| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 29280, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 1.801757574081421, | |
| "learning_rate": 6.25e-05, | |
| "loss": 6.1993, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 1.014549970626831, | |
| "learning_rate": 0.000125, | |
| "loss": 3.6289, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 1.0354454517364502, | |
| "learning_rate": 0.0001875, | |
| "loss": 3.3439, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 0.8663645386695862, | |
| "learning_rate": 0.00025, | |
| "loss": 3.1569, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 0.9717487692832947, | |
| "learning_rate": 0.0003125, | |
| "loss": 3.0413, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.4380028648723519, | |
| "eval_loss": 3.010645866394043, | |
| "eval_runtime": 3.3126, | |
| "eval_samples_per_second": 1355.734, | |
| "eval_steps_per_second": 10.868, | |
| "step": 2928 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "grad_norm": 0.8488046526908875, | |
| "learning_rate": 0.000375, | |
| "loss": 2.9511, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 0.8335849046707153, | |
| "learning_rate": 0.00043750000000000006, | |
| "loss": 2.8601, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 0.8527885675430298, | |
| "learning_rate": 0.0005, | |
| "loss": 2.8103, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 0.8196571469306946, | |
| "learning_rate": 0.0005625000000000001, | |
| "loss": 2.7757, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 0.7712026238441467, | |
| "learning_rate": 0.000625, | |
| "loss": 2.7326, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 0.7557732462882996, | |
| "learning_rate": 0.0006875, | |
| "loss": 2.706, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.4614636096339661, | |
| "eval_loss": 2.7756640911102295, | |
| "eval_runtime": 3.3754, | |
| "eval_samples_per_second": 1330.505, | |
| "eval_steps_per_second": 10.665, | |
| "step": 5856 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 0.7290401458740234, | |
| "learning_rate": 0.00075, | |
| "loss": 2.6636, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "grad_norm": 0.7280641794204712, | |
| "learning_rate": 0.0008125, | |
| "loss": 2.6292, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "grad_norm": 0.7040073871612549, | |
| "learning_rate": 0.0008750000000000001, | |
| "loss": 2.6193, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "grad_norm": 0.6477857828140259, | |
| "learning_rate": 0.0009375, | |
| "loss": 2.5962, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "grad_norm": 0.6744890213012695, | |
| "learning_rate": 0.001, | |
| "loss": 2.587, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 0.567164957523346, | |
| "learning_rate": 0.0010625, | |
| "loss": 2.5796, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.4708138937542627, | |
| "eval_loss": 2.6893413066864014, | |
| "eval_runtime": 3.4141, | |
| "eval_samples_per_second": 1315.426, | |
| "eval_steps_per_second": 10.544, | |
| "step": 8784 | |
| }, | |
| { | |
| "epoch": 3.07, | |
| "grad_norm": 0.6116344332695007, | |
| "learning_rate": 0.0011250000000000001, | |
| "loss": 2.5478, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.24, | |
| "grad_norm": 0.6196800470352173, | |
| "learning_rate": 0.0011875, | |
| "loss": 2.5082, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "grad_norm": 0.5813851356506348, | |
| "learning_rate": 0.00125, | |
| "loss": 2.5135, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.59, | |
| "grad_norm": 0.510899543762207, | |
| "learning_rate": 0.0013125, | |
| "loss": 2.5229, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 3.76, | |
| "grad_norm": 0.4730781614780426, | |
| "learning_rate": 0.001375, | |
| "loss": 2.5106, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 3.93, | |
| "grad_norm": 0.4921322166919708, | |
| "learning_rate": 0.0014375000000000002, | |
| "loss": 2.5174, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.4766470824413481, | |
| "eval_loss": 2.634284019470215, | |
| "eval_runtime": 3.4269, | |
| "eval_samples_per_second": 1310.502, | |
| "eval_steps_per_second": 10.505, | |
| "step": 11712 | |
| }, | |
| { | |
| "epoch": 4.1, | |
| "grad_norm": 0.4425903558731079, | |
| "learning_rate": 0.0015, | |
| "loss": 2.4712, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 4.27, | |
| "grad_norm": 0.43356040120124817, | |
| "learning_rate": 0.0015625, | |
| "loss": 2.4604, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "grad_norm": 0.4289402663707733, | |
| "learning_rate": 0.001625, | |
| "loss": 2.4623, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 4.61, | |
| "grad_norm": 0.4142000675201416, | |
| "learning_rate": 0.0016875, | |
| "loss": 2.4576, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 4.78, | |
| "grad_norm": 0.36214813590049744, | |
| "learning_rate": 0.0017500000000000003, | |
| "loss": 2.4678, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "grad_norm": 0.365395188331604, | |
| "learning_rate": 0.0018124999999999999, | |
| "loss": 2.4609, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.47951370808107907, | |
| "eval_loss": 2.6072192192077637, | |
| "eval_runtime": 3.4065, | |
| "eval_samples_per_second": 1318.36, | |
| "eval_steps_per_second": 10.568, | |
| "step": 14640 | |
| }, | |
| { | |
| "epoch": 5.12, | |
| "grad_norm": 0.36923331022262573, | |
| "learning_rate": 0.001875, | |
| "loss": 2.4109, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 5.29, | |
| "grad_norm": 0.337181031703949, | |
| "learning_rate": 0.0019375000000000002, | |
| "loss": 2.4102, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 5.46, | |
| "grad_norm": 0.39215803146362305, | |
| "learning_rate": 0.002, | |
| "loss": 2.4135, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 5.64, | |
| "grad_norm": 0.35966119170188904, | |
| "learning_rate": 0.0020625, | |
| "loss": 2.4291, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 5.81, | |
| "grad_norm": 0.4037434458732605, | |
| "learning_rate": 0.002125, | |
| "loss": 2.4236, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 5.98, | |
| "grad_norm": 0.3071579039096832, | |
| "learning_rate": 0.0021874999999999998, | |
| "loss": 2.4299, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.4825889749753225, | |
| "eval_loss": 2.585719108581543, | |
| "eval_runtime": 3.4287, | |
| "eval_samples_per_second": 1309.825, | |
| "eval_steps_per_second": 10.5, | |
| "step": 17568 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "grad_norm": 0.3319147527217865, | |
| "learning_rate": 0.0022500000000000003, | |
| "loss": 2.3562, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 6.32, | |
| "grad_norm": 0.35867124795913696, | |
| "learning_rate": 0.0023125000000000003, | |
| "loss": 2.3702, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 6.49, | |
| "grad_norm": 0.3402807116508484, | |
| "learning_rate": 0.002375, | |
| "loss": 2.3783, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 6.66, | |
| "grad_norm": 0.33656784892082214, | |
| "learning_rate": 0.0024375, | |
| "loss": 2.3905, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 6.83, | |
| "grad_norm": 0.38668304681777954, | |
| "learning_rate": 0.0025, | |
| "loss": 2.3999, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.4859254817596698, | |
| "eval_loss": 2.5638771057128906, | |
| "eval_runtime": 3.4265, | |
| "eval_samples_per_second": 1310.666, | |
| "eval_steps_per_second": 10.506, | |
| "step": 20496 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 0.2988421320915222, | |
| "learning_rate": 0.0025625, | |
| "loss": 2.4033, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 7.17, | |
| "grad_norm": 0.34959354996681213, | |
| "learning_rate": 0.002625, | |
| "loss": 2.3211, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 7.34, | |
| "grad_norm": 0.3704765737056732, | |
| "learning_rate": 0.0026875000000000002, | |
| "loss": 2.3397, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 7.51, | |
| "grad_norm": 0.3484719693660736, | |
| "learning_rate": 0.00275, | |
| "loss": 2.3566, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 7.68, | |
| "grad_norm": 0.3266860246658325, | |
| "learning_rate": 0.0028125, | |
| "loss": 2.3694, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "grad_norm": 0.3507145941257477, | |
| "learning_rate": 0.0028750000000000004, | |
| "loss": 2.3645, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.4874315560254367, | |
| "eval_loss": 2.5552921295166016, | |
| "eval_runtime": 3.3792, | |
| "eval_samples_per_second": 1329.004, | |
| "eval_steps_per_second": 10.653, | |
| "step": 23424 | |
| }, | |
| { | |
| "epoch": 8.03, | |
| "grad_norm": 0.3075660467147827, | |
| "learning_rate": 0.0029375, | |
| "loss": 2.3579, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 8.2, | |
| "grad_norm": 0.32643941044807434, | |
| "learning_rate": 0.003, | |
| "loss": 2.2952, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 8.37, | |
| "grad_norm": 0.31350839138031006, | |
| "learning_rate": 0.002715909090909091, | |
| "loss": 2.3125, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 8.54, | |
| "grad_norm": 0.36248230934143066, | |
| "learning_rate": 0.0024318181818181817, | |
| "loss": 2.3118, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 8.71, | |
| "grad_norm": 0.27529385685920715, | |
| "learning_rate": 0.002147727272727273, | |
| "loss": 2.3016, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 8.88, | |
| "grad_norm": 0.27803343534469604, | |
| "learning_rate": 0.0018636363636363638, | |
| "loss": 2.2938, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.4938678757339701, | |
| "eval_loss": 2.5048580169677734, | |
| "eval_runtime": 3.376, | |
| "eval_samples_per_second": 1330.268, | |
| "eval_steps_per_second": 10.663, | |
| "step": 26352 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "grad_norm": 0.33036813139915466, | |
| "learning_rate": 0.0015795454545454546, | |
| "loss": 2.2526, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 9.22, | |
| "grad_norm": 0.29800865054130554, | |
| "learning_rate": 0.0012954545454545456, | |
| "loss": 2.176, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 9.39, | |
| "grad_norm": 0.26307207345962524, | |
| "learning_rate": 0.0010113636363636364, | |
| "loss": 2.1712, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 9.56, | |
| "grad_norm": 0.2642175257205963, | |
| "learning_rate": 0.0007272727272727273, | |
| "loss": 2.1735, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 9.73, | |
| "grad_norm": 0.268196702003479, | |
| "learning_rate": 0.0004431818181818182, | |
| "loss": 2.1658, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 9.9, | |
| "grad_norm": 0.2505429983139038, | |
| "learning_rate": 0.0001590909090909091, | |
| "loss": 2.151, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.499566061256371, | |
| "eval_loss": 2.4763824939727783, | |
| "eval_runtime": 3.4206, | |
| "eval_samples_per_second": 1312.914, | |
| "eval_steps_per_second": 10.524, | |
| "step": 29280 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 29280, | |
| "total_flos": 2273237316403200.0, | |
| "train_loss": 2.5702311156225988, | |
| "train_runtime": 763.0136, | |
| "train_samples_per_second": 613.947, | |
| "train_steps_per_second": 38.374 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 29280, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 2000, | |
| "total_flos": 2273237316403200.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |