| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.3161794977356937, | |
| "eval_steps": 500, | |
| "global_step": 24, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.013174145738987238, | |
| "grad_norm": 0.5370480418205261, | |
| "learning_rate": 1.25e-07, | |
| "logits/chosen": 10.071717262268066, | |
| "logits/rejected": 10.610974311828613, | |
| "logps/chosen": -0.5208129286766052, | |
| "logps/rejected": -0.5466158390045166, | |
| "loss": 1.3666, | |
| "rewards/accuracies": 0.5703125, | |
| "rewards/chosen": -1.0416258573532104, | |
| "rewards/margins": 0.05160584673285484, | |
| "rewards/rejected": -1.0932316780090332, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.026348291477974475, | |
| "grad_norm": 0.6556285619735718, | |
| "learning_rate": 2.5e-07, | |
| "logits/chosen": 10.222262382507324, | |
| "logits/rejected": 10.754176139831543, | |
| "logps/chosen": -0.5501401424407959, | |
| "logps/rejected": -0.580172598361969, | |
| "loss": 1.363, | |
| "rewards/accuracies": 0.546875, | |
| "rewards/chosen": -1.1002802848815918, | |
| "rewards/margins": 0.06006482243537903, | |
| "rewards/rejected": -1.160345196723938, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.03952243721696171, | |
| "grad_norm": 0.5852232575416565, | |
| "learning_rate": 3.75e-07, | |
| "logits/chosen": 10.029905319213867, | |
| "logits/rejected": 10.635485649108887, | |
| "logps/chosen": -0.6488087773323059, | |
| "logps/rejected": -0.6685145497322083, | |
| "loss": 1.3891, | |
| "rewards/accuracies": 0.484375, | |
| "rewards/chosen": -1.2976175546646118, | |
| "rewards/margins": 0.03941144049167633, | |
| "rewards/rejected": -1.3370290994644165, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.05269658295594895, | |
| "grad_norm": 0.880297839641571, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": 9.950821876525879, | |
| "logits/rejected": 10.651039123535156, | |
| "logps/chosen": -0.5544578433036804, | |
| "logps/rejected": -0.5768871307373047, | |
| "loss": 1.3895, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -1.1089156866073608, | |
| "rewards/margins": 0.04485849663615227, | |
| "rewards/rejected": -1.1537742614746094, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.06587072869493618, | |
| "grad_norm": 1.0345523357391357, | |
| "learning_rate": 6.249999999999999e-07, | |
| "logits/chosen": 10.145177841186523, | |
| "logits/rejected": 10.952865600585938, | |
| "logps/chosen": -0.5600388646125793, | |
| "logps/rejected": -0.5631997585296631, | |
| "loss": 1.3982, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -1.1200777292251587, | |
| "rewards/margins": 0.006321802735328674, | |
| "rewards/rejected": -1.1263995170593262, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.07904487443392343, | |
| "grad_norm": 0.9534112811088562, | |
| "learning_rate": 7.5e-07, | |
| "logits/chosen": 10.677449226379395, | |
| "logits/rejected": 11.309142112731934, | |
| "logps/chosen": -0.5716575384140015, | |
| "logps/rejected": -0.6011527180671692, | |
| "loss": 1.3753, | |
| "rewards/accuracies": 0.5859375, | |
| "rewards/chosen": -1.143315076828003, | |
| "rewards/margins": 0.05899051949381828, | |
| "rewards/rejected": -1.2023054361343384, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.09221902017291066, | |
| "grad_norm": 0.7576708793640137, | |
| "learning_rate": 8.75e-07, | |
| "logits/chosen": 9.934803009033203, | |
| "logits/rejected": 10.656447410583496, | |
| "logps/chosen": -0.5172385573387146, | |
| "logps/rejected": -0.5407333374023438, | |
| "loss": 1.3735, | |
| "rewards/accuracies": 0.5234375, | |
| "rewards/chosen": -1.0344771146774292, | |
| "rewards/margins": 0.04698962718248367, | |
| "rewards/rejected": -1.0814666748046875, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.1053931659118979, | |
| "grad_norm": 0.6589964628219604, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": 10.398344039916992, | |
| "logits/rejected": 11.005778312683105, | |
| "logps/chosen": -0.5838003754615784, | |
| "logps/rejected": -0.5947949290275574, | |
| "loss": 1.396, | |
| "rewards/accuracies": 0.5546875, | |
| "rewards/chosen": -1.1676007509231567, | |
| "rewards/margins": 0.02198910340666771, | |
| "rewards/rejected": -1.1895898580551147, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.11856731165088513, | |
| "grad_norm": 0.7832041382789612, | |
| "learning_rate": 9.994504457428556e-07, | |
| "logits/chosen": 10.207743644714355, | |
| "logits/rejected": 11.032607078552246, | |
| "logps/chosen": -0.6102317571640015, | |
| "logps/rejected": -0.5968730449676514, | |
| "loss": 1.4486, | |
| "rewards/accuracies": 0.5234375, | |
| "rewards/chosen": -1.220463514328003, | |
| "rewards/margins": -0.02671748585999012, | |
| "rewards/rejected": -1.1937460899353027, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.13174145738987236, | |
| "grad_norm": 1.3487299680709839, | |
| "learning_rate": 9.97802991010949e-07, | |
| "logits/chosen": 9.977300643920898, | |
| "logits/rejected": 10.761427879333496, | |
| "logps/chosen": -0.49488401412963867, | |
| "logps/rejected": -0.47264525294303894, | |
| "loss": 1.4363, | |
| "rewards/accuracies": 0.5078125, | |
| "rewards/chosen": -0.9897680282592773, | |
| "rewards/margins": -0.04447745531797409, | |
| "rewards/rejected": -0.9452905058860779, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.14491560312885962, | |
| "grad_norm": 0.9786595702171326, | |
| "learning_rate": 9.950612572673255e-07, | |
| "logits/chosen": 10.096546173095703, | |
| "logits/rejected": 10.999103546142578, | |
| "logps/chosen": -0.6210685968399048, | |
| "logps/rejected": -0.6260945796966553, | |
| "loss": 1.4145, | |
| "rewards/accuracies": 0.4453125, | |
| "rewards/chosen": -1.2421371936798096, | |
| "rewards/margins": 0.010051850229501724, | |
| "rewards/rejected": -1.2521891593933105, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.15808974886784685, | |
| "grad_norm": 0.9959055781364441, | |
| "learning_rate": 9.912312714377879e-07, | |
| "logits/chosen": 10.130002975463867, | |
| "logits/rejected": 10.666738510131836, | |
| "logps/chosen": -0.5434242486953735, | |
| "logps/rejected": -0.5788334608078003, | |
| "loss": 1.3496, | |
| "rewards/accuracies": 0.5546875, | |
| "rewards/chosen": -1.086848497390747, | |
| "rewards/margins": 0.07081828266382217, | |
| "rewards/rejected": -1.1576669216156006, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.17126389460683408, | |
| "grad_norm": 0.7984468340873718, | |
| "learning_rate": 9.863214526624063e-07, | |
| "logits/chosen": 9.885704040527344, | |
| "logits/rejected": 10.887924194335938, | |
| "logps/chosen": -0.5868184566497803, | |
| "logps/rejected": -0.5810582041740417, | |
| "loss": 1.4268, | |
| "rewards/accuracies": 0.4453125, | |
| "rewards/chosen": -1.1736369132995605, | |
| "rewards/margins": -0.011520491912961006, | |
| "rewards/rejected": -1.1621164083480835, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.1844380403458213, | |
| "grad_norm": 1.042403221130371, | |
| "learning_rate": 9.8034259378842e-07, | |
| "logits/chosen": 10.426769256591797, | |
| "logits/rejected": 11.37084674835205, | |
| "logps/chosen": -0.5749006271362305, | |
| "logps/rejected": -0.6211482882499695, | |
| "loss": 1.3485, | |
| "rewards/accuracies": 0.5703125, | |
| "rewards/chosen": -1.149801254272461, | |
| "rewards/margins": 0.09249527752399445, | |
| "rewards/rejected": -1.242296576499939, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.19761218608480857, | |
| "grad_norm": 1.288346290588379, | |
| "learning_rate": 9.73307837645217e-07, | |
| "logits/chosen": 9.93845272064209, | |
| "logits/rejected": 10.669981002807617, | |
| "logps/chosen": -0.5796323418617249, | |
| "logps/rejected": -0.5988457202911377, | |
| "loss": 1.3761, | |
| "rewards/accuracies": 0.4765625, | |
| "rewards/chosen": -1.1592646837234497, | |
| "rewards/margins": 0.03842674568295479, | |
| "rewards/rejected": -1.1976914405822754, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.2107863318237958, | |
| "grad_norm": 1.540350079536438, | |
| "learning_rate": 9.652326481535433e-07, | |
| "logits/chosen": 10.660343170166016, | |
| "logits/rejected": 11.23757553100586, | |
| "logps/chosen": -0.5483916997909546, | |
| "logps/rejected": -0.5503613948822021, | |
| "loss": 1.3984, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": -1.0967833995819092, | |
| "rewards/margins": 0.003939321264624596, | |
| "rewards/rejected": -1.1007227897644043, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.22396047756278303, | |
| "grad_norm": 1.0571225881576538, | |
| "learning_rate": 9.561347763324483e-07, | |
| "logits/chosen": 10.136300086975098, | |
| "logits/rejected": 10.82028579711914, | |
| "logps/chosen": -0.5688516497612, | |
| "logps/rejected": -0.553267240524292, | |
| "loss": 1.4383, | |
| "rewards/accuracies": 0.546875, | |
| "rewards/chosen": -1.1377032995224, | |
| "rewards/margins": -0.031168799847364426, | |
| "rewards/rejected": -1.106534481048584, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.23713462330177026, | |
| "grad_norm": 1.095320463180542, | |
| "learning_rate": 9.460342212786932e-07, | |
| "logits/chosen": 10.693653106689453, | |
| "logits/rejected": 11.115797996520996, | |
| "logps/chosen": -0.6066192388534546, | |
| "logps/rejected": -0.5900355577468872, | |
| "loss": 1.4382, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -1.2132384777069092, | |
| "rewards/margins": -0.03316720202565193, | |
| "rewards/rejected": -1.1800711154937744, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.2503087690407575, | |
| "grad_norm": 0.6557356119155884, | |
| "learning_rate": 9.349531862043951e-07, | |
| "logits/chosen": 10.337396621704102, | |
| "logits/rejected": 10.86961841583252, | |
| "logps/chosen": -0.525456964969635, | |
| "logps/rejected": -0.5541884899139404, | |
| "loss": 1.3703, | |
| "rewards/accuracies": 0.609375, | |
| "rewards/chosen": -1.05091392993927, | |
| "rewards/margins": 0.057463016360998154, | |
| "rewards/rejected": -1.1083769798278809, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.2634829147797447, | |
| "grad_norm": 0.7737817168235779, | |
| "learning_rate": 9.229160296295487e-07, | |
| "logits/chosen": 10.50332260131836, | |
| "logits/rejected": 11.313106536865234, | |
| "logps/chosen": -0.6586668491363525, | |
| "logps/rejected": -0.6300007104873657, | |
| "loss": 1.4657, | |
| "rewards/accuracies": 0.5390625, | |
| "rewards/chosen": -1.317333698272705, | |
| "rewards/margins": -0.05733209103345871, | |
| "rewards/rejected": -1.2600014209747314, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.276657060518732, | |
| "grad_norm": 0.5888201594352722, | |
| "learning_rate": 9.099492118367122e-07, | |
| "logits/chosen": 9.65850830078125, | |
| "logits/rejected": 10.547868728637695, | |
| "logps/chosen": -0.6071429252624512, | |
| "logps/rejected": -0.6548156142234802, | |
| "loss": 1.3526, | |
| "rewards/accuracies": 0.5390625, | |
| "rewards/chosen": -1.2142858505249023, | |
| "rewards/margins": 0.09534531086683273, | |
| "rewards/rejected": -1.3096312284469604, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.28983120625771924, | |
| "grad_norm": 1.3495151996612549, | |
| "learning_rate": 8.960812367055646e-07, | |
| "logits/chosen": 9.991419792175293, | |
| "logits/rejected": 10.768990516662598, | |
| "logps/chosen": -0.6086853742599487, | |
| "logps/rejected": -0.6002129912376404, | |
| "loss": 1.4365, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -1.2173707485198975, | |
| "rewards/margins": -0.01694471389055252, | |
| "rewards/rejected": -1.2004259824752808, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.3030053519967065, | |
| "grad_norm": 0.40277352929115295, | |
| "learning_rate": 8.813425890551909e-07, | |
| "logits/chosen": 9.981124877929688, | |
| "logits/rejected": 10.845569610595703, | |
| "logps/chosen": -0.6795557141304016, | |
| "logps/rejected": -0.630460262298584, | |
| "loss": 1.5098, | |
| "rewards/accuracies": 0.5078125, | |
| "rewards/chosen": -1.3591114282608032, | |
| "rewards/margins": -0.0981910228729248, | |
| "rewards/rejected": -1.260920524597168, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.3161794977356937, | |
| "grad_norm": 0.738304853439331, | |
| "learning_rate": 8.657656676318345e-07, | |
| "logits/chosen": 10.05403995513916, | |
| "logits/rejected": 10.649165153503418, | |
| "logps/chosen": -0.5632920265197754, | |
| "logps/rejected": -0.5841237902641296, | |
| "loss": 1.3813, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": -1.1265840530395508, | |
| "rewards/margins": 0.0416635125875473, | |
| "rewards/rejected": -1.1682475805282593, | |
| "step": 24 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 75, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 12, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |