| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 200, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2, | |
| "eval_loss": 0.05579125136137009, | |
| "eval_runtime": 37.7203, | |
| "eval_samples_per_second": 424.174, | |
| "eval_steps_per_second": 6.628, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_loss": 0.19435200095176697, | |
| "eval_runtime": 37.6839, | |
| "eval_samples_per_second": 424.584, | |
| "eval_steps_per_second": 6.634, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 148661.390625, | |
| "learning_rate": 4.75e-05, | |
| "loss": 0.2826, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_loss": 0.3969748318195343, | |
| "eval_runtime": 37.6897, | |
| "eval_samples_per_second": 424.519, | |
| "eval_steps_per_second": 6.633, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_loss": 0.6244608759880066, | |
| "eval_runtime": 37.6841, | |
| "eval_samples_per_second": 424.583, | |
| "eval_steps_per_second": 6.634, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 5462.595703125, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.8928, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 2.0544872283935547, | |
| "eval_runtime": 37.666, | |
| "eval_samples_per_second": 424.787, | |
| "eval_steps_per_second": 6.637, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "eval_loss": 0.3789268136024475, | |
| "eval_runtime": 37.8068, | |
| "eval_samples_per_second": 423.204, | |
| "eval_steps_per_second": 6.613, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "eval_loss": 0.41200506687164307, | |
| "eval_runtime": 37.7019, | |
| "eval_samples_per_second": 424.381, | |
| "eval_steps_per_second": 6.631, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 26773808.0, | |
| "learning_rate": 4.25e-05, | |
| "loss": 0.5735, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_loss": 0.9738002419471741, | |
| "eval_runtime": 37.6855, | |
| "eval_samples_per_second": 424.567, | |
| "eval_steps_per_second": 6.634, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "eval_loss": 1.4284050464630127, | |
| "eval_runtime": 37.676, | |
| "eval_samples_per_second": 424.673, | |
| "eval_steps_per_second": 6.636, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1449586.25, | |
| "learning_rate": 4e-05, | |
| "loss": 3.2584, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 3.862783193588257, | |
| "eval_runtime": 37.86, | |
| "eval_samples_per_second": 422.609, | |
| "eval_steps_per_second": 6.603, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_loss": 0.6802809834480286, | |
| "eval_runtime": 37.6736, | |
| "eval_samples_per_second": 424.701, | |
| "eval_steps_per_second": 6.636, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "eval_loss": 0.41678962111473083, | |
| "eval_runtime": 37.886, | |
| "eval_samples_per_second": 422.319, | |
| "eval_steps_per_second": 6.599, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 3005.915771484375, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 1.1454, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "eval_loss": 0.0627516433596611, | |
| "eval_runtime": 37.6977, | |
| "eval_samples_per_second": 424.429, | |
| "eval_steps_per_second": 6.632, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "eval_loss": 0.03529125079512596, | |
| "eval_runtime": 37.6892, | |
| "eval_samples_per_second": 424.525, | |
| "eval_steps_per_second": 6.633, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.08225303888320923, | |
| "learning_rate": 3.5e-05, | |
| "loss": 0.0693, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.030110161751508713, | |
| "eval_runtime": 37.7252, | |
| "eval_samples_per_second": 424.12, | |
| "eval_steps_per_second": 6.627, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "eval_loss": 0.029404468834400177, | |
| "eval_runtime": 37.6821, | |
| "eval_samples_per_second": 424.605, | |
| "eval_steps_per_second": 6.634, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "eval_loss": 0.028393320739269257, | |
| "eval_runtime": 37.9071, | |
| "eval_samples_per_second": 422.084, | |
| "eval_steps_per_second": 6.595, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 3.5, | |
| "grad_norm": 0.11493222415447235, | |
| "learning_rate": 3.2500000000000004e-05, | |
| "loss": 0.0299, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "eval_loss": 0.02786264941096306, | |
| "eval_runtime": 37.6844, | |
| "eval_samples_per_second": 424.578, | |
| "eval_steps_per_second": 6.634, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "eval_loss": 0.027447132393717766, | |
| "eval_runtime": 37.6815, | |
| "eval_samples_per_second": 424.611, | |
| "eval_steps_per_second": 6.635, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.09786217659711838, | |
| "learning_rate": 3e-05, | |
| "loss": 0.0287, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.02736382931470871, | |
| "eval_runtime": 37.6836, | |
| "eval_samples_per_second": 424.588, | |
| "eval_steps_per_second": 6.634, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.2, | |
| "eval_loss": 0.02714085392653942, | |
| "eval_runtime": 37.7033, | |
| "eval_samples_per_second": 424.366, | |
| "eval_steps_per_second": 6.631, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "eval_loss": 0.026028065010905266, | |
| "eval_runtime": 37.7142, | |
| "eval_samples_per_second": 424.243, | |
| "eval_steps_per_second": 6.629, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "grad_norm": 0.10962454974651337, | |
| "learning_rate": 2.7500000000000004e-05, | |
| "loss": 0.0274, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 4.6, | |
| "eval_loss": 0.026036322116851807, | |
| "eval_runtime": 37.9005, | |
| "eval_samples_per_second": 422.157, | |
| "eval_steps_per_second": 6.596, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 4.8, | |
| "eval_loss": 0.026090338826179504, | |
| "eval_runtime": 37.8184, | |
| "eval_samples_per_second": 423.074, | |
| "eval_steps_per_second": 6.611, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.15162990987300873, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.0267, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.025744769722223282, | |
| "eval_runtime": 37.9279, | |
| "eval_samples_per_second": 421.853, | |
| "eval_steps_per_second": 6.591, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.2, | |
| "eval_loss": 0.025514526292681694, | |
| "eval_runtime": 37.8847, | |
| "eval_samples_per_second": 422.334, | |
| "eval_steps_per_second": 6.599, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 5.4, | |
| "eval_loss": 0.025529591366648674, | |
| "eval_runtime": 37.6892, | |
| "eval_samples_per_second": 424.525, | |
| "eval_steps_per_second": 6.633, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 5.5, | |
| "grad_norm": 0.05882571265101433, | |
| "learning_rate": 2.25e-05, | |
| "loss": 0.0263, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 5.6, | |
| "eval_loss": 0.02538459375500679, | |
| "eval_runtime": 37.6706, | |
| "eval_samples_per_second": 424.735, | |
| "eval_steps_per_second": 6.636, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 5.8, | |
| "eval_loss": 0.02504335343837738, | |
| "eval_runtime": 37.6763, | |
| "eval_samples_per_second": 424.671, | |
| "eval_steps_per_second": 6.635, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.09895586967468262, | |
| "learning_rate": 2e-05, | |
| "loss": 0.0259, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.02500820718705654, | |
| "eval_runtime": 37.6737, | |
| "eval_samples_per_second": 424.699, | |
| "eval_steps_per_second": 6.636, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.2, | |
| "eval_loss": 0.025201652199029922, | |
| "eval_runtime": 37.9114, | |
| "eval_samples_per_second": 422.036, | |
| "eval_steps_per_second": 6.594, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 6.4, | |
| "eval_loss": 0.02534925751388073, | |
| "eval_runtime": 37.6963, | |
| "eval_samples_per_second": 424.445, | |
| "eval_steps_per_second": 6.632, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "grad_norm": 0.1827753484249115, | |
| "learning_rate": 1.75e-05, | |
| "loss": 0.0256, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 6.6, | |
| "eval_loss": 0.02500862069427967, | |
| "eval_runtime": 37.8861, | |
| "eval_samples_per_second": 422.318, | |
| "eval_steps_per_second": 6.599, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 6.8, | |
| "eval_loss": 0.024682270362973213, | |
| "eval_runtime": 37.762, | |
| "eval_samples_per_second": 423.706, | |
| "eval_steps_per_second": 6.62, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 0.08037039637565613, | |
| "learning_rate": 1.5e-05, | |
| "loss": 0.0253, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 0.025552883744239807, | |
| "eval_runtime": 37.6888, | |
| "eval_samples_per_second": 424.53, | |
| "eval_steps_per_second": 6.633, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.2, | |
| "eval_loss": 0.024695081636309624, | |
| "eval_runtime": 37.8117, | |
| "eval_samples_per_second": 423.15, | |
| "eval_steps_per_second": 6.612, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 7.4, | |
| "eval_loss": 0.024455612525343895, | |
| "eval_runtime": 37.8335, | |
| "eval_samples_per_second": 422.906, | |
| "eval_steps_per_second": 6.608, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "grad_norm": 0.12335552275180817, | |
| "learning_rate": 1.25e-05, | |
| "loss": 0.0251, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 7.6, | |
| "eval_loss": 0.02451241761445999, | |
| "eval_runtime": 37.6804, | |
| "eval_samples_per_second": 424.624, | |
| "eval_steps_per_second": 6.635, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 7.8, | |
| "eval_loss": 0.024500148370862007, | |
| "eval_runtime": 37.6703, | |
| "eval_samples_per_second": 424.738, | |
| "eval_steps_per_second": 6.637, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.18342430889606476, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0251, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.02456432767212391, | |
| "eval_runtime": 37.918, | |
| "eval_samples_per_second": 421.963, | |
| "eval_steps_per_second": 6.593, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.2, | |
| "eval_loss": 0.02438100427389145, | |
| "eval_runtime": 37.7187, | |
| "eval_samples_per_second": 424.193, | |
| "eval_steps_per_second": 6.628, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 8.4, | |
| "eval_loss": 0.02455132268369198, | |
| "eval_runtime": 37.679, | |
| "eval_samples_per_second": 424.64, | |
| "eval_steps_per_second": 6.635, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 8.5, | |
| "grad_norm": 0.09001246094703674, | |
| "learning_rate": 7.5e-06, | |
| "loss": 0.0252, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 8.6, | |
| "eval_loss": 0.024252494797110558, | |
| "eval_runtime": 37.731, | |
| "eval_samples_per_second": 424.054, | |
| "eval_steps_per_second": 6.626, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 8.8, | |
| "eval_loss": 0.02421058714389801, | |
| "eval_runtime": 37.6831, | |
| "eval_samples_per_second": 424.593, | |
| "eval_steps_per_second": 6.634, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 0.07999344915151596, | |
| "learning_rate": 5e-06, | |
| "loss": 0.0244, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 0.024233995005488396, | |
| "eval_runtime": 37.8728, | |
| "eval_samples_per_second": 422.467, | |
| "eval_steps_per_second": 6.601, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.2, | |
| "eval_loss": 0.02422364056110382, | |
| "eval_runtime": 37.7417, | |
| "eval_samples_per_second": 423.934, | |
| "eval_steps_per_second": 6.624, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 9.4, | |
| "eval_loss": 0.024163929745554924, | |
| "eval_runtime": 37.7865, | |
| "eval_samples_per_second": 423.432, | |
| "eval_steps_per_second": 6.616, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 9.5, | |
| "grad_norm": 0.10827407240867615, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.0247, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 9.6, | |
| "eval_loss": 0.024232398718595505, | |
| "eval_runtime": 37.8161, | |
| "eval_samples_per_second": 423.1, | |
| "eval_steps_per_second": 6.611, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 9.8, | |
| "eval_loss": 0.02414529025554657, | |
| "eval_runtime": 37.6993, | |
| "eval_samples_per_second": 424.411, | |
| "eval_steps_per_second": 6.631, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.13070951402187347, | |
| "learning_rate": 0.0, | |
| "loss": 0.0245, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.024126721546053886, | |
| "eval_runtime": 37.7053, | |
| "eval_samples_per_second": 424.344, | |
| "eval_steps_per_second": 6.63, | |
| "step": 10000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 10000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 1000, | |
| "total_flos": 4.180672512e+16, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |