| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.0, | |
| "eval_steps": 6, | |
| "global_step": 78, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 0.08957596868276596, | |
| "learning_rate": 2e-05, | |
| "loss": 1.0134, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "eval_loss": 1.0981205701828003, | |
| "eval_runtime": 2.5128, | |
| "eval_samples_per_second": 1.194, | |
| "eval_steps_per_second": 1.194, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 0.07771213352680206, | |
| "learning_rate": 4e-05, | |
| "loss": 0.9545, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 0.1224137470126152, | |
| "learning_rate": 6e-05, | |
| "loss": 1.1733, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 0.09190034121274948, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9954, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 0.08263542503118515, | |
| "learning_rate": 0.0001, | |
| "loss": 0.9486, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 0.09250061959028244, | |
| "learning_rate": 0.00012, | |
| "loss": 0.972, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "eval_loss": 1.0735869407653809, | |
| "eval_runtime": 2.5357, | |
| "eval_samples_per_second": 1.183, | |
| "eval_steps_per_second": 1.183, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 0.1398034691810608, | |
| "learning_rate": 0.00014, | |
| "loss": 1.0445, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 0.0993918851017952, | |
| "learning_rate": 0.00016, | |
| "loss": 0.9169, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 0.07937725633382797, | |
| "learning_rate": 0.00018, | |
| "loss": 0.8462, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 0.10001373291015625, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8708, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 0.1337287873029709, | |
| "learning_rate": 0.00019995690062269984, | |
| "loss": 0.86, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 0.11684636771678925, | |
| "learning_rate": 0.00019982763964192585, | |
| "loss": 0.7982, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "eval_loss": 0.8548387885093689, | |
| "eval_runtime": 2.5536, | |
| "eval_samples_per_second": 1.175, | |
| "eval_steps_per_second": 1.175, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 0.12103456258773804, | |
| "learning_rate": 0.0001996123284790336, | |
| "loss": 0.7906, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 0.1426106095314026, | |
| "learning_rate": 0.00019931115272956405, | |
| "loss": 0.7825, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 0.12367941439151764, | |
| "learning_rate": 0.0001989243720032624, | |
| "loss": 0.7341, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 0.10154826194047928, | |
| "learning_rate": 0.00019845231970029773, | |
| "loss": 0.7064, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 0.13628405332565308, | |
| "learning_rate": 0.0001978954027238763, | |
| "loss": 0.6988, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 0.11276472359895706, | |
| "learning_rate": 0.0001972541011294959, | |
| "loss": 0.6944, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_loss": 0.7151015400886536, | |
| "eval_runtime": 2.5734, | |
| "eval_samples_per_second": 1.166, | |
| "eval_steps_per_second": 1.166, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 0.13381372392177582, | |
| "learning_rate": 0.00019652896771114414, | |
| "loss": 0.6956, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 0.11248588562011719, | |
| "learning_rate": 0.00019572062752479683, | |
| "loss": 0.7155, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 0.17762312293052673, | |
| "learning_rate": 0.00019482977734962753, | |
| "loss": 0.7357, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 0.10546916723251343, | |
| "learning_rate": 0.00019385718508739262, | |
| "loss": 0.6691, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 0.3150898516178131, | |
| "learning_rate": 0.00019280368910050942, | |
| "loss": 0.7167, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 0.13151158392429352, | |
| "learning_rate": 0.00019167019748939846, | |
| "loss": 0.6808, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "eval_loss": 0.6942548751831055, | |
| "eval_runtime": 2.5831, | |
| "eval_samples_per_second": 1.161, | |
| "eval_steps_per_second": 1.161, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 0.14906296133995056, | |
| "learning_rate": 0.00019045768730971196, | |
| "loss": 0.6762, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 0.19484123587608337, | |
| "learning_rate": 0.00018916720373012426, | |
| "loss": 0.6854, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 0.12819896638393402, | |
| "learning_rate": 0.00018779985913140924, | |
| "loss": 0.6873, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 0.21385614573955536, | |
| "learning_rate": 0.00018635683214758214, | |
| "loss": 0.6874, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 0.12286895513534546, | |
| "learning_rate": 0.0001848393666499315, | |
| "loss": 0.6843, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 0.08534862101078033, | |
| "learning_rate": 0.00018324877067481783, | |
| "loss": 0.6763, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "eval_loss": 0.6821426749229431, | |
| "eval_runtime": 2.5911, | |
| "eval_samples_per_second": 1.158, | |
| "eval_steps_per_second": 1.158, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 0.17990928888320923, | |
| "learning_rate": 0.0001815864152961624, | |
| "loss": 0.6789, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 0.12137839943170547, | |
| "learning_rate": 0.0001798537334435986, | |
| "loss": 0.6877, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 0.10240964591503143, | |
| "learning_rate": 0.00017805221866730458, | |
| "loss": 0.6725, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 0.14333295822143555, | |
| "learning_rate": 0.00017618342385058145, | |
| "loss": 0.6745, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 0.0904482752084732, | |
| "learning_rate": 0.00017424895987128722, | |
| "loss": 0.6894, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 0.11753042787313461, | |
| "learning_rate": 0.00017225049421328023, | |
| "loss": 0.67, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "eval_loss": 0.6763580441474915, | |
| "eval_runtime": 2.5955, | |
| "eval_samples_per_second": 1.156, | |
| "eval_steps_per_second": 1.156, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 0.13719823956489563, | |
| "learning_rate": 0.00017018974952906884, | |
| "loss": 0.6589, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 0.1040361225605011, | |
| "learning_rate": 0.0001680685021549063, | |
| "loss": 0.666, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.07594098895788193, | |
| "learning_rate": 0.00016588858057961113, | |
| "loss": 0.645, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "grad_norm": 0.08139798045158386, | |
| "learning_rate": 0.0001636518638684325, | |
| "loss": 0.6542, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 0.07313457876443863, | |
| "learning_rate": 0.0001613602800433194, | |
| "loss": 0.6458, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 0.07903215289115906, | |
| "learning_rate": 0.00015901580442098968, | |
| "loss": 0.6424, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "eval_loss": 0.6730008125305176, | |
| "eval_runtime": 2.5989, | |
| "eval_samples_per_second": 1.154, | |
| "eval_steps_per_second": 1.154, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 0.09322352707386017, | |
| "learning_rate": 0.00015662045791023173, | |
| "loss": 0.6567, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 0.07249985635280609, | |
| "learning_rate": 0.00015417630526990615, | |
| "loss": 0.6384, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 0.07686451077461243, | |
| "learning_rate": 0.0001516854533291494, | |
| "loss": 0.665, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 0.07324113696813583, | |
| "learning_rate": 0.00014915004917131344, | |
| "loss": 0.6297, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 0.09203895926475525, | |
| "learning_rate": 0.00014657227828320635, | |
| "loss": 0.6539, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 0.09338624030351639, | |
| "learning_rate": 0.00014395436267123016, | |
| "loss": 0.6552, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "eval_loss": 0.6780009269714355, | |
| "eval_runtime": 2.6045, | |
| "eval_samples_per_second": 1.152, | |
| "eval_steps_per_second": 1.152, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "grad_norm": 0.0812142863869667, | |
| "learning_rate": 0.00014129855894603886, | |
| "loss": 0.6319, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "grad_norm": 0.19316132366657257, | |
| "learning_rate": 0.00013860715637736818, | |
| "loss": 0.7, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 0.10698059946298599, | |
| "learning_rate": 0.0001358824749207136, | |
| "loss": 0.6725, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "grad_norm": 0.14100198447704315, | |
| "learning_rate": 0.00013312686321755761, | |
| "loss": 0.6766, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 0.09599179029464722, | |
| "learning_rate": 0.00013034269657086992, | |
| "loss": 0.645, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "grad_norm": 0.08999059349298477, | |
| "learning_rate": 0.000127532374897626, | |
| "loss": 0.6527, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "eval_loss": 0.6689873337745667, | |
| "eval_runtime": 2.6108, | |
| "eval_samples_per_second": 1.149, | |
| "eval_steps_per_second": 1.149, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 0.13835830986499786, | |
| "learning_rate": 0.00012469832066010843, | |
| "loss": 0.6561, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "grad_norm": 0.10695886611938477, | |
| "learning_rate": 0.00012184297677777463, | |
| "loss": 0.6668, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "grad_norm": 0.0739368349313736, | |
| "learning_rate": 0.00011896880452149077, | |
| "loss": 0.643, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "grad_norm": 0.21791452169418335, | |
| "learning_rate": 0.00011607828139194683, | |
| "loss": 0.6768, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "grad_norm": 0.06241246312856674, | |
| "learning_rate": 0.00011317389898408189, | |
| "loss": 0.6252, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 0.1302526593208313, | |
| "learning_rate": 0.00011025816083936036, | |
| "loss": 0.6624, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "eval_loss": 0.6632375121116638, | |
| "eval_runtime": 2.6043, | |
| "eval_samples_per_second": 1.152, | |
| "eval_steps_per_second": 1.152, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "grad_norm": 0.11702455580234528, | |
| "learning_rate": 0.0001073335802877504, | |
| "loss": 0.6522, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "grad_norm": 0.08904154598712921, | |
| "learning_rate": 0.00010440267828126478, | |
| "loss": 0.6472, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "grad_norm": 0.08021406084299088, | |
| "learning_rate": 0.00010146798122093166, | |
| "loss": 0.6279, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "grad_norm": 0.07384659349918365, | |
| "learning_rate": 9.853201877906836e-05, | |
| "loss": 0.6262, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "grad_norm": 0.06457240134477615, | |
| "learning_rate": 9.559732171873523e-05, | |
| "loss": 0.64, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 0.07967618852853775, | |
| "learning_rate": 9.266641971224963e-05, | |
| "loss": 0.6228, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "eval_loss": 0.6625072360038757, | |
| "eval_runtime": 2.6047, | |
| "eval_samples_per_second": 1.152, | |
| "eval_steps_per_second": 1.152, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 0.09555868804454803, | |
| "learning_rate": 8.974183916063968e-05, | |
| "loss": 0.635, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "grad_norm": 0.07187359035015106, | |
| "learning_rate": 8.682610101591814e-05, | |
| "loss": 0.6277, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 0.091610848903656, | |
| "learning_rate": 8.392171860805319e-05, | |
| "loss": 0.6649, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 0.065833680331707, | |
| "learning_rate": 8.103119547850924e-05, | |
| "loss": 0.6262, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 0.09459354728460312, | |
| "learning_rate": 7.815702322222538e-05, | |
| "loss": 0.6359, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 0.06780053675174713, | |
| "learning_rate": 7.530167933989161e-05, | |
| "loss": 0.6447, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "eval_loss": 0.6616933941841125, | |
| "eval_runtime": 2.607, | |
| "eval_samples_per_second": 1.151, | |
| "eval_steps_per_second": 1.151, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "grad_norm": 0.0954224094748497, | |
| "learning_rate": 7.246762510237403e-05, | |
| "loss": 0.6636, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 0.0937703400850296, | |
| "learning_rate": 6.96573034291301e-05, | |
| "loss": 0.6381, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 0.10935033112764359, | |
| "learning_rate": 6.687313678244242e-05, | |
| "loss": 0.628, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 0.08154003322124481, | |
| "learning_rate": 6.411752507928642e-05, | |
| "loss": 0.6386, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 0.12196218967437744, | |
| "learning_rate": 6.139284362263185e-05, | |
| "loss": 0.6317, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.11538293212652206, | |
| "learning_rate": 5.870144105396118e-05, | |
| "loss": 0.6409, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.6598871350288391, | |
| "eval_runtime": 2.6073, | |
| "eval_samples_per_second": 1.151, | |
| "eval_steps_per_second": 1.151, | |
| "step": 78 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 117, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 1.2819075742826496e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |