| { | |
| "best_metric": 0.4170692563056946, | |
| "best_model_checkpoint": "mikhail-panzo/zlm-fil_b64_le5_s8000/checkpoint-3000", | |
| "epoch": 133.33333333333334, | |
| "eval_steps": 500, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 2.2222222222222223, | |
| "grad_norm": 4.803515434265137, | |
| "learning_rate": 2.5000000000000004e-07, | |
| "loss": 0.8968, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 4.444444444444445, | |
| "grad_norm": 2.6444506645202637, | |
| "learning_rate": 5.000000000000001e-07, | |
| "loss": 0.8539, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 6.666666666666667, | |
| "grad_norm": 1.9820537567138672, | |
| "learning_rate": 7.5e-07, | |
| "loss": 0.7856, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 8.88888888888889, | |
| "grad_norm": 1.2031773328781128, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 0.7306, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 11.11111111111111, | |
| "grad_norm": 1.728398084640503, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.6821, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 13.333333333333334, | |
| "grad_norm": 1.232366681098938, | |
| "learning_rate": 1.5e-06, | |
| "loss": 0.6348, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 15.555555555555555, | |
| "grad_norm": 1.4039260149002075, | |
| "learning_rate": 1.75e-06, | |
| "loss": 0.5962, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 17.77777777777778, | |
| "grad_norm": 1.1848849058151245, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 0.5758, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 0.9365352392196655, | |
| "learning_rate": 2.25e-06, | |
| "loss": 0.5636, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 22.22222222222222, | |
| "grad_norm": 0.9893051981925964, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.5529, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 22.22222222222222, | |
| "eval_loss": 0.4999743402004242, | |
| "eval_runtime": 6.8701, | |
| "eval_samples_per_second": 23.144, | |
| "eval_steps_per_second": 2.911, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 24.444444444444443, | |
| "grad_norm": 0.7535696029663086, | |
| "learning_rate": 2.7500000000000004e-06, | |
| "loss": 0.5476, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 26.666666666666668, | |
| "grad_norm": 0.7487667202949524, | |
| "learning_rate": 3e-06, | |
| "loss": 0.536, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 28.88888888888889, | |
| "grad_norm": 0.9480589628219604, | |
| "learning_rate": 3.2500000000000002e-06, | |
| "loss": 0.5271, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 31.11111111111111, | |
| "grad_norm": 1.0027987957000732, | |
| "learning_rate": 3.5e-06, | |
| "loss": 0.5202, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 33.333333333333336, | |
| "grad_norm": 0.7353075742721558, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.5128, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 35.55555555555556, | |
| "grad_norm": 1.2223567962646484, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.5159, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 37.77777777777778, | |
| "grad_norm": 0.835096001625061, | |
| "learning_rate": 4.25e-06, | |
| "loss": 0.5051, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "grad_norm": 0.8878389000892639, | |
| "learning_rate": 4.5e-06, | |
| "loss": 0.5044, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 42.22222222222222, | |
| "grad_norm": 0.8464574813842773, | |
| "learning_rate": 4.75e-06, | |
| "loss": 0.4984, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 44.44444444444444, | |
| "grad_norm": 0.8119956254959106, | |
| "learning_rate": 5e-06, | |
| "loss": 0.4974, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 44.44444444444444, | |
| "eval_loss": 0.4557439386844635, | |
| "eval_runtime": 6.6345, | |
| "eval_samples_per_second": 23.965, | |
| "eval_steps_per_second": 3.015, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 46.666666666666664, | |
| "grad_norm": 1.2427846193313599, | |
| "learning_rate": 5.2500000000000006e-06, | |
| "loss": 0.4943, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 48.888888888888886, | |
| "grad_norm": 0.9691639542579651, | |
| "learning_rate": 5.500000000000001e-06, | |
| "loss": 0.4879, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 51.111111111111114, | |
| "grad_norm": 0.815934419631958, | |
| "learning_rate": 5.75e-06, | |
| "loss": 0.4825, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 53.333333333333336, | |
| "grad_norm": 1.090553641319275, | |
| "learning_rate": 6e-06, | |
| "loss": 0.4875, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 55.55555555555556, | |
| "grad_norm": 1.090770959854126, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.4821, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 57.77777777777778, | |
| "grad_norm": 1.2514219284057617, | |
| "learning_rate": 6.5000000000000004e-06, | |
| "loss": 0.4752, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 60.0, | |
| "grad_norm": 1.0580047369003296, | |
| "learning_rate": 6.750000000000001e-06, | |
| "loss": 0.4749, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 62.22222222222222, | |
| "grad_norm": 1.2864220142364502, | |
| "learning_rate": 7e-06, | |
| "loss": 0.4727, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 64.44444444444444, | |
| "grad_norm": 1.0031774044036865, | |
| "learning_rate": 7.25e-06, | |
| "loss": 0.4706, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 66.66666666666667, | |
| "grad_norm": 0.7133861184120178, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.4716, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 66.66666666666667, | |
| "eval_loss": 0.43592509627342224, | |
| "eval_runtime": 6.6514, | |
| "eval_samples_per_second": 23.905, | |
| "eval_steps_per_second": 3.007, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 68.88888888888889, | |
| "grad_norm": 0.8695252537727356, | |
| "learning_rate": 7.75e-06, | |
| "loss": 0.464, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 71.11111111111111, | |
| "grad_norm": 2.2705039978027344, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.4677, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 73.33333333333333, | |
| "grad_norm": 0.8814280033111572, | |
| "learning_rate": 8.25e-06, | |
| "loss": 0.4628, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 75.55555555555556, | |
| "grad_norm": 0.6686401963233948, | |
| "learning_rate": 8.5e-06, | |
| "loss": 0.4642, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 77.77777777777777, | |
| "grad_norm": 1.1037653684616089, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 0.4582, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 80.0, | |
| "grad_norm": 1.0339733362197876, | |
| "learning_rate": 9e-06, | |
| "loss": 0.4595, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 82.22222222222223, | |
| "grad_norm": 0.715323269367218, | |
| "learning_rate": 9.250000000000001e-06, | |
| "loss": 0.4621, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 84.44444444444444, | |
| "grad_norm": 0.872926652431488, | |
| "learning_rate": 9.5e-06, | |
| "loss": 0.4553, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 86.66666666666667, | |
| "grad_norm": 0.8996181488037109, | |
| "learning_rate": 9.75e-06, | |
| "loss": 0.4528, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 88.88888888888889, | |
| "grad_norm": 1.3289371728897095, | |
| "learning_rate": 1e-05, | |
| "loss": 0.453, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 88.88888888888889, | |
| "eval_loss": 0.4245700240135193, | |
| "eval_runtime": 6.6886, | |
| "eval_samples_per_second": 23.772, | |
| "eval_steps_per_second": 2.99, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 91.11111111111111, | |
| "grad_norm": 0.9035242795944214, | |
| "learning_rate": 9.916666666666668e-06, | |
| "loss": 0.4504, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 93.33333333333333, | |
| "grad_norm": 1.091410756111145, | |
| "learning_rate": 9.833333333333333e-06, | |
| "loss": 0.4528, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 95.55555555555556, | |
| "grad_norm": 0.7232839465141296, | |
| "learning_rate": 9.75e-06, | |
| "loss": 0.4481, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 97.77777777777777, | |
| "grad_norm": 0.9006613492965698, | |
| "learning_rate": 9.666666666666667e-06, | |
| "loss": 0.444, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 100.0, | |
| "grad_norm": 0.9088640809059143, | |
| "learning_rate": 9.583333333333335e-06, | |
| "loss": 0.4445, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 102.22222222222223, | |
| "grad_norm": 1.2409088611602783, | |
| "learning_rate": 9.5e-06, | |
| "loss": 0.4457, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 104.44444444444444, | |
| "grad_norm": 0.749260663986206, | |
| "learning_rate": 9.416666666666667e-06, | |
| "loss": 0.4453, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 106.66666666666667, | |
| "grad_norm": 0.7552911639213562, | |
| "learning_rate": 9.333333333333334e-06, | |
| "loss": 0.4402, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 108.88888888888889, | |
| "grad_norm": 1.0751075744628906, | |
| "learning_rate": 9.250000000000001e-06, | |
| "loss": 0.4417, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 111.11111111111111, | |
| "grad_norm": 0.8774231672286987, | |
| "learning_rate": 9.166666666666666e-06, | |
| "loss": 0.4428, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 111.11111111111111, | |
| "eval_loss": 0.41958874464035034, | |
| "eval_runtime": 6.7008, | |
| "eval_samples_per_second": 23.728, | |
| "eval_steps_per_second": 2.985, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 113.33333333333333, | |
| "grad_norm": 1.3239281177520752, | |
| "learning_rate": 9.083333333333333e-06, | |
| "loss": 0.4399, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 115.55555555555556, | |
| "grad_norm": 0.9855658411979675, | |
| "learning_rate": 9e-06, | |
| "loss": 0.4376, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 117.77777777777777, | |
| "grad_norm": 0.9498355984687805, | |
| "learning_rate": 8.916666666666667e-06, | |
| "loss": 0.4427, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 120.0, | |
| "grad_norm": 1.3080662488937378, | |
| "learning_rate": 8.833333333333334e-06, | |
| "loss": 0.437, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 122.22222222222223, | |
| "grad_norm": 0.6745712757110596, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 0.4405, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 124.44444444444444, | |
| "grad_norm": 1.136643409729004, | |
| "learning_rate": 8.666666666666668e-06, | |
| "loss": 0.4329, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 126.66666666666667, | |
| "grad_norm": 0.8538027405738831, | |
| "learning_rate": 8.583333333333333e-06, | |
| "loss": 0.4313, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 128.88888888888889, | |
| "grad_norm": 0.8776795268058777, | |
| "learning_rate": 8.5e-06, | |
| "loss": 0.43, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 131.11111111111111, | |
| "grad_norm": 0.987410306930542, | |
| "learning_rate": 8.416666666666667e-06, | |
| "loss": 0.4346, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 133.33333333333334, | |
| "grad_norm": 1.061345100402832, | |
| "learning_rate": 8.333333333333334e-06, | |
| "loss": 0.4332, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 133.33333333333334, | |
| "eval_loss": 0.4170692563056946, | |
| "eval_runtime": 6.8806, | |
| "eval_samples_per_second": 23.108, | |
| "eval_steps_per_second": 2.907, | |
| "step": 3000 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 8000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 364, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.364575967899699e+16, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |