| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9965977492802931, | |
| "eval_steps": 100, | |
| "global_step": 238, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -2.852463483810425, | |
| "logits/rejected": -2.8067848682403564, | |
| "logps/chosen": -307.2070617675781, | |
| "logps/rejected": -292.9268493652344, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0833333333333334e-06, | |
| "logits/chosen": -2.7673232555389404, | |
| "logits/rejected": -2.714327096939087, | |
| "logps/chosen": -278.9463806152344, | |
| "logps/rejected": -269.96435546875, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.5086805820465088, | |
| "rewards/chosen": 0.0003373799263499677, | |
| "rewards/margins": 0.0005398019566200674, | |
| "rewards/rejected": -0.0002024220593739301, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.166666666666667e-06, | |
| "logits/chosen": -2.789074659347534, | |
| "logits/rejected": -2.739534378051758, | |
| "logps/chosen": -288.80340576171875, | |
| "logps/rejected": -252.3999481201172, | |
| "loss": 0.6892, | |
| "rewards/accuracies": 0.6546875238418579, | |
| "rewards/chosen": 0.003567395731806755, | |
| "rewards/margins": 0.00781105924397707, | |
| "rewards/rejected": -0.004243663977831602, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.99030821197584e-06, | |
| "logits/chosen": -2.7394039630889893, | |
| "logits/rejected": -2.6750001907348633, | |
| "logps/chosen": -277.71734619140625, | |
| "logps/rejected": -260.70751953125, | |
| "loss": 0.6798, | |
| "rewards/accuracies": 0.6953125, | |
| "rewards/chosen": 0.015827985480427742, | |
| "rewards/margins": 0.02965703047811985, | |
| "rewards/rejected": -0.013829047791659832, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.931352528237398e-06, | |
| "logits/chosen": -2.73751163482666, | |
| "logits/rejected": -2.684286594390869, | |
| "logps/chosen": -285.98248291015625, | |
| "logps/rejected": -277.09478759765625, | |
| "loss": 0.6678, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.02657519280910492, | |
| "rewards/margins": 0.0554216094315052, | |
| "rewards/rejected": -0.028846416622400284, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.820092227512736e-06, | |
| "logits/chosen": -2.729034900665283, | |
| "logits/rejected": -2.6729769706726074, | |
| "logps/chosen": -275.21514892578125, | |
| "logps/rejected": -280.52362060546875, | |
| "loss": 0.6521, | |
| "rewards/accuracies": 0.714062511920929, | |
| "rewards/chosen": -0.01229151152074337, | |
| "rewards/margins": 0.09265317022800446, | |
| "rewards/rejected": -0.10494468361139297, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.658920803689553e-06, | |
| "logits/chosen": -2.6756510734558105, | |
| "logits/rejected": -2.621499538421631, | |
| "logps/chosen": -280.72991943359375, | |
| "logps/rejected": -274.62103271484375, | |
| "loss": 0.6387, | |
| "rewards/accuracies": 0.692187488079071, | |
| "rewards/chosen": -0.04894893616437912, | |
| "rewards/margins": 0.1295410692691803, | |
| "rewards/rejected": -0.17849001288414001, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.451305466682615e-06, | |
| "logits/chosen": -2.6422786712646484, | |
| "logits/rejected": -2.585517644882202, | |
| "logps/chosen": -285.8612365722656, | |
| "logps/rejected": -291.7132873535156, | |
| "loss": 0.6281, | |
| "rewards/accuracies": 0.7015625238418579, | |
| "rewards/chosen": -0.10629353672266006, | |
| "rewards/margins": 0.18995100259780884, | |
| "rewards/rejected": -0.2962445318698883, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.2017125538726574e-06, | |
| "logits/chosen": -2.660236120223999, | |
| "logits/rejected": -2.5804734230041504, | |
| "logps/chosen": -292.82135009765625, | |
| "logps/rejected": -288.45306396484375, | |
| "loss": 0.605, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.13801410794258118, | |
| "rewards/margins": 0.2476145476102829, | |
| "rewards/rejected": -0.3856286406517029, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.915511447755793e-06, | |
| "logits/chosen": -2.6195216178894043, | |
| "logits/rejected": -2.5578157901763916, | |
| "logps/chosen": -293.1012878417969, | |
| "logps/rejected": -303.3466796875, | |
| "loss": 0.6143, | |
| "rewards/accuracies": 0.6734374761581421, | |
| "rewards/chosen": -0.19188150763511658, | |
| "rewards/margins": 0.25040262937545776, | |
| "rewards/rejected": -0.44228416681289673, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.5988590667807542e-06, | |
| "logits/chosen": -2.614781618118286, | |
| "logits/rejected": -2.5338780879974365, | |
| "logps/chosen": -307.21429443359375, | |
| "logps/rejected": -317.61480712890625, | |
| "loss": 0.5916, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": -0.21577294170856476, | |
| "rewards/margins": 0.3157234787940979, | |
| "rewards/rejected": -0.5314964056015015, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": -2.6026527881622314, | |
| "eval_logits/rejected": -2.5165884494781494, | |
| "eval_logps/chosen": -309.4196472167969, | |
| "eval_logps/rejected": -308.91461181640625, | |
| "eval_loss": 0.6025363802909851, | |
| "eval_rewards/accuracies": 0.6940000057220459, | |
| "eval_rewards/chosen": -0.25379911065101624, | |
| "eval_rewards/margins": 0.26018837094306946, | |
| "eval_rewards/rejected": -0.5139874815940857, | |
| "eval_runtime": 384.168, | |
| "eval_samples_per_second": 5.206, | |
| "eval_steps_per_second": 0.651, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.2585674142717483e-06, | |
| "logits/chosen": -2.6301052570343018, | |
| "logits/rejected": -2.5181527137756348, | |
| "logps/chosen": -304.8443603515625, | |
| "logps/rejected": -305.76165771484375, | |
| "loss": 0.6052, | |
| "rewards/accuracies": 0.6953125, | |
| "rewards/chosen": -0.2113591432571411, | |
| "rewards/margins": 0.2960760295391083, | |
| "rewards/rejected": -0.507435142993927, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.901957034798671e-06, | |
| "logits/chosen": -2.571101665496826, | |
| "logits/rejected": -2.4931883811950684, | |
| "logps/chosen": -299.7984313964844, | |
| "logps/rejected": -306.89056396484375, | |
| "loss": 0.6, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -0.281256765127182, | |
| "rewards/margins": 0.25873565673828125, | |
| "rewards/rejected": -0.5399924516677856, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.536699530523292e-06, | |
| "logits/chosen": -2.564572811126709, | |
| "logits/rejected": -2.4486539363861084, | |
| "logps/chosen": -308.9500732421875, | |
| "logps/rejected": -306.9311828613281, | |
| "loss": 0.582, | |
| "rewards/accuracies": 0.7421875, | |
| "rewards/chosen": -0.2418045997619629, | |
| "rewards/margins": 0.36538395285606384, | |
| "rewards/rejected": -0.6071885228157043, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.1706525253979533e-06, | |
| "logits/chosen": -2.544403553009033, | |
| "logits/rejected": -2.417354106903076, | |
| "logps/chosen": -327.3999328613281, | |
| "logps/rejected": -330.9579162597656, | |
| "loss": 0.5807, | |
| "rewards/accuracies": 0.7109375, | |
| "rewards/chosen": -0.29655271768569946, | |
| "rewards/margins": 0.36625009775161743, | |
| "rewards/rejected": -0.6628028750419617, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.811690627559351e-06, | |
| "logits/chosen": -2.4691693782806396, | |
| "logits/rejected": -2.3947396278381348, | |
| "logps/chosen": -303.43988037109375, | |
| "logps/rejected": -326.5160217285156, | |
| "loss": 0.5886, | |
| "rewards/accuracies": 0.7046874761581421, | |
| "rewards/chosen": -0.25864046812057495, | |
| "rewards/margins": 0.34002798795700073, | |
| "rewards/rejected": -0.5986684560775757, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.4675360263490296e-06, | |
| "logits/chosen": -2.4801511764526367, | |
| "logits/rejected": -2.3981406688690186, | |
| "logps/chosen": -303.1156005859375, | |
| "logps/rejected": -314.0051574707031, | |
| "loss": 0.577, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.25856736302375793, | |
| "rewards/margins": 0.3619126081466675, | |
| "rewards/rejected": -0.620479941368103, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.1455923682523476e-06, | |
| "logits/chosen": -2.462503671646118, | |
| "logits/rejected": -2.3446600437164307, | |
| "logps/chosen": -325.72021484375, | |
| "logps/rejected": -326.17205810546875, | |
| "loss": 0.5832, | |
| "rewards/accuracies": 0.7015625238418579, | |
| "rewards/chosen": -0.3383588194847107, | |
| "rewards/margins": 0.36354926228523254, | |
| "rewards/rejected": -0.7019080519676208, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.527854855097226e-07, | |
| "logits/chosen": -2.4440627098083496, | |
| "logits/rejected": -2.290944814682007, | |
| "logps/chosen": -310.25933837890625, | |
| "logps/rejected": -306.5115966796875, | |
| "loss": 0.5763, | |
| "rewards/accuracies": 0.7203124761581421, | |
| "rewards/chosen": -0.32932430505752563, | |
| "rewards/margins": 0.3963031768798828, | |
| "rewards/rejected": -0.7256274223327637, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 5.954144037354645e-07, | |
| "logits/chosen": -2.4063987731933594, | |
| "logits/rejected": -2.3209547996520996, | |
| "logps/chosen": -316.28802490234375, | |
| "logps/rejected": -324.861083984375, | |
| "loss": 0.5673, | |
| "rewards/accuracies": 0.7421875, | |
| "rewards/chosen": -0.3298575282096863, | |
| "rewards/margins": 0.40440693497657776, | |
| "rewards/rejected": -0.7342644929885864, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.7901583375171277e-07, | |
| "logits/chosen": -2.37199068069458, | |
| "logits/rejected": -2.2828264236450195, | |
| "logps/chosen": -322.8636169433594, | |
| "logps/rejected": -347.17742919921875, | |
| "loss": 0.5667, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.3475632071495056, | |
| "rewards/margins": 0.3942633271217346, | |
| "rewards/rejected": -0.7418265342712402, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": -2.372931718826294, | |
| "eval_logits/rejected": -2.2597038745880127, | |
| "eval_logps/chosen": -323.44244384765625, | |
| "eval_logps/rejected": -332.209228515625, | |
| "eval_loss": 0.5787754058837891, | |
| "eval_rewards/accuracies": 0.7200000286102295, | |
| "eval_rewards/chosen": -0.39402660727500916, | |
| "eval_rewards/margins": 0.35290706157684326, | |
| "eval_rewards/rejected": -0.74693363904953, | |
| "eval_runtime": 384.1102, | |
| "eval_samples_per_second": 5.207, | |
| "eval_steps_per_second": 0.651, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.0824506276503898e-07, | |
| "logits/chosen": -2.346809148788452, | |
| "logits/rejected": -2.259718656539917, | |
| "logps/chosen": -330.906005859375, | |
| "logps/rejected": -344.9519348144531, | |
| "loss": 0.5802, | |
| "rewards/accuracies": 0.667187511920929, | |
| "rewards/chosen": -0.37789514660835266, | |
| "rewards/margins": 0.33273065090179443, | |
| "rewards/rejected": -0.7106258273124695, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 8.677580722139673e-08, | |
| "logits/chosen": -2.3068695068359375, | |
| "logits/rejected": -2.2727885246276855, | |
| "logps/chosen": -312.369140625, | |
| "logps/rejected": -334.8897399902344, | |
| "loss": 0.5761, | |
| "rewards/accuracies": 0.7046874761581421, | |
| "rewards/chosen": -0.40535077452659607, | |
| "rewards/margins": 0.3713846206665039, | |
| "rewards/rejected": -0.7767353653907776, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 1.7221181760899153e-08, | |
| "logits/chosen": -2.347468852996826, | |
| "logits/rejected": -2.2807364463806152, | |
| "logps/chosen": -321.6934814453125, | |
| "logps/rejected": -344.81976318359375, | |
| "loss": 0.5749, | |
| "rewards/accuracies": 0.698437511920929, | |
| "rewards/chosen": -0.3814999759197235, | |
| "rewards/margins": 0.3564358651638031, | |
| "rewards/rejected": -0.7379359006881714, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 238, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6080600586758942, | |
| "train_runtime": 20828.337, | |
| "train_samples_per_second": 2.935, | |
| "train_steps_per_second": 0.011 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 238, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |