| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 50.0, | |
| "eval_steps": 500, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 74.0, | |
| "learning_rate": 1.3333333333333333e-05, | |
| "loss": 17.168, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 15.087593078613281, | |
| "eval_runtime": 1.3097, | |
| "eval_samples_per_second": 1.527, | |
| "eval_steps_per_second": 0.764, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "grad_norm": 67.5, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 14.9207, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 8.764444351196289, | |
| "eval_runtime": 1.3178, | |
| "eval_samples_per_second": 1.518, | |
| "eval_steps_per_second": 0.759, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 4.842519760131836, | |
| "eval_runtime": 1.321, | |
| "eval_samples_per_second": 1.514, | |
| "eval_steps_per_second": 0.757, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "grad_norm": 15.875, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 7.3214, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 3.023890495300293, | |
| "eval_runtime": 1.3205, | |
| "eval_samples_per_second": 1.515, | |
| "eval_steps_per_second": 0.757, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 0.0002, | |
| "loss": 2.9627, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 2.256518602371216, | |
| "eval_runtime": 1.3203, | |
| "eval_samples_per_second": 1.515, | |
| "eval_steps_per_second": 0.757, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 1.8792051076889038, | |
| "eval_runtime": 1.3249, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 6.67, | |
| "grad_norm": 1.640625, | |
| "learning_rate": 0.00019932383577419432, | |
| "loss": 1.7971, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 1.7647607326507568, | |
| "eval_runtime": 1.324, | |
| "eval_samples_per_second": 1.511, | |
| "eval_steps_per_second": 0.755, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 1.7012015581130981, | |
| "eval_runtime": 1.3248, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "grad_norm": 2.234375, | |
| "learning_rate": 0.00019730448705798239, | |
| "loss": 1.4939, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 1.547886610031128, | |
| "eval_runtime": 1.3259, | |
| "eval_samples_per_second": 1.508, | |
| "eval_steps_per_second": 0.754, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 1.78125, | |
| "learning_rate": 0.00019396926207859084, | |
| "loss": 1.2756, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.5050724744796753, | |
| "eval_runtime": 1.3222, | |
| "eval_samples_per_second": 1.513, | |
| "eval_steps_per_second": 0.756, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_loss": 1.3975391387939453, | |
| "eval_runtime": 1.3256, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.754, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 11.67, | |
| "grad_norm": 1.1796875, | |
| "learning_rate": 0.00018936326403234125, | |
| "loss": 1.0884, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 1.444014549255371, | |
| "eval_runtime": 1.325, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.755, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_loss": 1.413475513458252, | |
| "eval_runtime": 1.3258, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.754, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "grad_norm": 1.3046875, | |
| "learning_rate": 0.00018354878114129367, | |
| "loss": 0.9429, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.4587093591690063, | |
| "eval_runtime": 1.3247, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 2.21875, | |
| "learning_rate": 0.0001766044443118978, | |
| "loss": 0.7653, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_loss": 1.487448811531067, | |
| "eval_runtime": 1.3231, | |
| "eval_samples_per_second": 1.512, | |
| "eval_steps_per_second": 0.756, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.5958000421524048, | |
| "eval_runtime": 1.3242, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 16.67, | |
| "grad_norm": 0.9921875, | |
| "learning_rate": 0.0001686241637868734, | |
| "loss": 0.6424, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_loss": 1.592842698097229, | |
| "eval_runtime": 1.3242, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_loss": 1.683807373046875, | |
| "eval_runtime": 1.3252, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.755, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 18.33, | |
| "grad_norm": 4.78125, | |
| "learning_rate": 0.00015971585917027862, | |
| "loss": 0.5346, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_loss": 1.8263951539993286, | |
| "eval_runtime": 1.3248, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 0.99609375, | |
| "learning_rate": 0.00015000000000000001, | |
| "loss": 0.4249, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_loss": 1.9654637575149536, | |
| "eval_runtime": 1.3233, | |
| "eval_samples_per_second": 1.511, | |
| "eval_steps_per_second": 0.756, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_loss": 2.137000560760498, | |
| "eval_runtime": 1.3247, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 21.67, | |
| "grad_norm": 1.484375, | |
| "learning_rate": 0.0001396079766039157, | |
| "loss": 0.3347, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_loss": 2.698075294494629, | |
| "eval_runtime": 1.3255, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.754, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_loss": 2.713052272796631, | |
| "eval_runtime": 1.3259, | |
| "eval_samples_per_second": 1.508, | |
| "eval_steps_per_second": 0.754, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 23.33, | |
| "grad_norm": 1.71875, | |
| "learning_rate": 0.00012868032327110904, | |
| "loss": 0.2655, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_loss": 2.7668490409851074, | |
| "eval_runtime": 1.3252, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.755, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 0.734375, | |
| "learning_rate": 0.00011736481776669306, | |
| "loss": 0.2026, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_loss": 2.8614566326141357, | |
| "eval_runtime": 1.3226, | |
| "eval_samples_per_second": 1.512, | |
| "eval_steps_per_second": 0.756, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_loss": 3.1595633029937744, | |
| "eval_runtime": 1.3267, | |
| "eval_samples_per_second": 1.507, | |
| "eval_steps_per_second": 0.754, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 26.67, | |
| "grad_norm": 0.73828125, | |
| "learning_rate": 0.00010581448289104758, | |
| "loss": 0.1588, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_loss": 3.3285796642303467, | |
| "eval_runtime": 1.3243, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_loss": 3.546278953552246, | |
| "eval_runtime": 1.3248, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 28.33, | |
| "grad_norm": 1.015625, | |
| "learning_rate": 9.418551710895243e-05, | |
| "loss": 0.1319, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_loss": 3.3686463832855225, | |
| "eval_runtime": 1.3245, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "grad_norm": 0.56640625, | |
| "learning_rate": 8.263518223330697e-05, | |
| "loss": 0.1111, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_loss": 3.685863733291626, | |
| "eval_runtime": 1.3215, | |
| "eval_samples_per_second": 1.513, | |
| "eval_steps_per_second": 0.757, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "eval_loss": 3.780993700027466, | |
| "eval_runtime": 1.3257, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.754, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 31.67, | |
| "grad_norm": 0.4375, | |
| "learning_rate": 7.131967672889101e-05, | |
| "loss": 0.0939, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_loss": 3.7559256553649902, | |
| "eval_runtime": 1.3244, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "eval_loss": 3.916355848312378, | |
| "eval_runtime": 1.3252, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.755, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "grad_norm": 0.390625, | |
| "learning_rate": 6.039202339608432e-05, | |
| "loss": 0.082, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_loss": 3.9693491458892822, | |
| "eval_runtime": 1.3217, | |
| "eval_samples_per_second": 1.513, | |
| "eval_steps_per_second": 0.757, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "grad_norm": 0.283203125, | |
| "learning_rate": 5.000000000000002e-05, | |
| "loss": 0.0709, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "eval_loss": 4.04301118850708, | |
| "eval_runtime": 1.3186, | |
| "eval_samples_per_second": 1.517, | |
| "eval_steps_per_second": 0.758, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "eval_loss": 4.101677417755127, | |
| "eval_runtime": 1.3238, | |
| "eval_samples_per_second": 1.511, | |
| "eval_steps_per_second": 0.755, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 36.67, | |
| "grad_norm": 0.27734375, | |
| "learning_rate": 4.028414082972141e-05, | |
| "loss": 0.0638, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "eval_loss": 4.144949913024902, | |
| "eval_runtime": 1.3248, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "eval_loss": 4.1638994216918945, | |
| "eval_runtime": 1.3218, | |
| "eval_samples_per_second": 1.513, | |
| "eval_steps_per_second": 0.757, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 38.33, | |
| "grad_norm": 0.3125, | |
| "learning_rate": 3.137583621312665e-05, | |
| "loss": 0.0597, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 39.0, | |
| "eval_loss": 4.187974452972412, | |
| "eval_runtime": 1.3246, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "grad_norm": 0.318359375, | |
| "learning_rate": 2.339555568810221e-05, | |
| "loss": 0.0556, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "eval_loss": 4.212304592132568, | |
| "eval_runtime": 1.3229, | |
| "eval_samples_per_second": 1.512, | |
| "eval_steps_per_second": 0.756, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 41.0, | |
| "eval_loss": 4.219560623168945, | |
| "eval_runtime": 1.3269, | |
| "eval_samples_per_second": 1.507, | |
| "eval_steps_per_second": 0.754, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 41.67, | |
| "grad_norm": 0.201171875, | |
| "learning_rate": 1.6451218858706374e-05, | |
| "loss": 0.0535, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 42.0, | |
| "eval_loss": 4.226192951202393, | |
| "eval_runtime": 1.3243, | |
| "eval_samples_per_second": 1.51, | |
| "eval_steps_per_second": 0.755, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 43.0, | |
| "eval_loss": 4.230067729949951, | |
| "eval_runtime": 1.3252, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.755, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 43.33, | |
| "grad_norm": 0.2490234375, | |
| "learning_rate": 1.0636735967658784e-05, | |
| "loss": 0.0521, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 44.0, | |
| "eval_loss": 4.231449604034424, | |
| "eval_runtime": 1.3262, | |
| "eval_samples_per_second": 1.508, | |
| "eval_steps_per_second": 0.754, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 45.0, | |
| "grad_norm": 0.10986328125, | |
| "learning_rate": 6.030737921409169e-06, | |
| "loss": 0.0521, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 45.0, | |
| "eval_loss": 4.2365241050720215, | |
| "eval_runtime": 1.3238, | |
| "eval_samples_per_second": 1.511, | |
| "eval_steps_per_second": 0.755, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 46.0, | |
| "eval_loss": 4.234958171844482, | |
| "eval_runtime": 1.3255, | |
| "eval_samples_per_second": 1.509, | |
| "eval_steps_per_second": 0.754, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 46.67, | |
| "grad_norm": 0.12451171875, | |
| "learning_rate": 2.6955129420176196e-06, | |
| "loss": 0.0525, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 47.0, | |
| "eval_loss": 4.236445426940918, | |
| "eval_runtime": 1.3261, | |
| "eval_samples_per_second": 1.508, | |
| "eval_steps_per_second": 0.754, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 48.0, | |
| "eval_loss": 4.231955051422119, | |
| "eval_runtime": 1.3237, | |
| "eval_samples_per_second": 1.511, | |
| "eval_steps_per_second": 0.755, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 48.33, | |
| "grad_norm": 0.1357421875, | |
| "learning_rate": 6.761642258056978e-07, | |
| "loss": 0.0509, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 49.0, | |
| "eval_loss": 4.236112594604492, | |
| "eval_runtime": 1.3264, | |
| "eval_samples_per_second": 1.508, | |
| "eval_steps_per_second": 0.754, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "grad_norm": 0.09326171875, | |
| "learning_rate": 0.0, | |
| "loss": 0.0505, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "eval_loss": 4.238922595977783, | |
| "eval_runtime": 1.3223, | |
| "eval_samples_per_second": 1.513, | |
| "eval_steps_per_second": 0.756, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "step": 150, | |
| "total_flos": 4.601366279314473e+17, | |
| "train_loss": 1.2187181319793066, | |
| "train_runtime": 1336.0597, | |
| "train_samples_per_second": 3.256, | |
| "train_steps_per_second": 0.112 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 150, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 100, | |
| "total_flos": 4.601366279314473e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |