| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 391, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 422.289956912934, | |
| "learning_rate": 1.25e-09, | |
| "logits/chosen": -4.623842239379883, | |
| "logits/rejected": -4.85917854309082, | |
| "logps/chosen": -239.31422424316406, | |
| "logps/rejected": -207.56365966796875, | |
| "loss": 0.6959, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 397.335507450448, | |
| "learning_rate": 1.25e-08, | |
| "logits/chosen": -4.334493160247803, | |
| "logits/rejected": -4.643917560577393, | |
| "logps/chosen": -265.1265563964844, | |
| "logps/rejected": -215.76223754882812, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.0019649432506412268, | |
| "rewards/margins": 0.0034448718652129173, | |
| "rewards/rejected": -0.005409814417362213, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 411.8624893441942, | |
| "learning_rate": 2.5e-08, | |
| "logits/chosen": -4.506826400756836, | |
| "logits/rejected": -4.740732192993164, | |
| "logps/chosen": -267.86932373046875, | |
| "logps/rejected": -216.64578247070312, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 0.007061410695314407, | |
| "rewards/margins": 0.010701502673327923, | |
| "rewards/rejected": -0.003640091512352228, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 426.4548766919431, | |
| "learning_rate": 3.75e-08, | |
| "logits/chosen": -4.585576057434082, | |
| "logits/rejected": -4.762608528137207, | |
| "logps/chosen": -258.25, | |
| "logps/rejected": -214.71231079101562, | |
| "loss": 0.6756, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.028281379491090775, | |
| "rewards/margins": 0.04623327776789665, | |
| "rewards/rejected": -0.017951902002096176, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 368.66232685986097, | |
| "learning_rate": 5e-08, | |
| "logits/chosen": -4.62213659286499, | |
| "logits/rejected": -4.706842422485352, | |
| "logps/chosen": -252.6122283935547, | |
| "logps/rejected": -220.41427612304688, | |
| "loss": 0.6177, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.0986957773566246, | |
| "rewards/margins": 0.1730131059885025, | |
| "rewards/rejected": -0.07431730628013611, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 294.81806277707193, | |
| "learning_rate": 4.989992961303737e-08, | |
| "logits/chosen": -4.516692161560059, | |
| "logits/rejected": -4.714283466339111, | |
| "logps/chosen": -269.63470458984375, | |
| "logps/rejected": -227.5962371826172, | |
| "loss": 0.5368, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.2334650307893753, | |
| "rewards/margins": 0.4059367775917053, | |
| "rewards/rejected": -0.17247170209884644, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 263.2732793061953, | |
| "learning_rate": 4.960051957873725e-08, | |
| "logits/chosen": -4.610293388366699, | |
| "logits/rejected": -4.734058856964111, | |
| "logps/chosen": -237.3783721923828, | |
| "logps/rejected": -218.6456298828125, | |
| "loss": 0.4381, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.3744103014469147, | |
| "rewards/margins": 0.7368327975273132, | |
| "rewards/rejected": -0.3624224364757538, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 243.25478495437125, | |
| "learning_rate": 4.910416686333906e-08, | |
| "logits/chosen": -4.496267795562744, | |
| "logits/rejected": -4.724743843078613, | |
| "logps/chosen": -248.19540405273438, | |
| "logps/rejected": -222.69140625, | |
| "loss": 0.4014, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 0.49465441703796387, | |
| "rewards/margins": 0.9875491261482239, | |
| "rewards/rejected": -0.49289458990097046, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 203.40516001004536, | |
| "learning_rate": 4.841484508350678e-08, | |
| "logits/chosen": -4.488529682159424, | |
| "logits/rejected": -4.711686611175537, | |
| "logps/chosen": -254.6974639892578, | |
| "logps/rejected": -235.7099151611328, | |
| "loss": 0.3429, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.652252197265625, | |
| "rewards/margins": 1.1913691759109497, | |
| "rewards/rejected": -0.5391170978546143, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 237.79736326938396, | |
| "learning_rate": 4.7538072695020406e-08, | |
| "logits/chosen": -4.58192777633667, | |
| "logits/rejected": -4.797459602355957, | |
| "logps/chosen": -243.06143188476562, | |
| "logps/rejected": -210.63308715820312, | |
| "loss": 0.3085, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 0.8151354789733887, | |
| "rewards/margins": 1.595442533493042, | |
| "rewards/rejected": -0.7803069353103638, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 222.07031689896021, | |
| "learning_rate": 4.6480868814055416e-08, | |
| "logits/chosen": -4.459914207458496, | |
| "logits/rejected": -4.747165679931641, | |
| "logps/chosen": -247.2704620361328, | |
| "logps/rejected": -236.6487274169922, | |
| "loss": 0.2984, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 0.9447771906852722, | |
| "rewards/margins": 1.8987438678741455, | |
| "rewards/rejected": -0.953966498374939, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_logits/chosen": -4.55070686340332, | |
| "eval_logits/rejected": -4.762002944946289, | |
| "eval_logps/chosen": -390.7516174316406, | |
| "eval_logps/rejected": -515.337158203125, | |
| "eval_loss": 0.9402573704719543, | |
| "eval_rewards/accuracies": 0.40625, | |
| "eval_rewards/chosen": -0.027714576572179794, | |
| "eval_rewards/margins": -0.3327641487121582, | |
| "eval_rewards/rejected": 0.3050495684146881, | |
| "eval_runtime": 97.8238, | |
| "eval_samples_per_second": 20.445, | |
| "eval_steps_per_second": 0.327, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 179.38114597248955, | |
| "learning_rate": 4.525169702472916e-08, | |
| "logits/chosen": -4.51773738861084, | |
| "logits/rejected": -4.777291297912598, | |
| "logps/chosen": -238.6410369873047, | |
| "logps/rejected": -227.3874053955078, | |
| "loss": 0.2711, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 1.0544707775115967, | |
| "rewards/margins": 2.1246237754821777, | |
| "rewards/rejected": -1.070152997970581, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 163.17108435846185, | |
| "learning_rate": 4.386039762276975e-08, | |
| "logits/chosen": -4.462746620178223, | |
| "logits/rejected": -4.7056145668029785, | |
| "logps/chosen": -262.86920166015625, | |
| "logps/rejected": -238.99801635742188, | |
| "loss": 0.267, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 1.3120397329330444, | |
| "rewards/margins": 2.3509771823883057, | |
| "rewards/rejected": -1.0389372110366821, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 229.70994586547334, | |
| "learning_rate": 4.231810883773999e-08, | |
| "logits/chosen": -4.532160758972168, | |
| "logits/rejected": -4.803128719329834, | |
| "logps/chosen": -246.3385009765625, | |
| "logps/rejected": -227.83792114257812, | |
| "loss": 0.2577, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 1.2105365991592407, | |
| "rewards/margins": 2.461458921432495, | |
| "rewards/rejected": -1.2509223222732544, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 152.87001146329087, | |
| "learning_rate": 4.063717766448194e-08, | |
| "logits/chosen": -4.559675216674805, | |
| "logits/rejected": -4.842075824737549, | |
| "logps/chosen": -272.3346252441406, | |
| "logps/rejected": -246.54464721679688, | |
| "loss": 0.2453, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.1860424280166626, | |
| "rewards/margins": 2.6231608390808105, | |
| "rewards/rejected": -1.4371181726455688, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 176.7311824941399, | |
| "learning_rate": 3.8831061017632845e-08, | |
| "logits/chosen": -4.557957172393799, | |
| "logits/rejected": -4.812293529510498, | |
| "logps/chosen": -239.32144165039062, | |
| "logps/rejected": -232.82479858398438, | |
| "loss": 0.2425, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 1.3096827268600464, | |
| "rewards/margins": 2.680234432220459, | |
| "rewards/rejected": -1.3705517053604126, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 179.5862213559593, | |
| "learning_rate": 3.691421800053269e-08, | |
| "logits/chosen": -4.614952564239502, | |
| "logits/rejected": -4.799678802490234, | |
| "logps/chosen": -238.4506378173828, | |
| "logps/rejected": -229.4785614013672, | |
| "loss": 0.2216, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 1.300586462020874, | |
| "rewards/margins": 2.687537908554077, | |
| "rewards/rejected": -1.3869514465332031, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 160.70849517962517, | |
| "learning_rate": 3.490199415097892e-08, | |
| "logits/chosen": -4.499081611633301, | |
| "logits/rejected": -4.779529571533203, | |
| "logps/chosen": -244.9915771484375, | |
| "logps/rejected": -235.46743774414062, | |
| "loss": 0.2108, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.6082652807235718, | |
| "rewards/margins": 3.174710988998413, | |
| "rewards/rejected": -1.5664453506469727, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 149.11333079529007, | |
| "learning_rate": 3.2810498590513937e-08, | |
| "logits/chosen": -4.59390926361084, | |
| "logits/rejected": -4.832152366638184, | |
| "logps/chosen": -222.95986938476562, | |
| "logps/rejected": -211.53585815429688, | |
| "loss": 0.2274, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.3903782367706299, | |
| "rewards/margins": 2.9419426918029785, | |
| "rewards/rejected": -1.5515644550323486, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 171.77003361632143, | |
| "learning_rate": 3.065647506074306e-08, | |
| "logits/chosen": -4.56182861328125, | |
| "logits/rejected": -4.7075724601745605, | |
| "logps/chosen": -245.95556640625, | |
| "logps/rejected": -247.3394012451172, | |
| "loss": 0.2299, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.4937294721603394, | |
| "rewards/margins": 2.711061477661133, | |
| "rewards/rejected": -1.2173320055007935, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 152.3542212939215, | |
| "learning_rate": 2.8457167879118325e-08, | |
| "logits/chosen": -4.556639194488525, | |
| "logits/rejected": -4.735670566558838, | |
| "logps/chosen": -237.8849639892578, | |
| "logps/rejected": -229.5240020751953, | |
| "loss": 0.2338, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.4254279136657715, | |
| "rewards/margins": 2.843172073364258, | |
| "rewards/rejected": -1.4177442789077759, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "eval_logits/chosen": -4.538640975952148, | |
| "eval_logits/rejected": -4.758352279663086, | |
| "eval_logps/chosen": -391.07916259765625, | |
| "eval_logps/rejected": -514.2457885742188, | |
| "eval_loss": 1.0996507406234741, | |
| "eval_rewards/accuracies": 0.38671875, | |
| "eval_rewards/chosen": -0.06046929210424423, | |
| "eval_rewards/margins": -0.4746614694595337, | |
| "eval_rewards/rejected": 0.41419219970703125, | |
| "eval_runtime": 98.0841, | |
| "eval_samples_per_second": 20.391, | |
| "eval_steps_per_second": 0.326, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 139.65082338502864, | |
| "learning_rate": 2.6230183887296952e-08, | |
| "logits/chosen": -4.619709014892578, | |
| "logits/rejected": -4.859663963317871, | |
| "logps/chosen": -249.32113647460938, | |
| "logps/rejected": -246.33468627929688, | |
| "loss": 0.1967, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.6563961505889893, | |
| "rewards/margins": 3.425267457962036, | |
| "rewards/rejected": -1.768871545791626, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 147.42119588032455, | |
| "learning_rate": 2.3993351497264626e-08, | |
| "logits/chosen": -4.466108798980713, | |
| "logits/rejected": -4.793113708496094, | |
| "logps/chosen": -247.28756713867188, | |
| "logps/rejected": -244.1797637939453, | |
| "loss": 0.2159, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.5523773431777954, | |
| "rewards/margins": 3.5030410289764404, | |
| "rewards/rejected": -1.9506635665893555, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 158.88183877851495, | |
| "learning_rate": 2.1764577963648613e-08, | |
| "logits/chosen": -4.541924476623535, | |
| "logits/rejected": -4.858447074890137, | |
| "logps/chosen": -245.6726531982422, | |
| "logps/rejected": -243.28677368164062, | |
| "loss": 0.2197, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.5296446084976196, | |
| "rewards/margins": 3.487210750579834, | |
| "rewards/rejected": -1.957566499710083, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 157.203527489415, | |
| "learning_rate": 1.9561706024845818e-08, | |
| "logits/chosen": -4.5143561363220215, | |
| "logits/rejected": -4.772491455078125, | |
| "logps/chosen": -256.7393798828125, | |
| "logps/rejected": -240.91226196289062, | |
| "loss": 0.2123, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.6083800792694092, | |
| "rewards/margins": 3.60624623298645, | |
| "rewards/rejected": -1.9978656768798828, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 159.9556615524972, | |
| "learning_rate": 1.740237106064383e-08, | |
| "logits/chosen": -4.641883850097656, | |
| "logits/rejected": -4.893360614776611, | |
| "logps/chosen": -234.55264282226562, | |
| "logps/rejected": -224.25631713867188, | |
| "loss": 0.2024, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.414884090423584, | |
| "rewards/margins": 3.079598903656006, | |
| "rewards/rejected": -1.6647151708602905, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 198.60636587673295, | |
| "learning_rate": 1.530385990987863e-08, | |
| "logits/chosen": -4.517378330230713, | |
| "logits/rejected": -4.799233913421631, | |
| "logps/chosen": -238.4044189453125, | |
| "logps/rejected": -242.4611358642578, | |
| "loss": 0.2025, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.5644124746322632, | |
| "rewards/margins": 3.8356785774230957, | |
| "rewards/rejected": -2.271266460418701, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 186.10967020286805, | |
| "learning_rate": 1.3282972478382409e-08, | |
| "logits/chosen": -4.603947639465332, | |
| "logits/rejected": -4.826247215270996, | |
| "logps/chosen": -245.811279296875, | |
| "logps/rejected": -236.3955841064453, | |
| "loss": 0.2127, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 1.5221991539001465, | |
| "rewards/margins": 3.6592516899108887, | |
| "rewards/rejected": -2.137052536010742, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 144.59147008918274, | |
| "learning_rate": 1.1355887245137383e-08, | |
| "logits/chosen": -4.557550430297852, | |
| "logits/rejected": -4.780216217041016, | |
| "logps/chosen": -248.1402587890625, | |
| "logps/rejected": -262.3576354980469, | |
| "loss": 0.1884, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 1.636661171913147, | |
| "rewards/margins": 3.578484296798706, | |
| "rewards/rejected": -1.9418232440948486, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 331.8489693457681, | |
| "learning_rate": 9.538031743343628e-09, | |
| "logits/chosen": -4.426544666290283, | |
| "logits/rejected": -4.761611461639404, | |
| "logps/chosen": -255.4153289794922, | |
| "logps/rejected": -234.84487915039062, | |
| "loss": 0.1966, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.7193052768707275, | |
| "rewards/margins": 3.6408512592315674, | |
| "rewards/rejected": -1.9215457439422607, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 177.15049850318, | |
| "learning_rate": 7.843959053281662e-09, | |
| "logits/chosen": -4.541173934936523, | |
| "logits/rejected": -4.741909503936768, | |
| "logps/chosen": -236.41854858398438, | |
| "logps/rejected": -236.483642578125, | |
| "loss": 0.2158, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": 1.6989631652832031, | |
| "rewards/margins": 3.7507786750793457, | |
| "rewards/rejected": -2.0518155097961426, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "eval_logits/chosen": -4.539734840393066, | |
| "eval_logits/rejected": -4.759631633758545, | |
| "eval_logps/chosen": -392.2895812988281, | |
| "eval_logps/rejected": -515.406494140625, | |
| "eval_loss": 1.1359957456588745, | |
| "eval_rewards/accuracies": 0.41015625, | |
| "eval_rewards/chosen": -0.18151262402534485, | |
| "eval_rewards/margins": -0.47963014245033264, | |
| "eval_rewards/rejected": 0.2981175184249878, | |
| "eval_runtime": 97.905, | |
| "eval_samples_per_second": 20.428, | |
| "eval_steps_per_second": 0.327, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 159.8961908197972, | |
| "learning_rate": 6.28723129572247e-09, | |
| "logits/chosen": -4.461672782897949, | |
| "logits/rejected": -4.776505470275879, | |
| "logps/chosen": -244.0063934326172, | |
| "logps/rejected": -244.4510955810547, | |
| "loss": 0.2028, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 1.7133222818374634, | |
| "rewards/margins": 3.9386374950408936, | |
| "rewards/rejected": -2.2253153324127197, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 185.55678167306448, | |
| "learning_rate": 4.880311058593617e-09, | |
| "logits/chosen": -4.521292209625244, | |
| "logits/rejected": -4.848372936248779, | |
| "logps/chosen": -230.04397583007812, | |
| "logps/rejected": -226.9331817626953, | |
| "loss": 0.2095, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.5120347738265991, | |
| "rewards/margins": 3.560204267501831, | |
| "rewards/rejected": -2.0481698513031006, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 155.26912676521727, | |
| "learning_rate": 3.6344616260994942e-09, | |
| "logits/chosen": -4.473151683807373, | |
| "logits/rejected": -4.766911029815674, | |
| "logps/chosen": -251.7646484375, | |
| "logps/rejected": -254.7379913330078, | |
| "loss": 0.1928, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.638421654701233, | |
| "rewards/margins": 3.8422539234161377, | |
| "rewards/rejected": -2.2038321495056152, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 151.70242269299357, | |
| "learning_rate": 2.5596568090246547e-09, | |
| "logits/chosen": -4.4894232749938965, | |
| "logits/rejected": -4.816695213317871, | |
| "logps/chosen": -255.334716796875, | |
| "logps/rejected": -228.4678192138672, | |
| "loss": 0.1963, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.8811309337615967, | |
| "rewards/margins": 4.049709320068359, | |
| "rewards/rejected": -2.1685783863067627, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 150.47860807724385, | |
| "learning_rate": 1.6645010980854079e-09, | |
| "logits/chosen": -4.505433559417725, | |
| "logits/rejected": -4.675290584564209, | |
| "logps/chosen": -240.54714965820312, | |
| "logps/rejected": -246.03665161132812, | |
| "loss": 0.2168, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.4924026727676392, | |
| "rewards/margins": 3.3190674781799316, | |
| "rewards/rejected": -1.8266645669937134, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 165.46679700251414, | |
| "learning_rate": 9.561607795526006e-10, | |
| "logits/chosen": -4.49678373336792, | |
| "logits/rejected": -4.711674690246582, | |
| "logps/chosen": -246.95388793945312, | |
| "logps/rejected": -247.2928009033203, | |
| "loss": 0.2121, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 1.468100905418396, | |
| "rewards/margins": 3.451129913330078, | |
| "rewards/rejected": -1.9830286502838135, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 149.25359071163066, | |
| "learning_rate": 4.403065646083809e-10, | |
| "logits/chosen": -4.518364906311035, | |
| "logits/rejected": -4.680220603942871, | |
| "logps/chosen": -241.2300262451172, | |
| "logps/rejected": -255.6038818359375, | |
| "loss": 0.2103, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.5591168403625488, | |
| "rewards/margins": 3.434800386428833, | |
| "rewards/rejected": -1.8756835460662842, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 160.57572024314433, | |
| "learning_rate": 1.2106819172520434e-10, | |
| "logits/chosen": -4.618407726287842, | |
| "logits/rejected": -4.8883843421936035, | |
| "logps/chosen": -246.15853881835938, | |
| "logps/rejected": -243.1090087890625, | |
| "loss": 0.2034, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 1.5924150943756104, | |
| "rewards/margins": 3.8508572578430176, | |
| "rewards/rejected": -2.2584421634674072, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 235.40189038757265, | |
| "learning_rate": 1.0013655036916758e-12, | |
| "logits/chosen": -4.611303329467773, | |
| "logits/rejected": -4.857443809509277, | |
| "logps/chosen": -242.1800994873047, | |
| "logps/rejected": -241.1270294189453, | |
| "loss": 0.199, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.7492481470108032, | |
| "rewards/margins": 3.8104407787323, | |
| "rewards/rejected": -2.0611929893493652, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 391, | |
| "total_flos": 0.0, | |
| "train_loss": 0.29024992444935965, | |
| "train_runtime": 6148.7126, | |
| "train_samples_per_second": 8.132, | |
| "train_steps_per_second": 0.064 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 391, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |