| { | |
| "best_global_step": 120, | |
| "best_metric": 0.7120494842529297, | |
| "best_model_checkpoint": "/dss/dssfs05/pn39qo/pn39qo-dss-0001/tong/efficient_reasoning/extraction-vs-summary-efficient-cot-reasoning-perspective---Experiment-main/output/lora/Limo_qwen/checkpoint-120", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 120, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.16842105263157894, | |
| "grad_norm": 0.3850134313106537, | |
| "learning_rate": 1.3333333333333333e-05, | |
| "loss": 0.9614, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.3368421052631579, | |
| "grad_norm": 0.38739946484565735, | |
| "learning_rate": 4e-05, | |
| "loss": 0.9707, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.5052631578947369, | |
| "grad_norm": 0.3434392809867859, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.9543, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.6736842105263158, | |
| "grad_norm": 0.1770918369293213, | |
| "learning_rate": 7.998481228099806e-05, | |
| "loss": 0.865, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.8421052631578947, | |
| "grad_norm": 0.2255554050207138, | |
| "learning_rate": 7.98633797202668e-05, | |
| "loss": 0.8631, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.274989515542984, | |
| "learning_rate": 7.962088338550013e-05, | |
| "loss": 0.8842, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.8996532559394836, | |
| "eval_runtime": 12.9673, | |
| "eval_samples_per_second": 3.085, | |
| "eval_steps_per_second": 0.771, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.168421052631579, | |
| "grad_norm": 0.2877059280872345, | |
| "learning_rate": 7.925805973009672e-05, | |
| "loss": 0.856, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.3368421052631578, | |
| "grad_norm": 0.17793452739715576, | |
| "learning_rate": 7.877601063757323e-05, | |
| "loss": 0.8322, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 1.5052631578947369, | |
| "grad_norm": 0.09469418972730637, | |
| "learning_rate": 7.81762000751803e-05, | |
| "loss": 0.8178, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 1.6736842105263157, | |
| "grad_norm": 0.09645849466323853, | |
| "learning_rate": 7.74604496478822e-05, | |
| "loss": 0.8212, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.8421052631578947, | |
| "grad_norm": 0.11077257990837097, | |
| "learning_rate": 7.663093306620231e-05, | |
| "loss": 0.798, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.10621998459100723, | |
| "learning_rate": 7.569016954473577e-05, | |
| "loss": 0.8086, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.8223039507865906, | |
| "eval_runtime": 12.9571, | |
| "eval_samples_per_second": 3.087, | |
| "eval_steps_per_second": 0.772, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.168421052631579, | |
| "grad_norm": 0.09538944810628891, | |
| "learning_rate": 7.464101615137756e-05, | |
| "loss": 0.7815, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 2.336842105263158, | |
| "grad_norm": 0.07264512777328491, | |
| "learning_rate": 7.348665913050115e-05, | |
| "loss": 0.7909, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 2.5052631578947366, | |
| "grad_norm": 0.0889253318309784, | |
| "learning_rate": 7.223060422643914e-05, | |
| "loss": 0.795, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 2.6736842105263157, | |
| "grad_norm": 0.09207943081855774, | |
| "learning_rate": 7.087666603665284e-05, | |
| "loss": 0.7529, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 2.8421052631578947, | |
| "grad_norm": 0.07369054853916168, | |
| "learning_rate": 6.942895642692527e-05, | |
| "loss": 0.7329, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.06927139312028885, | |
| "learning_rate": 6.789187204375981e-05, | |
| "loss": 0.7502, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.7780648469924927, | |
| "eval_runtime": 12.9544, | |
| "eval_samples_per_second": 3.088, | |
| "eval_steps_per_second": 0.772, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 3.168421052631579, | |
| "grad_norm": 0.0669899433851242, | |
| "learning_rate": 6.627008096190938e-05, | |
| "loss": 0.7365, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 3.336842105263158, | |
| "grad_norm": 0.06613507866859436, | |
| "learning_rate": 6.456850850758673e-05, | |
| "loss": 0.7316, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 3.5052631578947366, | |
| "grad_norm": 0.06987571716308594, | |
| "learning_rate": 6.279232230041065e-05, | |
| "loss": 0.742, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 3.6736842105263157, | |
| "grad_norm": 0.057263512164354324, | |
| "learning_rate": 6.094691655951512e-05, | |
| "loss": 0.7204, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 3.8421052631578947, | |
| "grad_norm": 0.061604950577020645, | |
| "learning_rate": 5.903789572148295e-05, | |
| "loss": 0.7258, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.06501278281211853, | |
| "learning_rate": 5.707105741985615e-05, | |
| "loss": 0.7287, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.7513701319694519, | |
| "eval_runtime": 12.9656, | |
| "eval_samples_per_second": 3.085, | |
| "eval_steps_per_second": 0.771, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 4.168421052631579, | |
| "grad_norm": 0.0571233369410038, | |
| "learning_rate": 5.505237487791343e-05, | |
| "loss": 0.7132, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 4.336842105263158, | |
| "grad_norm": 0.059772882610559464, | |
| "learning_rate": 5.298797876818735e-05, | |
| "loss": 0.7349, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 4.505263157894737, | |
| "grad_norm": 0.05535353720188141, | |
| "learning_rate": 5.088413859381341e-05, | |
| "loss": 0.708, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 4.673684210526316, | |
| "grad_norm": 0.058156561106443405, | |
| "learning_rate": 4.874724364825504e-05, | |
| "loss": 0.7166, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 4.842105263157895, | |
| "grad_norm": 0.05242394655942917, | |
| "learning_rate": 4.658378361122936e-05, | |
| "loss": 0.6791, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.053081415593624115, | |
| "learning_rate": 4.440032883976318e-05, | |
| "loss": 0.6899, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.7341214418411255, | |
| "eval_runtime": 12.9486, | |
| "eval_samples_per_second": 3.089, | |
| "eval_steps_per_second": 0.772, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 5.168421052631579, | |
| "grad_norm": 0.05485621094703674, | |
| "learning_rate": 4.220351041423462e-05, | |
| "loss": 0.695, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 5.336842105263158, | |
| "grad_norm": 0.047592103481292725, | |
| "learning_rate": 4e-05, | |
| "loss": 0.696, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 5.505263157894737, | |
| "grad_norm": 0.05209505185484886, | |
| "learning_rate": 3.779648958576538e-05, | |
| "loss": 0.6842, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 5.673684210526316, | |
| "grad_norm": 0.05388766899704933, | |
| "learning_rate": 3.559967116023683e-05, | |
| "loss": 0.6967, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 5.842105263157895, | |
| "grad_norm": 0.05102350562810898, | |
| "learning_rate": 3.341621638877064e-05, | |
| "loss": 0.6816, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.059541136026382446, | |
| "learning_rate": 3.125275635174497e-05, | |
| "loss": 0.6934, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.7228327989578247, | |
| "eval_runtime": 12.9513, | |
| "eval_samples_per_second": 3.088, | |
| "eval_steps_per_second": 0.772, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 6.168421052631579, | |
| "grad_norm": 0.04430772364139557, | |
| "learning_rate": 2.9115861406186593e-05, | |
| "loss": 0.6659, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 6.336842105263158, | |
| "grad_norm": 0.04723000526428223, | |
| "learning_rate": 2.7012021231812666e-05, | |
| "loss": 0.6659, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 6.505263157894737, | |
| "grad_norm": 0.05059608444571495, | |
| "learning_rate": 2.4947625122086585e-05, | |
| "loss": 0.6888, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 6.673684210526316, | |
| "grad_norm": 0.05666281282901764, | |
| "learning_rate": 2.2928942580143855e-05, | |
| "loss": 0.6868, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 6.842105263157895, | |
| "grad_norm": 0.05316559597849846, | |
| "learning_rate": 2.096210427851706e-05, | |
| "loss": 0.7081, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 0.05505692958831787, | |
| "learning_rate": 1.9053083440484887e-05, | |
| "loss": 0.6727, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 0.7168187499046326, | |
| "eval_runtime": 12.9563, | |
| "eval_samples_per_second": 3.087, | |
| "eval_steps_per_second": 0.772, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 7.168421052631579, | |
| "grad_norm": 0.053103502839803696, | |
| "learning_rate": 1.7207677699589355e-05, | |
| "loss": 0.6816, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 7.336842105263158, | |
| "grad_norm": 0.046842310577631, | |
| "learning_rate": 1.5431491492413288e-05, | |
| "loss": 0.6642, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 7.505263157894737, | |
| "grad_norm": 0.04811316728591919, | |
| "learning_rate": 1.3729919038090627e-05, | |
| "loss": 0.6729, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 7.673684210526316, | |
| "grad_norm": 0.04840261861681938, | |
| "learning_rate": 1.2108127956240186e-05, | |
| "loss": 0.6697, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 7.842105263157895, | |
| "grad_norm": 0.048225287348032, | |
| "learning_rate": 1.0571043573074737e-05, | |
| "loss": 0.6757, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.05675382539629936, | |
| "learning_rate": 9.123333963347166e-06, | |
| "loss": 0.69, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.7133689522743225, | |
| "eval_runtime": 12.9684, | |
| "eval_samples_per_second": 3.084, | |
| "eval_steps_per_second": 0.771, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 8.16842105263158, | |
| "grad_norm": 0.04558368772268295, | |
| "learning_rate": 7.769395773560874e-06, | |
| "loss": 0.6674, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 8.336842105263157, | |
| "grad_norm": 0.049284905195236206, | |
| "learning_rate": 6.513340869498859e-06, | |
| "loss": 0.6833, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 8.505263157894737, | |
| "grad_norm": 0.04926304891705513, | |
| "learning_rate": 5.358983848622452e-06, | |
| "loss": 0.6743, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 8.673684210526316, | |
| "grad_norm": 0.04950882866978645, | |
| "learning_rate": 4.3098304552642385e-06, | |
| "loss": 0.6636, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 8.842105263157894, | |
| "grad_norm": 0.049746621400117874, | |
| "learning_rate": 3.3690669337977e-06, | |
| "loss": 0.6609, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 0.04946954548358917, | |
| "learning_rate": 2.5395503521178143e-06, | |
| "loss": 0.6892, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 0.7123732566833496, | |
| "eval_runtime": 12.9488, | |
| "eval_samples_per_second": 3.089, | |
| "eval_steps_per_second": 0.772, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 9.16842105263158, | |
| "grad_norm": 0.04948243126273155, | |
| "learning_rate": 1.8237999248197002e-06, | |
| "loss": 0.6742, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 9.336842105263157, | |
| "grad_norm": 0.05210984870791435, | |
| "learning_rate": 1.2239893624267852e-06, | |
| "loss": 0.6949, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 9.505263157894737, | |
| "grad_norm": 0.047761935740709305, | |
| "learning_rate": 7.419402699032852e-07, | |
| "loss": 0.6467, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 9.673684210526316, | |
| "grad_norm": 0.04957466945052147, | |
| "learning_rate": 3.791166144998704e-07, | |
| "loss": 0.6643, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 9.842105263157894, | |
| "grad_norm": 0.0482342354953289, | |
| "learning_rate": 1.3662027973320614e-07, | |
| "loss": 0.6764, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.04675516113638878, | |
| "learning_rate": 1.5187719001943378e-08, | |
| "loss": 0.6735, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.7120494842529297, | |
| "eval_runtime": 12.9487, | |
| "eval_samples_per_second": 3.089, | |
| "eval_steps_per_second": 0.772, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 120, | |
| "total_flos": 3.766644395416224e+18, | |
| "train_loss": 0.7336713840564092, | |
| "train_runtime": 9734.8369, | |
| "train_samples_per_second": 0.781, | |
| "train_steps_per_second": 0.012 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 120, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.766644395416224e+18, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |