| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984, | |
| "eval_steps": 500, | |
| "global_step": 156, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 24.144679616293203, | |
| "learning_rate": 6.25e-09, | |
| "logits/chosen": -0.4835050106048584, | |
| "logits/rejected": -0.45789963006973267, | |
| "logps/chosen": -214.22390747070312, | |
| "logps/rejected": -238.45899963378906, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 27.157833222439553, | |
| "learning_rate": 6.25e-08, | |
| "logits/chosen": -0.5191683769226074, | |
| "logits/rejected": -0.6054410338401794, | |
| "logps/chosen": -233.09689331054688, | |
| "logps/rejected": -247.0145721435547, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.4097222089767456, | |
| "rewards/chosen": -0.00034871429670602083, | |
| "rewards/margins": -0.0003790514019783586, | |
| "rewards/rejected": 3.03371598420199e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 22.695219914328067, | |
| "learning_rate": 9.979871469976195e-08, | |
| "logits/chosen": -0.5110132098197937, | |
| "logits/rejected": -0.5159324407577515, | |
| "logps/chosen": -226.8810272216797, | |
| "logps/rejected": -234.03579711914062, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.484375, | |
| "rewards/chosen": -0.012425726279616356, | |
| "rewards/margins": 0.0005857773358002305, | |
| "rewards/rejected": -0.013011504895985126, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 27.267303752253287, | |
| "learning_rate": 9.755282581475768e-08, | |
| "logits/chosen": -0.46132326126098633, | |
| "logits/rejected": -0.5038495063781738, | |
| "logps/chosen": -228.7631378173828, | |
| "logps/rejected": -243.3440399169922, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.5718749761581421, | |
| "rewards/chosen": -0.05522267147898674, | |
| "rewards/margins": 0.011394877918064594, | |
| "rewards/rejected": -0.06661754846572876, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 25.80441755927495, | |
| "learning_rate": 9.29224396800933e-08, | |
| "logits/chosen": -0.5605762600898743, | |
| "logits/rejected": -0.4145810008049011, | |
| "logps/chosen": -235.9694366455078, | |
| "logps/rejected": -245.53140258789062, | |
| "loss": 0.6877, | |
| "rewards/accuracies": 0.5406249761581421, | |
| "rewards/chosen": -0.12484677881002426, | |
| "rewards/margins": 0.01826881244778633, | |
| "rewards/rejected": -0.1431155800819397, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 29.148756992059326, | |
| "learning_rate": 8.613974319136957e-08, | |
| "logits/chosen": -0.33988311886787415, | |
| "logits/rejected": -0.4071916937828064, | |
| "logps/chosen": -225.6988525390625, | |
| "logps/rejected": -248.31716918945312, | |
| "loss": 0.6858, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.11917146295309067, | |
| "rewards/margins": 0.03840441256761551, | |
| "rewards/rejected": -0.15757587552070618, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 31.691697738483036, | |
| "learning_rate": 7.754484907260513e-08, | |
| "logits/chosen": -0.32609015703201294, | |
| "logits/rejected": -0.26230502128601074, | |
| "logps/chosen": -227.6748504638672, | |
| "logps/rejected": -247.0984649658203, | |
| "loss": 0.6813, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.07656601071357727, | |
| "rewards/margins": 0.052247386425733566, | |
| "rewards/rejected": -0.12881340086460114, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 29.21753574540554, | |
| "learning_rate": 6.756874120406714e-08, | |
| "logits/chosen": -0.46613264083862305, | |
| "logits/rejected": -0.40840277075767517, | |
| "logps/chosen": -227.2427520751953, | |
| "logps/rejected": -237.88174438476562, | |
| "loss": 0.6836, | |
| "rewards/accuracies": 0.5406249761581421, | |
| "rewards/chosen": -0.011777873151004314, | |
| "rewards/margins": 0.03103628195822239, | |
| "rewards/rejected": -0.04281415045261383, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 27.295936995560456, | |
| "learning_rate": 5.6711663290882774e-08, | |
| "logits/chosen": -0.3539288640022278, | |
| "logits/rejected": -0.4821072518825531, | |
| "logps/chosen": -227.5625, | |
| "logps/rejected": -246.0811767578125, | |
| "loss": 0.6843, | |
| "rewards/accuracies": 0.5093749761581421, | |
| "rewards/chosen": -0.02285691350698471, | |
| "rewards/margins": 0.04252464324235916, | |
| "rewards/rejected": -0.06538156419992447, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 37.94611808985469, | |
| "learning_rate": 4.551803455482833e-08, | |
| "logits/chosen": -0.31488946080207825, | |
| "logits/rejected": -0.2586648762226105, | |
| "logps/chosen": -235.135009765625, | |
| "logps/rejected": -250.8152618408203, | |
| "loss": 0.6811, | |
| "rewards/accuracies": 0.559374988079071, | |
| "rewards/chosen": -0.05702267214655876, | |
| "rewards/margins": 0.03373400494456291, | |
| "rewards/rejected": -0.09075668454170227, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 31.89166999144619, | |
| "learning_rate": 3.4549150281252633e-08, | |
| "logits/chosen": -0.2342231571674347, | |
| "logits/rejected": -0.11592201143503189, | |
| "logps/chosen": -241.8380126953125, | |
| "logps/rejected": -248.9578094482422, | |
| "loss": 0.6851, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -0.10655052959918976, | |
| "rewards/margins": 0.02655043639242649, | |
| "rewards/rejected": -0.1331009566783905, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 33.31375799730263, | |
| "learning_rate": 2.43550361297047e-08, | |
| "logits/chosen": -0.18439307808876038, | |
| "logits/rejected": -0.2467001974582672, | |
| "logps/chosen": -231.55880737304688, | |
| "logps/rejected": -247.73422241210938, | |
| "loss": 0.6818, | |
| "rewards/accuracies": 0.503125011920929, | |
| "rewards/chosen": -0.1505533754825592, | |
| "rewards/margins": 0.03467424958944321, | |
| "rewards/rejected": -0.185227632522583, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 47.472733400866, | |
| "learning_rate": 1.5446867550656767e-08, | |
| "logits/chosen": -0.33943504095077515, | |
| "logits/rejected": -0.27855488657951355, | |
| "logps/chosen": -237.60281372070312, | |
| "logps/rejected": -256.1316833496094, | |
| "loss": 0.6833, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.1447935551404953, | |
| "rewards/margins": 0.04320163279771805, | |
| "rewards/rejected": -0.18799519538879395, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 33.56575623501931, | |
| "learning_rate": 8.271337313934867e-09, | |
| "logits/chosen": -0.20994436740875244, | |
| "logits/rejected": -0.30067163705825806, | |
| "logps/chosen": -242.28720092773438, | |
| "logps/rejected": -260.9054260253906, | |
| "loss": 0.6824, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.13135352730751038, | |
| "rewards/margins": 0.033290181308984756, | |
| "rewards/rejected": -0.16464371979236603, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 36.04452677999868, | |
| "learning_rate": 3.1882564680131396e-09, | |
| "logits/chosen": -0.2294171154499054, | |
| "logits/rejected": -0.14980368316173553, | |
| "logps/chosen": -226.87939453125, | |
| "logps/rejected": -243.36831665039062, | |
| "loss": 0.6884, | |
| "rewards/accuracies": 0.503125011920929, | |
| "rewards/chosen": -0.11503396183252335, | |
| "rewards/margins": 0.014261065050959587, | |
| "rewards/rejected": -0.12929502129554749, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 34.70896434984119, | |
| "learning_rate": 4.52511911603265e-10, | |
| "logits/chosen": -0.28860437870025635, | |
| "logits/rejected": -0.18086276948451996, | |
| "logps/chosen": -245.88034057617188, | |
| "logps/rejected": -255.14749145507812, | |
| "loss": 0.6783, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -0.12854784727096558, | |
| "rewards/margins": 0.03599141910672188, | |
| "rewards/rejected": -0.16453926265239716, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 156, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6857114946230863, | |
| "train_runtime": 19132.8262, | |
| "train_samples_per_second": 1.045, | |
| "train_steps_per_second": 0.008 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 156, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 200, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |