steps_per_second': '0.838', 'epoch': '0.0573', 'train/total_time_seconds': '37.2', 'train/time_per_step_avg': '0.02026', 'train/epoch_time_elapsed': '216.6', 'train/estimated_remaining_minutes': '0.2918'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 176.40it/s]
+ 72%|███████▏  | 1800/2500 [03:49<00:33, 20.8[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.222', 'grad_norm': '1.367', 'learning_rate': '0.0003', 'epoch': '0.05797', 'train/total_time_seconds': '37.6', 'train/time_per_step_avg': '0.02028', 'train/epoch_time_elapsed': '217.5', 'train/estimated_remaining_minutes': '0.2842'}
+{'loss': '3.22', 'grad_norm': '1.344', 'learning_rate': '0.0003', 'epoch': '0.05865', 'train/total_time_seconds': '38', 'train/time_per_step_avg': '0.02022', 'train/epoch_time_elapsed': '218.4', 'train/estimated_remaining_minutes': '0.2766'}
+{'loss': '3.211', 'grad_norm': '1.352', 'learning_rate': '0.0003', 'epoch': '0.05932', 'train/total_time_seconds': '38.4', 'train/time_per_step_avg': '0.02017', 'train/epoch_time_elapsed': '219.3', 'train/estimated_remaining_minutes': '0.2691'}
+{'loss': '3.218', 'grad_norm': '1.477', 'learning_rate': '0.0003', 'epoch': '0.05999', 'train/total_time_seconds': '38.81', 'train/time_per_step_avg': '0.02012', 'train/epoch_time_elapsed': '220.2', 'train/estimated_remaining_minutes': '0.2616'}
+{'loss': '3.209', 'grad_norm': '1.32', 'learning_rate': '0.0003', 'epoch': '0.06067', 'train/total_time_seconds': '39.24', 'train/time_per_step_avg': '0.02037', 'train/epoch_time_elapsed': '221.1', 'train/estimated_remaining_minutes': '0.2543'}
+  - 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
+{'eval_loss': '3.209', 'eval_runtime': '7.973', 'eval_samples_per_second': '1195', 'eval_steps_per_second': '0.878', 'epoch': '0.06067', 'train/total_time_seconds': '39.24', 'train/time_per_step_avg': '0.02037', 'train/epoch_time_elapsed': '229.1', 'train/estimated_remaining_minutes': '0.2543'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 173.24it/s]
+ 76%|███████▌  | 1900/2500 [04:01<00:27, 21.6[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.197', 'grad_norm': '1.281', 'learning_rate': '0.0003', 'epoch': '0.06134', 'train/total_time_seconds': '39.64', 'train/time_per_step_avg': '0.02037', 'train/epoch_time_elapsed': '230', 'train/estimated_remaining_minutes': '0.2468'}
+{'loss': '3.197', 'grad_norm': '1.547', 'learning_rate': '0.0003', 'epoch': '0.06202', 'train/total_time_seconds': '40.04', 'train/time_per_step_avg': '0.02037', 'train/epoch_time_elapsed': '230.9', 'train/estimated_remaining_minutes': '0.2394'}
+{'loss': '3.208', 'grad_norm': '1.289', 'learning_rate': '0.0003', 'epoch': '0.06269', 'train/total_time_seconds': '40.45', 'train/time_per_step_avg': '0.02047', 'train/epoch_time_elapsed': '231.8', 'train/estimated_remaining_minutes': '0.232'}
+{'loss': '3.181', 'grad_norm': '1.344', 'learning_rate': '0.0003', 'epoch': '0.06336', 'train/total_time_seconds': '40.9', 'train/time_per_step_avg': '0.02096', 'train/epoch_time_elapsed': '232.8', 'train/estimated_remaining_minutes': '0.2248'}
+{'loss': '3.186', 'grad_norm': '1.305', 'learning_rate': '0.0003', 'epoch': '0.06404', 'train/total_time_seconds': '41.3', 'train/time_per_step_avg': '0.02066', 'train/epoch_time_elapsed': '233.7', 'train/estimated_remaining_minutes': '0.2174'}
+  - 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
+{'eval_loss': '3.191', 'eval_runtime': '7.907', 'eval_samples_per_second': '1205', 'eval_steps_per_second': '0.885', 'epoch': '0.06404', 'train/total_time_seconds': '41.3', 'train/time_per_step_avg': '0.02066', 'train/epoch_time_elapsed': '241.6', 'train/estimated_remaining_minutes': '0.2174'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 175.63it/s]
+ 80%|████████  | 2000/2500 [04:14<00:25, 19.9[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.174', 'grad_norm': '1.539', 'learning_rate': '0.0003', 'epoch': '0.06471', 'train/total_time_seconds': '41.71', 'train/time_per_step_avg': '0.02068', 'train/epoch_time_elapsed': '242.5', 'train/estimated_remaining_minutes': '0.21'}
+{'loss': '3.159', 'grad_norm': '1.227', 'learning_rate': '0.0003', 'epoch': '0.06539', 'train/total_time_seconds': '42.11', 'train/time_per_step_avg': '0.02071', 'train/epoch_time_elapsed': '243.4', 'train/estimated_remaining_minutes': '0.2026'}
+{'loss': '3.183', 'grad_norm': '1.43', 'learning_rate': '0.0003', 'epoch': '0.06606', 'train/total_time_seconds': '42.51', 'train/time_per_step_avg': '0.02056', 'train/epoch_time_elapsed': '244.3', 'train/estimated_remaining_minutes': '0.1952'}
+{'loss': '3.195', 'grad_norm': '1.266', 'learning_rate': '0.0003', 'epoch': '0.06673', 'train/total_time_seconds': '42.99', 'train/time_per_step_avg': '0.02084', 'train/epoch_time_elapsed': '245.4', 'train/estimated_remaining_minutes': '0.1882'}
+{'loss': '3.171', 'grad_norm': '1.578', 'learning_rate': '0.0003', 'epoch': '0.06741', 'train/total_time_seconds': '43.45', 'train/time_per_step_avg': '0.0215', 'train/epoch_time_elapsed': '246.5', 'train/estimated_remaining_minutes': '0.181'}
+  - 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
+{'eval_loss': '3.17', 'eval_runtime': '7.997', 'eval_samples_per_second': '1191', 'eval_steps_per_second': '0.875', 'epoch': '0.06741', 'train/total_time_seconds': '43.45', 'train/time_per_step_avg': '0.0215', 'train/epoch_time_elapsed': '254.5', 'train/estimated_remaining_minutes': '0.181'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 174.73it/s]
+ 84%|████████▍ | 2100/2500 [04:27<00:18, 21.7[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.16', 'grad_norm': '1.383', 'learning_rate': '0.0003', 'epoch': '0.06808', 'train/total_time_seconds': '43.85', 'train/time_per_step_avg': '0.02147', 'train/epoch_time_elapsed': '255.4', 'train/estimated_remaining_minutes': '0.1737'}
+{'loss': '3.161', 'grad_norm': '1.523', 'learning_rate': '0.0003', 'epoch': '0.06876', 'train/total_time_seconds': '44.26', 'train/time_per_step_avg': '0.02146', 'train/epoch_time_elapsed': '256.3', 'train/estimated_remaining_minutes': '0.1663'}
+{'loss': '3.15', 'grad_norm': '1.477', 'learning_rate': '0.0003', 'epoch': '0.06943', 'train/total_time_seconds': '44.66', 'train/time_per_step_avg': '0.02152', 'train/epoch_time_elapsed': '257.2', 'train/estimated_remaining_minutes': '0.159'}
+{'loss': '3.156', 'grad_norm': '1.461', 'learning_rate': '0.0003', 'epoch': '0.0701', 'train/total_time_seconds': '45.06', 'train/time_per_step_avg': '0.02071', 'train/epoch_time_elapsed': '258.1', 'train/estimated_remaining_minutes': '0.1516'}
+{'loss': '3.146', 'grad_norm': '1.305', 'learning_rate': '0.0003', 'epoch': '0.07078', 'train/total_time_seconds': '45.48', 'train/time_per_step_avg': '0.02027', 'train/epoch_time_elapsed': '259', 'train/estimated_remaining_minutes': '0.1444'}
+  - 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
+{'eval_loss': '3.147', 'eval_runtime': '8.008', 'eval_samples_per_second': '1190', 'eval_steps_per_second': '0.874', 'epoch': '0.07078', 'train/total_time_seconds': '45.48', 'train/time_per_step_avg': '0.02027', 'train/epoch_time_elapsed': '267', 'train/estimated_remaining_minutes': '0.1444'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 175.88it/s]
+ 88%|████████▊ | 2200/2500 [04:39<00:13, 22.6[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.146', 'grad_norm': '1.383', 'learning_rate': '0.0003', 'epoch': '0.07145', 'train/total_time_seconds': '45.88', 'train/time_per_step_avg': '0.02028', 'train/epoch_time_elapsed': '268', 'train/estimated_remaining_minutes': '0.1371'}
+{'loss': '3.148', 'grad_norm': '1.352', 'learning_rate': '0.0003', 'epoch': '0.07213', 'train/total_time_seconds': '46.28', 'train/time_per_step_avg': '0.02026', 'train/epoch_time_elapsed': '268.8', 'train/estimated_remaining_minutes': '0.1298'}
+{'loss': '3.132', 'grad_norm': '1.312', 'learning_rate': '0.0003', 'epoch': '0.0728', 'train/total_time_seconds': '46.72', 'train/time_per_step_avg': '0.02065', 'train/epoch_time_elapsed': '269.8', 'train/estimated_remaining_minutes': '0.1226'}
+{'loss': '3.14', 'grad_norm': '1.414', 'learning_rate': '0.0003', 'epoch': '0.07347', 'train/total_time_seconds': '47.18', 'train/time_per_step_avg': '0.02123', 'train/epoch_time_elapsed': '270.7', 'train/estimated_remaining_minutes': '0.1154'}
+{'loss': '3.139', 'grad_norm': '1.445', 'learning_rate': '0.0003', 'epoch': '0.07415', 'train/total_time_seconds': '47.57', 'train/time_per_step_avg': '0.02095', 'train/epoch_time_elapsed': '271.6', 'train/estimated_remaining_minutes': '0.1081'}
+  - 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
+{'eval_loss': '3.132', 'eval_runtime': '7.968', 'eval_samples_per_second': '1196', 'eval_steps_per_second': '0.879', 'epoch': '0.07415', 'train/total_time_seconds': '47.57', 'train/time_per_step_avg': '0.02095', 'train/epoch_time_elapsed': '279.6', 'train/estimated_remaining_minutes': '0.1081'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 106.40it/s]
+ 92%|█████████▏| 2300/2500 [04:52<00:09, 22.1[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.101', 'grad_norm': '1.367', 'learning_rate': '0.0003', 'epoch': '0.07482', 'train/total_time_seconds': '47.98', 'train/time_per_step_avg': '0.02099', 'train/epoch_time_elapsed': '280.5', 'train/estimated_remaining_minutes': '0.1009'}
+{'loss': '3.124', 'grad_norm': '1.305', 'learning_rate': '0.0003', 'epoch': '0.0755', 'train/total_time_seconds': '48.38', 'train/time_per_step_avg': '0.021', 'train/epoch_time_elapsed': '281.4', 'train/estimated_remaining_minutes': '0.09359'}
+{'loss': '3.097', 'grad_norm': '1.453', 'learning_rate': '0.0003', 'epoch': '0.07617', 'train/total_time_seconds': '48.8', 'train/time_per_step_avg': '0.02076', 'train/epoch_time_elapsed': '282.4', 'train/estimated_remaining_minutes': '0.08637'}
+{'loss': '3.118', 'grad_norm': '1.492', 'learning_rate': '0.0003', 'epoch': '0.07685', 'train/total_time_seconds': '49.22', 'train/time_per_step_avg': '0.02037', 'train/epoch_time_elapsed': '283.3', 'train/estimated_remaining_minutes': '0.07915'}
+{'loss': '3.112', 'grad_norm': '1.383', 'learning_rate': '0.0003', 'epoch': '0.07752', 'train/total_time_seconds': '49.62', 'train/time_per_step_avg': '0.02048', 'train/epoch_time_elapsed': '284.2', 'train/estimated_remaining_minutes': '0.07191'}
+  - 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
+{'eval_loss': '3.115', 'eval_runtime': '7.987', 'eval_samples_per_second': '1193', 'eval_steps_per_second': '0.876', 'epoch': '0.07752', 'train/total_time_seconds': '49.62', 'train/time_per_step_avg': '0.02048', 'train/epoch_time_elapsed': '292.2', 'train/estimated_remaining_minutes': '0.07191'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 69.40it/s]
+ 96%|█████████▌| 2400/2500 [05:04<00:04, 22.6[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.11', 'grad_norm': '1.414', 'learning_rate': '0.0003', 'epoch': '0.07819', 'train/total_time_seconds': '50.03', 'train/time_per_step_avg': '0.02046', 'train/epoch_time_elapsed': '293.1', 'train/estimated_remaining_minutes': '0.06469'}
+{'loss': '3.098', 'grad_norm': '1.352', 'learning_rate': '0.0003', 'epoch': '0.07887', 'train/total_time_seconds': '50.43', 'train/time_per_step_avg': '0.02045', 'train/epoch_time_elapsed': '294', 'train/estimated_remaining_minutes': '0.05747'}
+{'loss': '3.124', 'grad_norm': '1.375', 'learning_rate': '0.0003', 'epoch': '0.07954', 'train/total_time_seconds': '50.83', 'train/time_per_step_avg': '0.02029', 'train/epoch_time_elapsed': '294.9', 'train/estimated_remaining_minutes': '0.05025'}
+{'loss': '3.112', 'grad_norm': '1.32', 'learning_rate': '0.0003', 'epoch': '0.08022', 'train/total_time_seconds': '51.23', 'train/time_per_step_avg': '0.02011', 'train/epoch_time_elapsed': '295.8', 'train/estimated_remaining_minutes': '0.04305'}
+{'loss': '3.112', 'grad_norm': '1.508', 'learning_rate': '0.0003', 'epoch': '0.08089', 'train/total_time_seconds': '51.63', 'train/time_per_step_avg': '0.02007', 'train/epoch_time_elapsed': '296.7', 'train/estimated_remaining_minutes': '0.03585'}
+  - 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
+{'eval_loss': '3.104', 'eval_runtime': '7.976', 'eval_samples_per_second': '1194', 'eval_steps_per_second': '0.878', 'epoch': '0.08089', 'train/total_time_seconds': '51.63', 'train/time_per_step_avg': '0.02007', 'train/epoch_time_elapsed': '304.7', 'train/estimated_remaining_minutes': '0.03585'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 178.16it/s]
+100%|██████████| 2500/2500 [05:17<00:00, 21.5[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.128', 'grad_norm': '1.297', 'learning_rate': '0.0003', 'epoch': '0.08156', 'train/total_time_seconds': '52.04', 'train/time_per_step_avg': '0.02009', 'train/epoch_time_elapsed': '305.6', 'train/estimated_remaining_minutes': '0.02867'}
+{'loss': '3.074', 'grad_norm': '1.547', 'learning_rate': '0.0003', 'epoch': '0.08224', 'train/total_time_seconds': '52.44', 'train/time_per_step_avg': '0.02013', 'train/epoch_time_elapsed': '306.5', 'train/estimated_remaining_minutes': '0.02149'}
+{'loss': '3.109', 'grad_norm': '1.43', 'learning_rate': '0.0003', 'epoch': '0.08291', 'train/total_time_seconds': '52.86', 'train/time_per_step_avg': '0.02035', 'train/epoch_time_elapsed': '307.4', 'train/estimated_remaining_minutes': '0.01433'}
+{'loss': '3.086', 'grad_norm': '1.391', 'learning_rate': '0.0003', 'epoch': '0.08359', 'train/total_time_seconds': '53.3', 'train/time_per_step_avg': '0.02077', 'train/epoch_time_elapsed': '308.4', 'train/estimated_remaining_minutes': '0.007165'}
+{'loss': '3.078', 'grad_norm': '1.32', 'learning_rate': '0.0003', 'epoch': '0.08426', 'train/total_time_seconds': '53.73', 'train/time_per_step_avg': '0.02098', 'train/epoch_time_elapsed': '309.3', 'train/estimated_remaining_minutes': '0'}
+  - 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
+{'eval_loss': '3.086', 'eval_runtime': '8.04', 'eval_samples_per_second': '1185', 'eval_steps_per_second': '0.871', 'epoch': '0.08426', 'train/total_time_seconds': '53.73', 'train/time_per_step_avg': '0.02098', 'train/epoch_time_elapsed': '317.4', 'train/estimated_remaining_minutes': '0'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 180.80it/s]
+100%|██████████| 2500/2500 [05:17<00:00,  7.88it/s], ?it/s]
+{'train_runtime': '318.4', 'train_samples_per_second': '251.2', 'train_steps_per_second': '7.851', 'train_loss': '3.913', 'epoch': '0.08426', 'train/total_time_seconds': '53.73', 'train/time_per_step_avg': '0.02098', 'train/epoch_time_elapsed': '317.4', 'train/estimated_remaining_minutes': '0'}
+100%|██████████| 7/7 [00:05<00:00,  1.21it/s]
+[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - 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
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 177.88it/s]
+>>> FINISHED mlp-quadratic-9L successfully

zain/Activation/wandb/run-20260820_231613-z5ghx5k4/files/wandb-summary.json

@@ -0,0 +1 @@
+{"eval/loss":3.0855629444122314,"eval/runtime":8.2426,"_runtime":325,"train/learning_rate":0.0003,"total_flos":3.6298555392e+14,"eval/samples_per_second":1155.831,"_wandb":{"runtime":325},"train_samples_per_second":251.247,"train/loss":3.078124237060547,"train_runtime":318.412,"train_steps_per_second":7.851,"train/global_step":2500,"train/grad_norm":1.3203125,"train/epoch":0.08426019548365352,"eval/steps_per_second":0.849,"_timestamp":1.7872681003884547e+09,"train_loss":3.9130628158569336,"_step":151}

zain/Activation/wandb/run-20260820_231613-z5ghx5k4/logs/debug-core.log

@@ -63,3 +63,10 @@
 {"time":"2026-08-20T23:16:13.990443666Z","level":"INFO","msg":"handleInformInit: received","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:14.261935343Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"z5ghx5k4","id":"5(@)"}
 {"time":"2026-08-20T23:16:20.112222253Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:40.403706695Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.099860028Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.101376929Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.10192683Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.842145532Z","level":"INFO","msg":"handleInformInit: received","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:42.103852099Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:47.827855381Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"xrlr57kv5zv0"}

zain/Activation/wandb/run-20260820_231613-z5ghx5k4/logs/debug-internal.log

@@ -31,3 +31,29 @@
 {"time":"2026-08-20T23:18:59.964774028Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
 {"time":"2026-08-20T23:19:14.747180756Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":77,"history_lines":8,"events_offset":21,"events_lines":2,"console_offset":132,"console_lines":1}
 {"time":"2026-08-20T23:19:15.17955498Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:19:29.747471524Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":85,"history_lines":9,"events_offset":23,"events_lines":2,"console_offset":138,"console_lines":32}
+{"time":"2026-08-20T23:19:30.23116623Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:19:44.74805619Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":94,"history_lines":7,"events_offset":25,"events_lines":2,"console_offset":165,"console_lines":1}
+{"time":"2026-08-20T23:19:45.152842931Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:19:59.74745341Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":101,"history_lines":6,"events_offset":27,"events_lines":2,"console_offset":170,"console_lines":24}
+{"time":"2026-08-20T23:20:00.132277593Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:20:14.74731999Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":107,"history_lines":6,"events_offset":29,"events_lines":2,"console_offset":187,"console_lines":1}
+{"time":"2026-08-20T23:20:15.188088498Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:20:29.747833214Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":113,"history_lines":7,"events_offset":31,"events_lines":2,"console_offset":193,"console_lines":28}
+{"time":"2026-08-20T23:20:30.076042417Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:20:44.747762003Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":120,"history_lines":9,"events_offset":33,"events_lines":2,"console_offset":220,"console_lines":15}
+{"time":"2026-08-20T23:20:45.150400861Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:20:59.747169256Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":129,"history_lines":8,"events_offset":35,"events_lines":2,"console_offset":231,"console_lines":1}
+{"time":"2026-08-20T23:20:59.911866446Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:21:14.747180731Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":137,"history_lines":6,"events_offset":37,"events_lines":2,"console_offset":235,"console_lines":25}
+{"time":"2026-08-20T23:21:15.219560091Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:21:29.747404617Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":143,"history_lines":6,"events_offset":39,"events_lines":2,"console_offset":253,"console_lines":1}
+{"time":"2026-08-20T23:21:30.189937435Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:21:40.913145038Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-08-20T23:21:40.913401218Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":149,"history_lines":3,"events_offset":41,"events_lines":1,"console_offset":259,"console_lines":25,"uploaded_len":3,"complete":true,"exit_code":0}
+{"time":"2026-08-20T23:21:41.097022958Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:21:41.09850287Z","level":"INFO","msg":"handler: operation stats","stats":{}}
+{"time":"2026-08-20T23:21:41.101391781Z","level":"INFO","msg":"stream: finishing up"}
+{"time":"2026-08-20T23:21:41.101414864Z","level":"INFO","msg":"handler: closed"}
+{"time":"2026-08-20T23:21:41.101475909Z","level":"INFO","msg":"sender: closed"}
+{"time":"2026-08-20T23:21:41.101480969Z","level":"INFO","msg":"stream: all finished"}

zain/Activation/wandb/run-20260820_231613-z5ghx5k4/logs/debug.log

@@ -18,3 +18,8 @@ config: {'_wandb': {}}
 2026-08-20 23:16:14,743 INFO    MainThread:58046 [wandb_run.py:_config_callback():1346] config_cb None None {'transformers_version': '5.16.0.dev0', 'architectures': None, 'output_hidden_states': False, 'return_dict': True, 'dtype': None, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'problem_type': None, 'vocab_size': 4096, 'hidden_size': 128, 'intermediate_size': 256, 'num_hidden_layers': 9, 'num_attention_heads': 4, 'num_key_value_heads': 4, 'hidden_act': 'silu', 'max_position_embeddings': 512, 'initializer_range': 0.02, 'rms_norm_eps': 1e-06, 'use_cache': False, 'pad_token_id': 0, 'bos_token_id': 1, 'eos_token_id': 2, 'pretraining_tp': 1, 'tie_word_embeddings': True, 'rope_parameters': {'rope_theta': 10000.0, 'rope_type': 'default'}, 'attention_bias': False, 'attention_dropout': 0.0, 'mlp_bias': False, 'head_dim': 32, '_name_or_path': '', 'tokenizer_name': 'w-ahmad/tiny-stories-tokenizer', 'mlp_type': 'mlp', 'activation': 'quadratic', 'waleed_beta': 10.0, 'powlu_m': 3.0, 'model_type': 'tiny_llama', 'output_attentions': False, 'output_dir': 'out/mlp-quadratic-9L_run', 'per_device_train_batch_size': 32, 'num_train_epochs': 1, 'max_steps': 2500, 'learning_rate': 0.0003, 'lr_scheduler_type': 'constant_with_warmup', 'lr_scheduler_kwargs': None, 'warmup_steps': 500, 'optim': 'adamw_torch_fused', 'optim_args': None, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'optim_target_modules': None, 'gradient_accumulation_steps': 1, 'average_tokens_across_devices': True, 'max_grad_norm': 1.0, 'label_smoothing_factor': 0.0, 'bf16': True, 'fp16': False, 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'use_liger_kernel': False, 'liger_kernel_config': None, 'neftune_noise_alpha': None, 'torch_empty_cache_steps': None, 'auto_find_batch_size': False, 'logging_strategy': 'steps', 'logging_steps': 20, 'logging_first_step': False, 'log_on_each_node': True, 'logging_nan_inf_filter': True, 'include_num_input_tokens_seen': 'no', 'log_level': 'passive', 'log_level_replica': 'warning', 'disable_tqdm': False, 'report_to': ['wandb'], 'run_name': 'LM-mlp-quadratic-9L-2.0M-20260820-231613', 'project': 'huggingface', 'trackio_space_id': None, 'trackio_bucket_id': None, 'trackio_static_space_id': None, 'eval_strategy': 'steps', 'eval_steps': 100, 'eval_delay': 0, 'per_device_eval_batch_size': 1500, 'prediction_loss_only': False, 'eval_on_start': False, 'eval_do_concat_batches': True, 'eval_use_gather_object': False, 'eval_accumulation_steps': None, 'include_for_metrics': [], 'batch_eval_metrics': False, 'save_only_model': False, 'save_strategy': 'steps', 'save_steps': 100, 'save_on_each_node': False, 'save_total_limit': None, 'enable_jit_checkpoint': False, 'push_to_hub': False, 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': None, 'hub_model_id': 'w-ahmad/6L-mlp-quadratic-9L', 'hub_strategy': 'every_save', 'hub_always_push': False, 'hub_revision': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'restore_callback_states_from_checkpoint': False, 'full_determinism': False, 'seed': 42, 'data_seed': 42, 'use_cpu': False, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}, 'parallelism_config': None, 'dataloader_drop_last': False, 'dataloader_num_workers': 0, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'dataloader_prefetch_factor': None, 'dataloader_multiprocessing_context': None, 'dataloader_in_order': True, 'remove_unused_columns': False, 'label_names': None, 'train_sampling_strategy': 'random', 'length_column_name': 'length', 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'ddp_static_graph': None, 'ddp_backend': None, 'ddp_timeout': 1800, 'fsdp': None, 'fsdp_config': None, 'deepspeed': None, 'debug': [], 'skip_memory_metrics': True, 'do_train': False, 'do_eval': True, 'do_predict': False, 'resume_from_checkpoint': None, 'local_rank': -1}
 2026-08-20 23:16:14,745 INFO    MainThread:58046 [wandb_config.py:__setitem__():155] [no run ID] config set model/num_parameters = 2001280 - <bound method Run._config_callback of <wandb.sdk.wandb_run.Run object at 0x1526c0593f10>>
 2026-08-20 23:16:14,745 INFO    MainThread:58046 [wandb_run.py:_config_callback():1346] config_cb model/num_parameters 2001280 None
+2026-08-20 23:21:40,402 INFO    MainThread:58046 [wandb_run.py:_finish():2383] finishing run deepnevro-deepnevro/research-ultimate/z5ghx5k4
+2026-08-20 23:21:40,403 INFO    MainThread:58046 [wandb_run.py:_atexit_cleanup():2588] got exitcode: 0
+2026-08-20 23:21:40,403 INFO    MainThread:58046 [wandb_run.py:_restore():2570] restore
+2026-08-20 23:21:40,403 INFO    MainThread:58046 [wandb_run.py:_restore():2576] restore done
+2026-08-20 23:21:41,100 INFO    MainThread:58046 [wandb_run.py:_footer_sync_info():3993] logging synced files

zain/Activation/wandb/run-20260820_231613-z5ghx5k4/run-z5ghx5k4.wandb

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b64b01e0a5ff262175afde6e099c82196c638055271f39579a94d5521bd3d630
-size 262144
+oid sha256:1c0fba156b8c4bec34031213aeb65e51046dafaf6ffcf3fe9769aba1b97bea9a
+size 513135

zain/Activation/wandb/run-20260820_232141-62pb25xf/files/output.log

@@ -0,0 +1,139 @@
+  4%|▍         | 100/2500 [00:12<01:47, 22.37[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '8.315', 'grad_norm': '2.719', 'learning_rate': '1.14e-05', 'epoch': '0.0006741', 'train/total_time_seconds': '0.424', 'train/time_per_step_avg': '0.0212', 'train/epoch_time_elapsed': '0.9616', 'train/estimated_remaining_minutes': '0.8763'}
+{'loss': '8.251', 'grad_norm': '1.852', 'learning_rate': '2.34e-05', 'epoch': '0.001348', 'train/total_time_seconds': '0.8373', 'train/time_per_step_avg': '0.02093', 'train/epoch_time_elapsed': '1.856', 'train/estimated_remaining_minutes': '0.8583'}
+{'loss': '8.105', 'grad_norm': '1.32', 'learning_rate': '3.54e-05', 'epoch': '0.002022', 'train/total_time_seconds': '1.252', 'train/time_per_step_avg': '0.02086', 'train/epoch_time_elapsed': '2.749', 'train/estimated_remaining_minutes': '0.8483'}
+{'loss': '7.968', 'grad_norm': '1.297', 'learning_rate': '4.74e-05', 'epoch': '0.002696', 'train/total_time_seconds': '1.665', 'train/time_per_step_avg': '0.02081', 'train/epoch_time_elapsed': '3.645', 'train/estimated_remaining_minutes': '0.8395'}
+{'loss': '7.805', 'grad_norm': '1.258', 'learning_rate': '5.94e-05', 'epoch': '0.00337', 'train/total_time_seconds': '2.078', 'train/time_per_step_avg': '0.02078', 'train/epoch_time_elapsed': '4.54', 'train/estimated_remaining_minutes': '0.8312'}
+  - 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
+{'eval_loss': '7.704', 'eval_runtime': '8.033', 'eval_samples_per_second': '1186', 'eval_steps_per_second': '0.871', 'epoch': '0.00337', 'train/total_time_seconds': '2.078', 'train/time_per_step_avg': '0.02078', 'train/epoch_time_elapsed': '12.57', 'train/estimated_remaining_minutes': '0.8312'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 177.27it/s]
+  8%|▊         | 200/2500 [00:25<01:42, 22.45[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '7.6', 'grad_norm': '1.273', 'learning_rate': '7.14e-05', 'epoch': '0.004044', 'train/total_time_seconds': '2.498', 'train/time_per_step_avg': '0.02074', 'train/epoch_time_elapsed': '13.51', 'train/estimated_remaining_minutes': '0.8258'}
+{'loss': '7.368', 'grad_norm': '1.266', 'learning_rate': '8.34e-05', 'epoch': '0.004719', 'train/total_time_seconds': '2.913', 'train/time_per_step_avg': '0.02075', 'train/epoch_time_elapsed': '14.41', 'train/estimated_remaining_minutes': '0.8183'}
+{'loss': '7.133', 'grad_norm': '1.266', 'learning_rate': '9.54e-05', 'epoch': '0.005393', 'train/total_time_seconds': '3.34', 'train/time_per_step_avg': '0.02089', 'train/epoch_time_elapsed': '15.36', 'train/estimated_remaining_minutes': '0.8142'}
+{'loss': '6.885', 'grad_norm': '1.203', 'learning_rate': '0.0001074', 'epoch': '0.006067', 'train/total_time_seconds': '3.77', 'train/time_per_step_avg': '0.02105', 'train/epoch_time_elapsed': '16.31', 'train/estimated_remaining_minutes': '0.8099'}
+{'loss': '6.63', 'grad_norm': '1.156', 'learning_rate': '0.0001194', 'epoch': '0.006741', 'train/total_time_seconds': '4.181', 'train/time_per_step_avg': '0.02103', 'train/epoch_time_elapsed': '17.2', 'train/estimated_remaining_minutes': '0.8013'}
+  - 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
+{'eval_loss': '6.5', 'eval_runtime': '7.948', 'eval_samples_per_second': '1199', 'eval_steps_per_second': '0.881', 'epoch': '0.006741', 'train/total_time_seconds': '4.181', 'train/time_per_step_avg': '0.02103', 'train/epoch_time_elapsed': '25.15', 'train/estimated_remaining_minutes': '0.8013'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 179.32it/s]
+ 12%|█▏        | 300/2500 [00:37<01:38, 22.39[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '6.391', 'grad_norm': '1.109', 'learning_rate': '0.0001314', 'epoch': '0.007415', 'train/total_time_seconds': '4.593', 'train/time_per_step_avg': '0.02095', 'train/epoch_time_elapsed': '26.08', 'train/estimated_remaining_minutes': '0.7934'}
+{'loss': '6.139', 'grad_norm': '1.219', 'learning_rate': '0.0001434', 'epoch': '0.008089', 'train/total_time_seconds': '5.004', 'train/time_per_step_avg': '0.02091', 'train/epoch_time_elapsed': '26.97', 'train/estimated_remaining_minutes': '0.7853'}
+{'loss': '5.917', 'grad_norm': '1.062', 'learning_rate': '0.0001554', 'epoch': '0.008763', 'train/total_time_seconds': '5.418', 'train/time_per_step_avg': '0.02077', 'train/epoch_time_elapsed': '27.87', 'train/estimated_remaining_minutes': '0.7779'}
+{'loss': '5.702', 'grad_norm': '1.102', 'learning_rate': '0.0001674', 'epoch': '0.009437', 'train/total_time_seconds': '5.832', 'train/time_per_step_avg': '0.02062', 'train/epoch_time_elapsed': '28.76', 'train/estimated_remaining_minutes': '0.7707'}
+{'loss': '5.5', 'grad_norm': '0.9219', 'learning_rate': '0.0001794', 'epoch': '0.01011', 'train/total_time_seconds': '6.248', 'train/time_per_step_avg': '0.02067', 'train/epoch_time_elapsed': '29.65', 'train/estimated_remaining_minutes': '0.7636'}
+  - 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
+{'eval_loss': '5.394', 'eval_runtime': '8.045', 'eval_samples_per_second': '1184', 'eval_steps_per_second': '0.87', 'epoch': '0.01011', 'train/total_time_seconds': '6.248', 'train/time_per_step_avg': '0.02067', 'train/epoch_time_elapsed': '37.7', 'train/estimated_remaining_minutes': '0.7636'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 133.36it/s]
+ 16%|█▌        | 400/2500 [00:50<01:51, 18.83[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '5.319', 'grad_norm': '2.5', 'learning_rate': '0.0001914', 'epoch': '0.01079', 'train/total_time_seconds': '6.661', 'train/time_per_step_avg': '0.02068', 'train/epoch_time_elapsed': '38.62', 'train/estimated_remaining_minutes': '0.7563'}
+{'loss': '5.202', 'grad_norm': '1.078', 'learning_rate': '0.0002034', 'epoch': '0.01146', 'train/total_time_seconds': '7.076', 'train/time_per_step_avg': '0.02072', 'train/epoch_time_elapsed': '39.52', 'train/estimated_remaining_minutes': '0.7493'}
+{'loss': '5.021', 'grad_norm': '1.32', 'learning_rate': '0.0002154', 'epoch': '0.01213', 'train/total_time_seconds': '7.488', 'train/time_per_step_avg': '0.02071', 'train/epoch_time_elapsed': '40.41', 'train/estimated_remaining_minutes': '0.7419'}
+{'loss': '4.873', 'grad_norm': '1.117', 'learning_rate': '0.0002274', 'epoch': '0.01281', 'train/total_time_seconds': '7.906', 'train/time_per_step_avg': '0.02074', 'train/epoch_time_elapsed': '41.31', 'train/estimated_remaining_minutes': '0.7351'}
+{'loss': '4.742', 'grad_norm': '1.234', 'learning_rate': '0.0002394', 'epoch': '0.01348', 'train/total_time_seconds': '8.367', 'train/time_per_step_avg': '0.02119', 'train/epoch_time_elapsed': '42.31', 'train/estimated_remaining_minutes': '0.7321'}
+  - 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
+{'eval_loss': '4.676', 'eval_runtime': '8.046', 'eval_samples_per_second': '1184', 'eval_steps_per_second': '0.87', 'epoch': '0.01348', 'train/total_time_seconds': '8.367', 'train/time_per_step_avg': '0.02119', 'train/epoch_time_elapsed': '50.36', 'train/estimated_remaining_minutes': '0.7321'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 180.22it/s]
+ 20%|██        | 500/2500 [01:02<01:30, 22.08[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '4.607', 'grad_norm': '1.305', 'learning_rate': '0.0002514', 'epoch': '0.01416', 'train/total_time_seconds': '8.8', 'train/time_per_step_avg': '0.02139', 'train/epoch_time_elapsed': '51.32', 'train/estimated_remaining_minutes': '0.7263'}
+{'loss': '4.515', 'grad_norm': '1.258', 'learning_rate': '0.0002634', 'epoch': '0.01483', 'train/total_time_seconds': '9.216', 'train/time_per_step_avg': '0.02139', 'train/epoch_time_elapsed': '52.21', 'train/estimated_remaining_minutes': '0.7191'}
+{'loss': '4.417', 'grad_norm': '1.062', 'learning_rate': '0.0002754', 'epoch': '0.0155', 'train/total_time_seconds': '9.627', 'train/time_per_step_avg': '0.02139', 'train/epoch_time_elapsed': '53.1', 'train/estimated_remaining_minutes': '0.7116'}
+{'loss': '4.305', 'grad_norm': '1.367', 'learning_rate': '0.0002874', 'epoch': '0.01618', 'train/total_time_seconds': '10.04', 'train/time_per_step_avg': '0.02136', 'train/epoch_time_elapsed': '54', 'train/estimated_remaining_minutes': '0.7044'}
+{'loss': '4.239', 'grad_norm': '1.211', 'learning_rate': '0.0002994', 'epoch': '0.01685', 'train/total_time_seconds': '10.47', 'train/time_per_step_avg': '0.02099', 'train/epoch_time_elapsed': '54.91', 'train/estimated_remaining_minutes': '0.6978'}
+  - 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
+{'eval_loss': '4.196', 'eval_runtime': '7.93', 'eval_samples_per_second': '1201', 'eval_steps_per_second': '0.883', 'epoch': '0.01685', 'train/total_time_seconds': '10.47', 'train/time_per_step_avg': '0.02099', 'train/epoch_time_elapsed': '62.84', 'train/estimated_remaining_minutes': '0.6978'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 118.36it/s]
+ 24%|██▍       | 600/2500 [01:15<01:25, 22.25[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '4.176', 'grad_norm': '0.9727', 'learning_rate': '0.0003', 'epoch': '0.01753', 'train/total_time_seconds': '10.91', 'train/time_per_step_avg': '0.02111', 'train/epoch_time_elapsed': '63.8', 'train/estimated_remaining_minutes': '0.6924'}
+{'loss': '4.092', 'grad_norm': '0.9141', 'learning_rate': '0.0003', 'epoch': '0.0182', 'train/total_time_seconds': '11.33', 'train/time_per_step_avg': '0.02116', 'train/epoch_time_elapsed': '64.71', 'train/estimated_remaining_minutes': '0.6855'}
+{'loss': '4.032', 'grad_norm': '1.289', 'learning_rate': '0.0003', 'epoch': '0.01887', 'train/total_time_seconds': '11.75', 'train/time_per_step_avg': '0.02126', 'train/epoch_time_elapsed': '65.62', 'train/estimated_remaining_minutes': '0.6786'}
+{'loss': '3.976', 'grad_norm': '1.125', 'learning_rate': '0.0003', 'epoch': '0.01955', 'train/total_time_seconds': '12.17', 'train/time_per_step_avg': '0.02126', 'train/epoch_time_elapsed': '66.52', 'train/estimated_remaining_minutes': '0.6714'}
+{'loss': '3.935', 'grad_norm': '1.219', 'learning_rate': '0.0003', 'epoch': '0.02022', 'train/total_time_seconds': '12.59', 'train/time_per_step_avg': '0.02123', 'train/epoch_time_elapsed': '67.42', 'train/estimated_remaining_minutes': '0.6644'}
+  - 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
+{'eval_loss': '3.916', 'eval_runtime': '8.061', 'eval_samples_per_second': '1182', 'eval_steps_per_second': '0.868', 'epoch': '0.02022', 'train/total_time_seconds': '12.59', 'train/time_per_step_avg': '0.02123', 'train/epoch_time_elapsed': '75.48', 'train/estimated_remaining_minutes': '0.6644'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 176.32it/s]
+ 28%|██▊       | 700/2500 [01:28<01:20, 22.26[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.878', 'grad_norm': '1.164', 'learning_rate': '0.0003', 'epoch': '0.0209', 'train/total_time_seconds': '13', 'train/time_per_step_avg': '0.02094', 'train/epoch_time_elapsed': '76.41', 'train/estimated_remaining_minutes': '0.6572'}
+{'loss': '3.855', 'grad_norm': '1.484', 'learning_rate': '0.0003', 'epoch': '0.02157', 'train/total_time_seconds': '13.43', 'train/time_per_step_avg': '0.02095', 'train/epoch_time_elapsed': '77.32', 'train/estimated_remaining_minutes': '0.6504'}
+{'loss': '3.823', 'grad_norm': '1.336', 'learning_rate': '0.0003', 'epoch': '0.02224', 'train/total_time_seconds': '13.86', 'train/time_per_step_avg': '0.02104', 'train/epoch_time_elapsed': '78.27', 'train/estimated_remaining_minutes': '0.6439'}
+{'loss': '3.783', 'grad_norm': '1.164', 'learning_rate': '0.0003', 'epoch': '0.02292', 'train/total_time_seconds': '14.27', 'train/time_per_step_avg': '0.02106', 'train/epoch_time_elapsed': '79.2', 'train/estimated_remaining_minutes': '0.6368'}
+{'loss': '3.776', 'grad_norm': '1.133', 'learning_rate': '0.0003', 'epoch': '0.02359', 'train/total_time_seconds': '14.69', 'train/time_per_step_avg': '0.021', 'train/epoch_time_elapsed': '80.09', 'train/estimated_remaining_minutes': '0.6295'}
+  - 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
+{'eval_loss': '3.76', 'eval_runtime': '7.974', 'eval_samples_per_second': '1195', 'eval_steps_per_second': '0.878', 'epoch': '0.02359', 'train/total_time_seconds': '14.69', 'train/time_per_step_avg': '0.021', 'train/epoch_time_elapsed': '88.07', 'train/estimated_remaining_minutes': '0.6295'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 162.43it/s]
+ 32%|███▏      | 800/2500 [01:40<01:19, 21.36[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.743', 'grad_norm': '1.25', 'learning_rate': '0.0003', 'epoch': '0.02427', 'train/total_time_seconds': '15.12', 'train/time_per_step_avg': '0.02115', 'train/epoch_time_elapsed': '89.03', 'train/estimated_remaining_minutes': '0.623'}
+{'loss': '3.732', 'grad_norm': '1.312', 'learning_rate': '0.0003', 'epoch': '0.02494', 'train/total_time_seconds': '15.55', 'train/time_per_step_avg': '0.02123', 'train/epoch_time_elapsed': '89.95', 'train/estimated_remaining_minutes': '0.6164'}
+{'loss': '3.684', 'grad_norm': '1.266', 'learning_rate': '0.0003', 'epoch': '0.02562', 'train/total_time_seconds': '15.98', 'train/time_per_step_avg': '0.02122', 'train/epoch_time_elapsed': '90.86', 'train/estimated_remaining_minutes': '0.6097'}
+{'loss': '3.646', 'grad_norm': '1.164', 'learning_rate': '0.0003', 'epoch': '0.02629', 'train/total_time_seconds': '16.41', 'train/time_per_step_avg': '0.02139', 'train/epoch_time_elapsed': '91.78', 'train/estimated_remaining_minutes': '0.6032'}
+{'loss': '3.654', 'grad_norm': '1.117', 'learning_rate': '0.0003', 'epoch': '0.02696', 'train/total_time_seconds': '16.85', 'train/time_per_step_avg': '0.02165', 'train/epoch_time_elapsed': '92.73', 'train/estimated_remaining_minutes': '0.5969'}
+  - 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
+{'eval_loss': '3.652', 'eval_runtime': '7.975', 'eval_samples_per_second': '1195', 'eval_steps_per_second': '0.878', 'epoch': '0.02696', 'train/total_time_seconds': '16.85', 'train/time_per_step_avg': '0.02165', 'train/epoch_time_elapsed': '100.7', 'train/estimated_remaining_minutes': '0.5969'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 174.69it/s]
+ 36%|███▌      | 900/2500 [01:53<01:16, 20.91[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.616', 'grad_norm': '1.117', 'learning_rate': '0.0003', 'epoch': '0.02764', 'train/total_time_seconds': '17.28', 'train/time_per_step_avg': '0.0216', 'train/epoch_time_elapsed': '101.6', 'train/estimated_remaining_minutes': '0.59'}
+{'loss': '3.61', 'grad_norm': '1.078', 'learning_rate': '0.0003', 'epoch': '0.02831', 'train/total_time_seconds': '17.7', 'train/time_per_step_avg': '0.02148', 'train/epoch_time_elapsed': '102.6', 'train/estimated_remaining_minutes': '0.5829'}
+{'loss': '3.634', 'grad_norm': '1.273', 'learning_rate': '0.0003', 'epoch': '0.02899', 'train/total_time_seconds': '18.11', 'train/time_per_step_avg': '0.02136', 'train/epoch_time_elapsed': '103.5', 'train/estimated_remaining_minutes': '0.5757'}
+{'loss': '3.586', 'grad_norm': '1.141', 'learning_rate': '0.0003', 'epoch': '0.02966', 'train/total_time_seconds': '18.55', 'train/time_per_step_avg': '0.02136', 'train/epoch_time_elapsed': '104.4', 'train/estimated_remaining_minutes': '0.5691'}
+{'loss': '3.549', 'grad_norm': '1.109', 'learning_rate': '0.0003', 'epoch': '0.03033', 'train/total_time_seconds': '18.99', 'train/time_per_step_avg': '0.0214', 'train/epoch_time_elapsed': '105.4', 'train/estimated_remaining_minutes': '0.5628'}
+  - 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
+{'eval_loss': '3.565', 'eval_runtime': '8.321', 'eval_samples_per_second': '1145', 'eval_steps_per_second': '0.841', 'epoch': '0.03033', 'train/total_time_seconds': '18.99', 'train/time_per_step_avg': '0.0214', 'train/epoch_time_elapsed': '113.7', 'train/estimated_remaining_minutes': '0.5628'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 181.04it/s]
+ 40%|████      | 1000/2500 [02:06<01:09, 21.6[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.565', 'grad_norm': '1.102', 'learning_rate': '0.0003', 'epoch': '0.03101', 'train/total_time_seconds': '19.42', 'train/time_per_step_avg': '0.02141', 'train/epoch_time_elapsed': '114.6', 'train/estimated_remaining_minutes': '0.5559'}
+{'loss': '3.548', 'grad_norm': '1.25', 'learning_rate': '0.0003', 'epoch': '0.03168', 'train/total_time_seconds': '19.84', 'train/time_per_step_avg': '0.02141', 'train/epoch_time_elapsed': '115.5', 'train/estimated_remaining_minutes': '0.5487'}
+{'loss': '3.514', 'grad_norm': '1.141', 'learning_rate': '0.0003', 'epoch': '0.03236', 'train/total_time_seconds': '20.26', 'train/time_per_step_avg': '0.02143', 'train/epoch_time_elapsed': '116.4', 'train/estimated_remaining_minutes': '0.5416'}
+{'loss': '3.502', 'grad_norm': '1.141', 'learning_rate': '0.0003', 'epoch': '0.03303', 'train/total_time_seconds': '20.67', 'train/time_per_step_avg': '0.02125', 'train/epoch_time_elapsed': '117.3', 'train/estimated_remaining_minutes': '0.5344'}
+{'loss': '3.501', 'grad_norm': '1.359', 'learning_rate': '0.0003', 'epoch': '0.0337', 'train/total_time_seconds': '21.11', 'train/time_per_step_avg': '0.02119', 'train/epoch_time_elapsed': '118.3', 'train/estimated_remaining_minutes': '0.5278'}
+  - 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
+{'eval_loss': '3.498', 'eval_runtime': '8.101', 'eval_samples_per_second': '1176', 'eval_steps_per_second': '0.864', 'epoch': '0.0337', 'train/total_time_seconds': '21.11', 'train/time_per_step_avg': '0.02119', 'train/epoch_time_elapsed': '126.4', 'train/estimated_remaining_minutes': '0.5278'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 169.51it/s]
+ 44%|████▍     | 1100/2500 [02:19<01:03, 22.0[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.491', 'grad_norm': '1.133', 'learning_rate': '0.0003', 'epoch': '0.03438', 'train/total_time_seconds': '21.55', 'train/time_per_step_avg': '0.02131', 'train/epoch_time_elapsed': '127.3', 'train/estimated_remaining_minutes': '0.5212'}
+{'loss': '3.474', 'grad_norm': '1.117', 'learning_rate': '0.0003', 'epoch': '0.03505', 'train/total_time_seconds': '21.97', 'train/time_per_step_avg': '0.0213', 'train/epoch_time_elapsed': '128.2', 'train/estimated_remaining_minutes': '0.514'}
+{'loss': '3.455', 'grad_norm': '1.281', 'learning_rate': '0.0003', 'epoch': '0.03573', 'train/total_time_seconds': '22.39', 'train/time_per_step_avg': '0.02129', 'train/epoch_time_elapsed': '129.2', 'train/estimated_remaining_minutes': '0.5069'}
+{'loss': '3.441', 'grad_norm': '1.203', 'learning_rate': '0.0003', 'epoch': '0.0364', 'train/total_time_seconds': '22.8', 'train/time_per_step_avg': '0.02129', 'train/epoch_time_elapsed': '130.1', 'train/estimated_remaining_minutes': '0.4997'}
+{'loss': '3.436', 'grad_norm': '1.273', 'learning_rate': '0.0003', 'epoch': '0.03707', 'train/total_time_seconds': '23.22', 'train/time_per_step_avg': '0.02107', 'train/epoch_time_elapsed': '131', 'train/estimated_remaining_minutes': '0.4926'}
+  - 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
+{'eval_loss': '3.445', 'eval_runtime': '8.164', 'eval_samples_per_second': '1167', 'eval_steps_per_second': '0.857', 'epoch': '0.03707', 'train/total_time_seconds': '23.22', 'train/time_per_step_avg': '0.02107', 'train/epoch_time_elapsed': '139.1', 'train/estimated_remaining_minutes': '0.4926'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 161.68it/s]
+ 48%|████▊     | 1200/2500 [02:32<01:01, 21.1[transformers] TinyLlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+{'loss': '3.437', 'grad_norm': '1.031', 'learning_rate': '0.0003', 'epoch': '0.03775', 'train/total_time_seconds': '23.67', 'train/time_per_step_avg': '0.02123', 'train/epoch_time_elapsed': '140.1', 'train/estimated_remaining_minutes': '0.4862'}
+{'loss': '3.411', 'grad_norm': '1.242', 'learning_rate': '0.0003', 'epoch': '0.03842', 'train/total_time_seconds': '24.16', 'train/time_per_step_avg': '0.02193', 'train/epoch_time_elapsed': '141.1', 'train/estimated_remaining_minutes': '0.4804'}
+{'loss': '3.407', 'grad_norm': '1.07', 'learning_rate': '0.0003', 'epoch': '0.0391', 'train/total_time_seconds': '24.61', 'train/time_per_step_avg': '0.0222', 'train/epoch_time_elapsed': '142', 'train/estimated_remaining_minutes': '0.4738'}
+{'loss': '3.423', 'grad_norm': '1.031', 'learning_rate': '0.0003', 'epoch': '0.03977', 'train/total_time_seconds': '25.15', 'train/time_per_step_avg': '0.02346', 'train/epoch_time_elapsed': '143.1', 'train/estimated_remaining_minutes': '0.4689'}
+{'loss': '3.391', 'grad_norm': '1.203', 'learning_rate': '0.0003', 'epoch': '0.04044', 'train/total_time_seconds': '25.6', 'train/time_per_step_avg': '0.02381', 'train/epoch_time_elapsed': '144', 'train/estimated_remaining_minutes': '0.4622'}
+  - 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
+{'eval_loss': '3.395', 'eval_runtime': '8.043', 'eval_samples_per_second': '1184', 'eval_steps_per_second': '0.87', 'epoch': '0.04044', 'train/total_time_seconds': '25.6', 'train/time_per_step_avg': '0.02381', 'train/epoch_time_elapsed': '152', 'train/estimated_remaining_minutes': '0.4622'}
+  - 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).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+Writing model shards: 100%|██████████| 1/1 [00:00<00:00, 162.46it/s]
+ 52%|█████▏    | 1300/2500 [02:36<00:56, 21.37it/s], ?it/s]
+{'loss': '3.361', 'grad_norm': '1.055', 'learning_rate': '0.0003', 'epoch': '0.04112', 'train/total_time_seconds': '26.05', 'train/time_per_step_avg': '0.02371', 'train/epoch_time_elapsed': '153', 'train/estimated_remaining_minutes': '0.4554'}
+{'loss': '3.379', 'grad_norm': '1.094', 'learning_rate': '0.0003', 'epoch': '0.04179', 'train/total_time_seconds': '26.49', 'train/time_per_step_avg': '0.0233', 'train/epoch_time_elapsed': '153.9', 'train/estimated_remaining_minutes': '0.4486'}
+{'loss': '3.388', 'grad_norm': '1.352', 'learning_rate': '0.0003', 'epoch': '0.04247', 'train/total_time_seconds': '26.94', 'train/time_per_step_avg': '0.02335', 'train/epoch_time_elapsed': '154.9', 'train/estimated_remaining_minutes': '0.4419'}
+{'loss': '3.365', 'grad_norm': '1.281', 'learning_rate': '0.0003', 'epoch': '0.04314', 'train/total_time_seconds': '27.39', 'train/time_per_step_avg': '0.02239', 'train/epoch_time_elapsed': '155.8', 'train/estimated_remaining_minutes': '0.4351'}
+{'loss': '3.357', 'grad_norm': '1.18', 'learning_rate': '0.0003', 'epoch': '0.04382', 'train/total_time_seconds': '27.84', 'train/time_per_step_avg': '0.02234', 'train/epoch_time_elapsed': '156.8', 'train/estimated_remaining_minutes': '0.4282'}
+  0%|          | 0/7 [00:00<?, ?it/s]

zain/Activation/wandb/run-20260820_232141-62pb25xf/files/requirements.txt

@@ -0,0 +1,149 @@
+asttokens==3.0.1
+comm==0.2.3
+debugpy==1.8.21
+decorator==5.3.1
+executing==2.2.1
+nest-asyncio==1.6.0
+parso==0.8.7
+platformdirs==4.11.0
+psutil==7.2.2
+ptyprocess==0.7.0
+pure_eval==0.2.3
+Pygments==2.20.0
+pyzmq==27.1.0
+setuptools==83.0.0
+six==1.17.0
+tornado==6.5.7
+traitlets==5.15.0
+fsspec==2026.4.0
+wcwidth==0.8.2
+ipython_pygments_lexers==1.1.1
+jedi==0.20.0
+jupyter_core==5.9.1
+matplotlib-inline==0.2.2
+pexpect==4.9.0
+prompt_toolkit==3.0.53
+python-dateutil==2.9.0.post0
+stack_data==0.6.3
+wheel==0.47.0
+jupyter_client==8.9.1
+pip==26.1.2
+ipython==9.15.0
+ipykernel==7.2.0
+threadpoolctl==3.6.0
+pyparsing==3.3.2
+typing_extensions==4.15.0
+Jinja2==3.1.6
+narwhals==2.24.0
+kiwisolver==1.5.0
+joblib==1.5.3
+fonttools==4.63.0
+cycler==0.12.1
+scipy==1.17.1
+pandas==3.0.5
+contourpy==1.3.3
+scikit-learn==1.9.0
+matplotlib==3.11.1
+urllib3==2.7.0
+tqdm==4.70.0
+idna==3.18
+charset-normalizer==3.4.9
+certifi==2026.7.22
+requests==2.34.2
+seaborn==0.13.2
+uv==0.12.0
+shellingham==1.5.4
+mpmath==1.3.0
+attrs==26.1.0
+hf-xet==1.5.2
+nvidia-nccl-cu12==2.21.5
+MarkupSafe==3.0.3
+regex==2026.7.19
+importlib_metadata==9.0.0
+httpcore==1.0.9
+annotated-doc==0.0.5
+multidict==6.7.1
+aiohttp==3.14.3
+aiosignal==1.4.0
+xxhash==3.8.1
+aiohappyeyeballs==2.7.1
+mdurl==0.1.2
+cuda-toolkit==13.0.3.0
+networkx==3.6.1
+PyYAML==6.0.3
+nvidia-cufile==1.15.1.6
+typer==0.27.0
+torchaudio==2.6.0+cu124
+rich==15.0.0
+nvidia-cufft-cu12==11.2.1.3
+h11==0.16.0
+dill==0.4.1
+cuda-pathfinder==1.6.0
+filelock==3.29.0
+nvidia-nvtx-cu12==12.4.127
+httpx==0.28.1
+anyio==4.14.2
+numpy==2.4.4
+yarl==1.24.5
+click==8.4.2
+triton==3.2.0
+frozenlist==1.8.0
+zipp==4.1.0
+propcache==0.5.2
+markdown-it-py==4.2.0
+nvidia-cuda-runtime==13.0.96
+cuda-bindings==13.3.1
+nvidia-cuda-cupti==13.0.85
+torch==2.6.0+cu124
+multiprocess==0.70.19
+pillow==12.2.0
+transformers==5.16.0.dev0
+wandb==0.28.1
+nvidia-curand==10.4.0.35
+sympy==1.13.1
+nvidia-cusparse==12.6.3.3
+nvidia-cuda-nvrtc==13.0.88
+typing-inspection==0.4.2
+nvidia-cusolver==12.0.4.66
+nvidia-cufft==12.0.0.61
+nvidia-cudnn-cu13==9.20.0.48
+nvidia-cublas==13.1.1.3
+pyarrow==25.0.0
+evaluate==0.4.6
+diffusers==0.39.0
+pydantic==2.13.4
+annotated-types==0.8.0
+protobuf==7.35.1
+sentry-sdk==2.66.1
+einops==0.8.2
+packaging==26.2
+nvidia-nvjitlink-cu12==12.4.127
+nvidia-curand-cu12==10.3.5.147
+nvidia-cusparselt-cu12==0.6.2
+nvidia-cusparse-cu12==12.3.1.170
+nvidia-cuda-runtime-cu12==12.4.127
+torchvision==0.21.0+cu124
+nvidia-cuda-nvrtc-cu12==12.4.127
+nvidia-cuda-cupti-cu12==12.4.127
+nvidia-cusolver-cu12==11.6.1.9
+nvidia-cublas-cu12==12.4.5.8
+nvidia-cudnn-cu12==9.1.0.70
+huggingface_hub==1.26.0
+datasets==5.0.1
+safetensors==0.8.0
+accelerate==1.14.0
+pydantic_core==2.46.4
+ninja==1.13.0
+tokenizers==0.23.1
+autocommand==2.2.2
+backports.tarfile==1.2.0
+importlib_metadata==8.7.1
+jaraco.text==4.0.0
+jaraco.context==6.1.0
+jaraco.functools==4.4.0
+more-itertools==10.8.0
+packaging==26.0
+platformdirs==4.4.0
+tomli==2.4.0
+wheel==0.46.3
+zipp==3.23.0

zain/Activation/wandb/run-20260820_232141-62pb25xf/files/wandb-metadata.json

@@ -0,0 +1,97 @@
+{
+  "os":  "Linux-5.15.0-126-generic-x86_64-with-glibc2.35",
+  "python":  "CPython 3.11.15",
+  "startedAt":  "2026-08-20T23:21:41.839825Z",
+  "args":  [
+    "--config",
+    "/mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/configs/baseline100L.yaml",
+    "--variants",
+    "mlp-squared_relu-6L",
+    "mlp-quadratic-9L",
+    "mlp-squared_relu-9L"
+  ],
+  "program":  "/mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/sweep.py",
+  "codePath":  "sweep.py",
+  "codePathLocal":  "sweep.py",
+  "git":  {
+    "remote":  "https://github.com/w-ahmad1a10/Activation.git",
+    "commit":  "444fe7f656e69d3ec9e58941464da02fec8c6047"
+  },
+  "email":  "deepnevro@gmail.com",
+  "root":  "/mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation",
+  "host":  "deeplens-k3s-node1",
+  "executable":  "/mnt/data/zainulabideen/zain-exp/notebooks/my_env/bin/python",
+  "cpu_count":  112,
+  "cpu_count_logical":  224,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "1560765693952",
+      "used":  "707208392704"
+    }
+  },
+  "memory":  {
+    "total":  "2164089937920"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-39c684a5-fde6-83d7-1663-0859795881ae"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-68012e5a-38b6-b643-0ca6-62fb66720bf3"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-132944c4-b689-2b5f-89a4-d730401677ab"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-2df386cc-6d26-d0e2-7a2d-a057b0d95864"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-bfa16575-1d94-1aa2-4537-2c93433f42ef"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-bc6c3e3c-9b90-09ca-c034-774961847c54"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-00a441e1-7c95-e7d6-4c35-43d6b291aea9"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-1c4d29a2-4647-6fce-d8fc-0c5ecfbbd6ea"
+    }
+  ],
+  "cudaVersion":  "12.4",
+  "writerId":  "o85q53dfqibubnrijsj0z2eb8vlsr6lx"
+}

zain/Activation/wandb/run-20260820_232141-62pb25xf/logs/debug-core.log

@@ -0,0 +1,72 @@
+{"time":"2026-08-20T22:38:55.087253336Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpc4p_eh7y/port-3927710.txt","pid":3927710,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-08-20T22:38:55.099549305Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":3927710}
+{"time":"2026-08-20T22:38:55.099548641Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-3927710-3943595-532688331/socket","Net":"unix"}}
+{"time":"2026-08-20T22:38:55.265065437Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-08-20T22:44:46.859578587Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"2(@)"}
+{"time":"2026-08-20T22:44:46.948318933Z","level":"INFO","msg":"handleInformInit: received","streamId":"wyqgp22a","id":"2(@)"}
+{"time":"2026-08-20T22:44:47.193559758Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"wyqgp22a","id":"2(@)"}
+{"time":"2026-08-20T22:44:52.73238834Z","level":"INFO","msg":"connection: cancelling request","id":"2(@)","requestId":"giupwl3hq1vp"}
+{"time":"2026-08-20T22:44:53.634247433Z","level":"INFO","msg":"connection: cancelling request","id":"2(@)","requestId":"giupwl3hq1vp"}
+{"time":"2026-08-20T22:44:54.279567553Z","level":"INFO","msg":"connection: cancelling request","id":"2(@)","requestId":"giupwl3hq1vp"}
+{"time":"2026-08-20T22:44:54.281083724Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"wyqgp22a","id":"2(@)"}
+{"time":"2026-08-20T22:44:54.282064659Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"wyqgp22a","id":"2(@)"}
+{"time":"2026-08-20T22:44:55.890518909Z","level":"INFO","msg":"connection: closing","id":"2(@)"}
+{"time":"2026-08-20T22:44:55.890548888Z","level":"INFO","msg":"processOutgoingData: finished","id":"2(@)"}
+{"time":"2026-08-20T22:44:55.89059573Z","level":"INFO","msg":"connection: closed successfully","id":"2(@)"}
+{"time":"2026-08-20T22:44:55.890642487Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"2(@)"}
+{"time":"2026-08-20T22:45:28.679676481Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"3(@)"}
+{"time":"2026-08-20T22:45:28.757387797Z","level":"INFO","msg":"handleInformInit: received","streamId":"m7e0j3tz","id":"3(@)"}
+{"time":"2026-08-20T22:45:29.017787043Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"m7e0j3tz","id":"3(@)"}
+{"time":"2026-08-20T22:45:34.464960892Z","level":"INFO","msg":"connection: cancelling request","id":"3(@)","requestId":"iugix4ryqtiw"}
+{"time":"2026-08-20T22:50:12.953504175Z","level":"INFO","msg":"connection: cancelling request","id":"3(@)","requestId":"iugix4ryqtiw"}
+{"time":"2026-08-20T22:50:13.878830759Z","level":"INFO","msg":"connection: cancelling request","id":"3(@)","requestId":"iugix4ryqtiw"}
+{"time":"2026-08-20T22:50:13.880605105Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"m7e0j3tz","id":"3(@)"}
+{"time":"2026-08-20T22:50:13.881805699Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"m7e0j3tz","id":"3(@)"}
+{"time":"2026-08-20T22:50:14.486948166Z","level":"INFO","msg":"handleInformInit: received","streamId":"5ntg3lev","id":"3(@)"}
+{"time":"2026-08-20T22:50:14.746950652Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"5ntg3lev","id":"3(@)"}
+{"time":"2026-08-20T22:50:20.774556518Z","level":"INFO","msg":"connection: cancelling request","id":"3(@)","requestId":"y1z88f23de6u"}
+{"time":"2026-08-20T22:55:01.139166304Z","level":"INFO","msg":"connection: cancelling request","id":"3(@)","requestId":"y1z88f23de6u"}
+{"time":"2026-08-20T22:55:02.043968986Z","level":"INFO","msg":"connection: cancelling request","id":"3(@)","requestId":"y1z88f23de6u"}
+{"time":"2026-08-20T22:55:02.04556083Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"5ntg3lev","id":"3(@)"}
+{"time":"2026-08-20T22:55:02.046266886Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"5ntg3lev","id":"3(@)"}
+{"time":"2026-08-20T22:55:03.638718943Z","level":"INFO","msg":"connection: closing","id":"3(@)"}
+{"time":"2026-08-20T22:55:03.638727793Z","level":"INFO","msg":"processOutgoingData: finished","id":"3(@)"}
+{"time":"2026-08-20T22:55:03.638790655Z","level":"INFO","msg":"connection: closed successfully","id":"3(@)"}
+{"time":"2026-08-20T22:55:03.63880292Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"3(@)"}
+{"time":"2026-08-20T22:56:44.416382285Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"4(@)"}
+{"time":"2026-08-20T22:56:44.673251954Z","level":"INFO","msg":"handleInformInit: received","streamId":"vip8dia7","id":"4(@)"}
+{"time":"2026-08-20T22:56:44.93093962Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"vip8dia7","id":"4(@)"}
+{"time":"2026-08-20T22:56:50.359862325Z","level":"INFO","msg":"connection: cancelling request","id":"4(@)","requestId":"0iv6cfg6vcgu"}
+{"time":"2026-08-20T23:01:32.13011939Z","level":"INFO","msg":"connection: cancelling request","id":"4(@)","requestId":"0iv6cfg6vcgu"}
+{"time":"2026-08-20T23:01:33.019572675Z","level":"INFO","msg":"connection: cancelling request","id":"4(@)","requestId":"0iv6cfg6vcgu"}
+{"time":"2026-08-20T23:01:33.021095606Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"vip8dia7","id":"4(@)"}
+{"time":"2026-08-20T23:01:33.021720993Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"vip8dia7","id":"4(@)"}
+{"time":"2026-08-20T23:01:33.561702921Z","level":"INFO","msg":"handleInformInit: received","streamId":"asswl5zl","id":"4(@)"}
+{"time":"2026-08-20T23:01:33.821998617Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"asswl5zl","id":"4(@)"}
+{"time":"2026-08-20T23:01:39.470923864Z","level":"INFO","msg":"connection: cancelling request","id":"4(@)","requestId":"fs875qei1of6"}
+{"time":"2026-08-20T23:06:41.303239824Z","level":"INFO","msg":"connection: cancelling request","id":"4(@)","requestId":"fs875qei1of6"}
+{"time":"2026-08-20T23:06:42.13099092Z","level":"INFO","msg":"connection: cancelling request","id":"4(@)","requestId":"fs875qei1of6"}
+{"time":"2026-08-20T23:06:42.132504373Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"asswl5zl","id":"4(@)"}
+{"time":"2026-08-20T23:06:42.133111269Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"asswl5zl","id":"4(@)"}
+{"time":"2026-08-20T23:06:44.247650764Z","level":"INFO","msg":"connection: closing","id":"4(@)"}
+{"time":"2026-08-20T23:06:44.247750282Z","level":"INFO","msg":"connection: closed successfully","id":"4(@)"}
+{"time":"2026-08-20T23:06:44.247666352Z","level":"INFO","msg":"processOutgoingData: finished","id":"4(@)"}
+{"time":"2026-08-20T23:06:44.2477609Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"4(@)"}
+{"time":"2026-08-20T23:11:06.156880911Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"5(@)"}
+{"time":"2026-08-20T23:11:06.339048523Z","level":"INFO","msg":"handleInformInit: received","streamId":"yher6h2v","id":"5(@)"}
+{"time":"2026-08-20T23:11:06.578468942Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"yher6h2v","id":"5(@)"}
+{"time":"2026-08-20T23:11:13.028091133Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"be6uoaykyqwd"}
+{"time":"2026-08-20T23:16:12.40631562Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"be6uoaykyqwd"}
+{"time":"2026-08-20T23:16:13.170328682Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"be6uoaykyqwd"}
+{"time":"2026-08-20T23:16:13.171991708Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"yher6h2v","id":"5(@)"}
+{"time":"2026-08-20T23:16:13.172887025Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"yher6h2v","id":"5(@)"}
+{"time":"2026-08-20T23:16:13.990443666Z","level":"INFO","msg":"handleInformInit: received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:16:14.261935343Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:16:20.112222253Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:40.403706695Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.099860028Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"0x298lmexgzq"}
+{"time":"2026-08-20T23:21:41.101376929Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.10192683Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"z5ghx5k4","id":"5(@)"}
+{"time":"2026-08-20T23:21:41.842145532Z","level":"INFO","msg":"handleInformInit: received","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:42.103852099Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"62pb25xf","id":"5(@)"}
+{"time":"2026-08-20T23:21:47.827855381Z","level":"INFO","msg":"connection: cancelling request","id":"5(@)","requestId":"xrlr57kv5zv0"}

zain/Activation/wandb/run-20260820_232141-62pb25xf/logs/debug-internal.log

@@ -0,0 +1,29 @@
+{"time":"2026-08-20T23:21:41.84229624Z","level":"INFO","msg":"wandb-core"}
+{"time":"2026-08-20T23:21:41.842431914Z","level":"INFO","msg":"stream: starting","core version":"0.28.1"}
+{"time":"2026-08-20T23:21:42.103652466Z","level":"INFO","msg":"stream: created new stream","id":"62pb25xf"}
+{"time":"2026-08-20T23:21:42.103749367Z","level":"INFO","msg":"handler: started"}
+{"time":"2026-08-20T23:21:42.103842743Z","level":"INFO","msg":"stream: started"}
+{"time":"2026-08-20T23:21:42.103855547Z","level":"INFO","msg":"writer: started","stream_id":"62pb25xf"}
+{"time":"2026-08-20T23:21:42.10388216Z","level":"INFO","msg":"sender: started"}
+{"time":"2026-08-20T23:21:42.612690788Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1}
+{"time":"2026-08-20T23:21:42.737898167Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:21:57.612950032Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":0,"history_lines":8,"events_offset":0,"events_lines":1,"console_offset":0,"console_lines":14,"uploaded_len":2}
+{"time":"2026-08-20T23:21:57.820912206Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:22:12.613688Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":8,"history_lines":9,"events_offset":1,"events_lines":2,"console_offset":11,"console_lines":1}
+{"time":"2026-08-20T23:22:13.235818035Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:22:27.612957229Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":17,"history_lines":6,"events_offset":3,"events_lines":2,"console_offset":14,"console_lines":26}
+{"time":"2026-08-20T23:22:28.0548937Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:22:42.613459262Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":23,"history_lines":6,"events_offset":5,"events_lines":2,"console_offset":33,"console_lines":1}
+{"time":"2026-08-20T23:22:43.163167398Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:22:57.613501341Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":29,"history_lines":6,"events_offset":7,"events_lines":2,"console_offset":39,"console_lines":23}
+{"time":"2026-08-20T23:22:57.787457711Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:23:12.614756489Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":35,"history_lines":9,"events_offset":9,"events_lines":2,"console_offset":55,"console_lines":1}
+{"time":"2026-08-20T23:23:12.910989461Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:23:27.613285577Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":44,"history_lines":8,"events_offset":11,"events_lines":2,"console_offset":61,"console_lines":32}
+{"time":"2026-08-20T23:23:27.704621158Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:23:42.61384811Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":52,"history_lines":7,"events_offset":13,"events_lines":2,"console_offset":88,"console_lines":1}
+{"time":"2026-08-20T23:23:43.160977529Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:23:57.613567795Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":59,"history_lines":6,"events_offset":15,"events_lines":2,"console_offset":93,"console_lines":24}
+{"time":"2026-08-20T23:23:57.908977513Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-08-20T23:24:12.613516868Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":65,"history_lines":6,"events_offset":17,"events_lines":2,"console_offset":110,"console_lines":1}
+{"time":"2026-08-20T23:24:12.938411406Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}

zain/Activation/wandb/run-20260820_232141-62pb25xf/logs/debug.log

@@ -0,0 +1,20 @@
+2026-08-20 23:21:41,840 INFO    MainThread:58046 [wandb_init.py:setup_run_log_directory():729] Logging user logs to /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/wandb/run-20260820_232141-62pb25xf/logs/debug.log
+2026-08-20 23:21:41,840 INFO    MainThread:58046 [wandb_init.py:setup_run_log_directory():730] Logging internal logs to /mnt/data/zainulabideen/zain-exp/notebooks/zain/Activation/wandb/run-20260820_232141-62pb25xf/logs/debug-internal.log
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():772] calling init triggers
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():777] wandb.init called with sweep_config: {}
+config: {'_wandb': {}}
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():820] starting backend
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():826] Connected to an existing wandb-core service via WANDB_SERVICE
+2026-08-20 23:21:41,841 INFO    MainThread:58046 [wandb_init.py:init():835] sending inform_init request
+2026-08-20 23:21:42,104 INFO    MainThread:58046 [wandb_init.py:init():840] backend started and connected
+2026-08-20 23:21:42,105 INFO    MainThread:58046 [wandb_init.py:init():910] updated telemetry
+2026-08-20 23:21:42,112 INFO    MainThread:58046 [wandb_init.py:init():933] communicating run to backend with 90.0 second timeout
+2026-08-20 23:21:42,530 INFO    MainThread:58046 [wandb_init.py:init():978] starting run threads in backend
+2026-08-20 23:21:42,607 INFO    MainThread:58046 [wandb_run.py:_console_start():2621] atexit reg
+2026-08-20 23:21:42,607 INFO    MainThread:58046 [wandb_run.py:_redirect():2471] redirect: wrap_raw
+2026-08-20 23:21:42,607 INFO    MainThread:58046 [wandb_run.py:_redirect():2540] Wrapping output streams.
+2026-08-20 23:21:42,607 INFO    MainThread:58046 [wandb_run.py:_redirect():2563] Redirects installed.
+2026-08-20 23:21:42,608 INFO    MainThread:58046 [wandb_init.py:init():1016] run started, returning control to user process
+2026-08-20 23:21:42,609 INFO    MainThread:58046 [wandb_run.py:_config_callback():1346] config_cb None None {'transformers_version': '5.16.0.dev0', 'architectures': None, 'output_hidden_states': False, 'return_dict': True, 'dtype': None, 'chunk_size_feed_forward': 0, 'is_encoder_decoder': False, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'problem_type': None, 'vocab_size': 4096, 'hidden_size': 128, 'intermediate_size': 256, 'num_hidden_layers': 9, 'num_attention_heads': 4, 'num_key_value_heads': 4, 'hidden_act': 'silu', 'max_position_embeddings': 512, 'initializer_range': 0.02, 'rms_norm_eps': 1e-06, 'use_cache': False, 'pad_token_id': 0, 'bos_token_id': 1, 'eos_token_id': 2, 'pretraining_tp': 1, 'tie_word_embeddings': True, 'rope_parameters': {'rope_theta': 10000.0, 'rope_type': 'default'}, 'attention_bias': False, 'attention_dropout': 0.0, 'mlp_bias': False, 'head_dim': 32, '_name_or_path': '', 'tokenizer_name': 'w-ahmad/tiny-stories-tokenizer', 'mlp_type': 'mlp', 'activation': 'squared_relu', 'waleed_beta': 10.0, 'powlu_m': 3.0, 'model_type': 'tiny_llama', 'output_attentions': False, 'output_dir': 'out/mlp-squared_relu-9L_run', 'per_device_train_batch_size': 32, 'num_train_epochs': 1, 'max_steps': 2500, 'learning_rate': 0.0003, 'lr_scheduler_type': 'constant_with_warmup', 'lr_scheduler_kwargs': None, 'warmup_steps': 500, 'optim': 'adamw_torch_fused', 'optim_args': None, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'optim_target_modules': None, 'gradient_accumulation_steps': 1, 'average_tokens_across_devices': True, 'max_grad_norm': 1.0, 'label_smoothing_factor': 0.0, 'bf16': True, 'fp16': False, 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'gradient_checkpointing': False, 'gradient_checkpointing_kwargs': None, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'use_liger_kernel': False, 'liger_kernel_config': None, 'neftune_noise_alpha': None, 'torch_empty_cache_steps': None, 'auto_find_batch_size': False, 'logging_strategy': 'steps', 'logging_steps': 20, 'logging_first_step': False, 'log_on_each_node': True, 'logging_nan_inf_filter': True, 'include_num_input_tokens_seen': 'no', 'log_level': 'passive', 'log_level_replica': 'warning', 'disable_tqdm': False, 'report_to': ['wandb'], 'run_name': 'LM-mlp-squared_relu-9L-2.0M-20260820-232141', 'project': 'huggingface', 'trackio_space_id': None, 'trackio_bucket_id': None, 'trackio_static_space_id': None, 'eval_strategy': 'steps', 'eval_steps': 100, 'eval_delay': 0, 'per_device_eval_batch_size': 1500, 'prediction_loss_only': False, 'eval_on_start': False, 'eval_do_concat_batches': True, 'eval_use_gather_object': False, 'eval_accumulation_steps': None, 'include_for_metrics': [], 'batch_eval_metrics': False, 'save_only_model': False, 'save_strategy': 'steps', 'save_steps': 100, 'save_on_each_node': False, 'save_total_limit': None, 'enable_jit_checkpoint': False, 'push_to_hub': False, 'hub_token': '<HUB_TOKEN>', 'hub_private_repo': None, 'hub_model_id': 'w-ahmad/6L-mlp-squared_relu-9L', 'hub_strategy': 'every_save', 'hub_always_push': False, 'hub_revision': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'restore_callback_states_from_checkpoint': False, 'full_determinism': False, 'seed': 42, 'data_seed': 42, 'use_cpu': False, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}, 'parallelism_config': None, 'dataloader_drop_last': False, 'dataloader_num_workers': 0, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'dataloader_prefetch_factor': None, 'dataloader_multiprocessing_context': None, 'dataloader_in_order': True, 'remove_unused_columns': False, 'label_names': None, 'train_sampling_strategy': 'random', 'length_column_name': 'length', 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'ddp_static_graph': None, 'ddp_backend': None, 'ddp_timeout': 1800, 'fsdp': None, 'fsdp_config': None, 'deepspeed': None, 'debug': [], 'skip_memory_metrics': True, 'do_train': False, 'do_eval': True, 'do_predict': False, 'resume_from_checkpoint': None, 'local_rank': -1}
+2026-08-20 23:21:42,611 INFO    MainThread:58046 [wandb_config.py:__setitem__():155] [no run ID] config set model/num_parameters = 2001280 - <bound method Run._config_callback of <wandb.sdk.wandb_run.Run object at 0x1526c05aeb50>>
+2026-08-20 23:21:42,611 INFO    MainThread:58046 [wandb_run.py:_config_callback():1346] config_cb model/num_parameters 2001280 None

zain/Activation/wandb/run-20260820_232141-62pb25xf/run-62pb25xf.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e2f595192bce4574d0de338db4c5843798fcc7ce4dfd53331c704e9fbac86cf
+size 229376