| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 20.0, | |
| "eval_steps": 500, | |
| "global_step": 1100, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 4.545454545454545e-07, | |
| "loss": 2.7297, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 9.090909090909091e-06, | |
| "loss": 2.7244, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.8181818181818182e-05, | |
| "loss": 2.7671, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "gpt4_scores": 0.43333333333333335, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 2.312265396118164, | |
| "eval_runtime": 4.9266, | |
| "eval_samples_per_second": 4.669, | |
| "eval_steps_per_second": 1.218, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 2.7272727272727273e-05, | |
| "loss": 2.4172, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 3.6363636363636364e-05, | |
| "loss": 2.1566, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 4.545454545454546e-05, | |
| "loss": 2.0319, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "gpt4_scores": 0.6, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.9679468870162964, | |
| "eval_runtime": 4.9521, | |
| "eval_samples_per_second": 4.644, | |
| "eval_steps_per_second": 1.212, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 4.9987413559579636e-05, | |
| "loss": 1.8348, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 4.988679806432712e-05, | |
| "loss": 1.8492, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "learning_rate": 4.968597221690986e-05, | |
| "loss": 1.7972, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "gpt4_scores": 0.5166666666666666, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 1.942632794380188, | |
| "eval_runtime": 4.9711, | |
| "eval_samples_per_second": 4.627, | |
| "eval_steps_per_second": 1.207, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 3.27, | |
| "learning_rate": 4.938574467213518e-05, | |
| "loss": 1.6588, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 4.898732434036244e-05, | |
| "loss": 1.6187, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 4.849231551964771e-05, | |
| "loss": 1.5841, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "gpt4_scores": 0.6833333333333332, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 2.0110199451446533, | |
| "eval_runtime": 4.9304, | |
| "eval_samples_per_second": 4.665, | |
| "eval_steps_per_second": 1.217, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "learning_rate": 4.790271143580174e-05, | |
| "loss": 1.2998, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 4.73, | |
| "learning_rate": 4.722088621637309e-05, | |
| "loss": 1.2842, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "gpt4_scores": 0.5166666666666667, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 2.2670648097991943, | |
| "eval_runtime": 4.9304, | |
| "eval_samples_per_second": 4.665, | |
| "eval_steps_per_second": 1.217, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 5.09, | |
| "learning_rate": 4.644958533087443e-05, | |
| "loss": 1.1442, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "learning_rate": 4.559191453574582e-05, | |
| "loss": 0.9076, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 5.82, | |
| "learning_rate": 4.465132736856969e-05, | |
| "loss": 0.9305, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "gpt4_scores": 0.75, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 2.5263493061065674, | |
| "eval_runtime": 4.9267, | |
| "eval_samples_per_second": 4.668, | |
| "eval_steps_per_second": 1.218, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 6.18, | |
| "learning_rate": 4.3631611241893874e-05, | |
| "loss": 0.7708, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 6.55, | |
| "learning_rate": 4.2536872192658036e-05, | |
| "loss": 0.6432, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 6.91, | |
| "learning_rate": 4.137151834863213e-05, | |
| "loss": 0.6734, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "gpt4_scores": 0.7333333333333334, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 2.7797744274139404, | |
| "eval_runtime": 4.982, | |
| "eval_samples_per_second": 4.617, | |
| "eval_steps_per_second": 1.204, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 7.27, | |
| "learning_rate": 4.014024217844167e-05, | |
| "loss": 0.5284, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 7.64, | |
| "learning_rate": 3.884800159665276e-05, | |
| "loss": 0.4544, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 0.4579, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "gpt4_scores": 0.5499999999999999, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 3.1051599979400635, | |
| "eval_runtime": 4.9596, | |
| "eval_samples_per_second": 4.637, | |
| "eval_steps_per_second": 1.21, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 8.36, | |
| "learning_rate": 3.610166531514436e-05, | |
| "loss": 0.3129, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 8.73, | |
| "learning_rate": 3.465862814232822e-05, | |
| "loss": 0.3091, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "gpt4_scores": 0.5666666666666668, | |
| "step": 495 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 3.3408806324005127, | |
| "eval_runtime": 4.9659, | |
| "eval_samples_per_second": 4.632, | |
| "eval_steps_per_second": 1.208, | |
| "step": 495 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "learning_rate": 3.3176699082935545e-05, | |
| "loss": 0.3082, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 9.45, | |
| "learning_rate": 3.166184534225087e-05, | |
| "loss": 0.2184, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 9.82, | |
| "learning_rate": 3.012016670162977e-05, | |
| "loss": 0.2418, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "gpt4_scores": 0.48333333333333334, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 3.499851703643799, | |
| "eval_runtime": 4.9663, | |
| "eval_samples_per_second": 4.631, | |
| "eval_steps_per_second": 1.208, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 10.18, | |
| "learning_rate": 2.8557870956832132e-05, | |
| "loss": 0.1996, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 10.55, | |
| "learning_rate": 2.698124892141971e-05, | |
| "loss": 0.1778, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 10.91, | |
| "learning_rate": 2.5396649095870202e-05, | |
| "loss": 0.1718, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "gpt4_scores": 0.6333333333333334, | |
| "step": 605 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_loss": 3.6687815189361572, | |
| "eval_runtime": 4.9484, | |
| "eval_samples_per_second": 4.648, | |
| "eval_steps_per_second": 1.213, | |
| "step": 605 | |
| }, | |
| { | |
| "epoch": 11.27, | |
| "learning_rate": 2.3810452104406444e-05, | |
| "loss": 0.1401, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 11.64, | |
| "learning_rate": 2.222904500247473e-05, | |
| "loss": 0.1344, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 2.0658795558326743e-05, | |
| "loss": 0.1555, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "gpt4_scores": 0.6, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 3.78193998336792, | |
| "eval_runtime": 4.9817, | |
| "eval_samples_per_second": 4.617, | |
| "eval_steps_per_second": 1.204, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 12.36, | |
| "learning_rate": 1.9106026612264316e-05, | |
| "loss": 0.1414, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 12.73, | |
| "learning_rate": 1.7576990616793137e-05, | |
| "loss": 0.1191, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "gpt4_scores": 0.3666666666666667, | |
| "step": 715 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_loss": 3.910775899887085, | |
| "eval_runtime": 4.9709, | |
| "eval_samples_per_second": 4.627, | |
| "eval_steps_per_second": 1.207, | |
| "step": 715 | |
| }, | |
| { | |
| "epoch": 13.09, | |
| "learning_rate": 1.6077844460203206e-05, | |
| "loss": 0.1191, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 13.45, | |
| "learning_rate": 1.4614624674952842e-05, | |
| "loss": 0.1069, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 13.82, | |
| "learning_rate": 1.3193223130682936e-05, | |
| "loss": 0.1291, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "gpt4_scores": 0.3666666666666667, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 3.995321750640869, | |
| "eval_runtime": 4.9833, | |
| "eval_samples_per_second": 4.615, | |
| "eval_steps_per_second": 1.204, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 14.18, | |
| "learning_rate": 1.181936330973744e-05, | |
| "loss": 0.1075, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 14.55, | |
| "learning_rate": 1.049857726072005e-05, | |
| "loss": 0.1058, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 14.91, | |
| "learning_rate": 9.236183322886945e-06, | |
| "loss": 0.1213, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "gpt4_scores": 0.3333333333333333, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_loss": 4.102020740509033, | |
| "eval_runtime": 4.931, | |
| "eval_samples_per_second": 4.664, | |
| "eval_steps_per_second": 1.217, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 15.27, | |
| "learning_rate": 8.0372647110717e-06, | |
| "loss": 0.1013, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 15.64, | |
| "learning_rate": 6.906649047373246e-06, | |
| "loss": 0.1176, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "learning_rate": 5.848888922025553e-06, | |
| "loss": 0.1, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "gpt4_scores": 0.43333333333333335, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 4.120510578155518, | |
| "eval_runtime": 4.9308, | |
| "eval_samples_per_second": 4.665, | |
| "eval_steps_per_second": 1.217, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 16.36, | |
| "learning_rate": 4.868243561723535e-06, | |
| "loss": 0.0974, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 16.73, | |
| "learning_rate": 3.968661679220468e-06, | |
| "loss": 0.115, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "gpt4_scores": 0.45, | |
| "step": 935 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_loss": 4.160642623901367, | |
| "eval_runtime": 4.9259, | |
| "eval_samples_per_second": 4.669, | |
| "eval_steps_per_second": 1.218, | |
| "step": 935 | |
| }, | |
| { | |
| "epoch": 17.09, | |
| "learning_rate": 3.1537655732553768e-06, | |
| "loss": 0.1013, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 17.45, | |
| "learning_rate": 2.4268365428344736e-06, | |
| "loss": 0.0971, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 17.82, | |
| "learning_rate": 1.790801674598186e-06, | |
| "loss": 0.1076, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "gpt4_scores": 0.31666666666666665, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_loss": 4.183867454528809, | |
| "eval_runtime": 4.9845, | |
| "eval_samples_per_second": 4.614, | |
| "eval_steps_per_second": 1.204, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 18.18, | |
| "learning_rate": 1.248222056476367e-06, | |
| "loss": 0.103, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 18.55, | |
| "learning_rate": 8.012824650910938e-07, | |
| "loss": 0.1082, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 18.91, | |
| "learning_rate": 4.517825684323324e-07, | |
| "loss": 0.0962, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "gpt4_scores": 0.4666666666666666, | |
| "step": 1045 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_loss": 4.187291622161865, | |
| "eval_runtime": 4.9682, | |
| "eval_samples_per_second": 4.629, | |
| "eval_steps_per_second": 1.208, | |
| "step": 1045 | |
| }, | |
| { | |
| "epoch": 19.27, | |
| "learning_rate": 2.011296792301165e-07, | |
| "loss": 0.0981, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 19.64, | |
| "learning_rate": 5.033308820289184e-08, | |
| "loss": 0.1154, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.0917, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "gpt4_scores": 0.31666666666666665, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_loss": 4.187403678894043, | |
| "eval_runtime": 4.9884, | |
| "eval_samples_per_second": 4.611, | |
| "eval_steps_per_second": 1.203, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 1100, | |
| "total_flos": 3.836907908090266e+16, | |
| "train_loss": 0.660087760145014, | |
| "train_runtime": 8612.1752, | |
| "train_samples_per_second": 0.504, | |
| "train_steps_per_second": 0.128 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 1100, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 20, | |
| "total_flos": 3.836907908090266e+16, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |