| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.148005148005148, | |
| "eval_steps": 2000, | |
| "global_step": 40000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 4.9679536679536684e-05, | |
| "loss": 2.6056, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.935778635778636e-05, | |
| "loss": 2.1389, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.9036036036036035e-05, | |
| "loss": 1.9778, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.8714285714285714e-05, | |
| "loss": 1.8847, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_accuracy": 0.6430994438209938, | |
| "eval_loss": 1.7282191514968872, | |
| "eval_runtime": 168.5943, | |
| "eval_samples_per_second": 163.131, | |
| "eval_steps_per_second": 20.392, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.8392535392535393e-05, | |
| "loss": 1.8149, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 4.807078507078507e-05, | |
| "loss": 1.7584, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 4.774903474903475e-05, | |
| "loss": 1.7177, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 4.742728442728443e-05, | |
| "loss": 1.6753, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "eval_accuracy": 0.6748989378707684, | |
| "eval_loss": 1.55081307888031, | |
| "eval_runtime": 171.3896, | |
| "eval_samples_per_second": 160.471, | |
| "eval_steps_per_second": 20.06, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 4.710553410553411e-05, | |
| "loss": 1.6493, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 4.678378378378379e-05, | |
| "loss": 1.6196, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 4.646203346203347e-05, | |
| "loss": 1.5848, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 4.614028314028314e-05, | |
| "loss": 1.5664, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "eval_accuracy": 0.6920701992459631, | |
| "eval_loss": 1.4559696912765503, | |
| "eval_runtime": 172.8272, | |
| "eval_samples_per_second": 159.136, | |
| "eval_steps_per_second": 19.893, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 4.581853281853282e-05, | |
| "loss": 1.5418, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 4.54967824967825e-05, | |
| "loss": 1.5232, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 4.517503217503218e-05, | |
| "loss": 1.5102, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "learning_rate": 4.4853281853281856e-05, | |
| "loss": 1.4973, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "eval_accuracy": 0.7051554952905896, | |
| "eval_loss": 1.3886678218841553, | |
| "eval_runtime": 170.2873, | |
| "eval_samples_per_second": 161.509, | |
| "eval_steps_per_second": 20.189, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 4.4531531531531535e-05, | |
| "loss": 1.4758, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 4.421042471042471e-05, | |
| "loss": 1.4534, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 4.388867438867439e-05, | |
| "loss": 1.45, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 4.356692406692407e-05, | |
| "loss": 1.4371, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "eval_accuracy": 0.7145240916645529, | |
| "eval_loss": 1.3381266593933105, | |
| "eval_runtime": 171.6166, | |
| "eval_samples_per_second": 160.258, | |
| "eval_steps_per_second": 20.033, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 4.324517374517374e-05, | |
| "loss": 1.4315, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "learning_rate": 4.292342342342343e-05, | |
| "loss": 1.4166, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "learning_rate": 4.2602316602316605e-05, | |
| "loss": 1.4071, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "learning_rate": 4.228056628056628e-05, | |
| "loss": 1.3949, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "eval_accuracy": 0.7217752604656028, | |
| "eval_loss": 1.3031212091445923, | |
| "eval_runtime": 171.5333, | |
| "eval_samples_per_second": 160.336, | |
| "eval_steps_per_second": 20.043, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 4.195881595881596e-05, | |
| "loss": 1.3929, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 4.163770913770914e-05, | |
| "loss": 1.3755, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 4.131595881595882e-05, | |
| "loss": 1.3731, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 4.09942084942085e-05, | |
| "loss": 1.3656, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "eval_accuracy": 0.7279450116990841, | |
| "eval_loss": 1.2706010341644287, | |
| "eval_runtime": 172.4413, | |
| "eval_samples_per_second": 159.492, | |
| "eval_steps_per_second": 19.937, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 4.067245817245817e-05, | |
| "loss": 1.3562, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 4.0350707850707856e-05, | |
| "loss": 1.345, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "learning_rate": 4.0028957528957535e-05, | |
| "loss": 1.3401, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "learning_rate": 3.970720720720721e-05, | |
| "loss": 1.3271, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "eval_accuracy": 0.7319083261136123, | |
| "eval_loss": 1.2469459772109985, | |
| "eval_runtime": 172.8063, | |
| "eval_samples_per_second": 159.155, | |
| "eval_steps_per_second": 19.895, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "learning_rate": 3.9385456885456886e-05, | |
| "loss": 1.3179, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "learning_rate": 3.9064350064350063e-05, | |
| "loss": 1.3207, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 3.874259974259974e-05, | |
| "loss": 1.3092, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 3.842084942084943e-05, | |
| "loss": 1.3016, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "eval_accuracy": 0.7369295282061049, | |
| "eval_loss": 1.2243499755859375, | |
| "eval_runtime": 171.841, | |
| "eval_samples_per_second": 160.049, | |
| "eval_steps_per_second": 20.007, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "learning_rate": 3.80990990990991e-05, | |
| "loss": 1.2989, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 3.777734877734878e-05, | |
| "loss": 1.2902, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "learning_rate": 3.745559845559846e-05, | |
| "loss": 1.2897, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 3.713384813384814e-05, | |
| "loss": 1.2867, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "eval_accuracy": 0.7412163432241851, | |
| "eval_loss": 1.1988307237625122, | |
| "eval_runtime": 173.4853, | |
| "eval_samples_per_second": 158.532, | |
| "eval_steps_per_second": 19.817, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "learning_rate": 3.681209781209781e-05, | |
| "loss": 1.2756, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "learning_rate": 3.6490990990990994e-05, | |
| "loss": 1.273, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "learning_rate": 3.616924066924067e-05, | |
| "loss": 1.2704, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 3.5847490347490345e-05, | |
| "loss": 1.2668, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "eval_accuracy": 0.7440720045251332, | |
| "eval_loss": 1.1828033924102783, | |
| "eval_runtime": 174.2101, | |
| "eval_samples_per_second": 157.873, | |
| "eval_steps_per_second": 19.735, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 3.552574002574003e-05, | |
| "loss": 1.2682, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 3.52039897039897e-05, | |
| "loss": 1.2624, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 3.488223938223938e-05, | |
| "loss": 1.2505, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "learning_rate": 3.4561132561132566e-05, | |
| "loss": 1.2426, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "eval_accuracy": 0.7478223153253316, | |
| "eval_loss": 1.166755199432373, | |
| "eval_runtime": 174.8548, | |
| "eval_samples_per_second": 157.29, | |
| "eval_steps_per_second": 19.662, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "learning_rate": 3.423938223938224e-05, | |
| "loss": 1.2337, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 3.22, | |
| "learning_rate": 3.391763191763192e-05, | |
| "loss": 1.2389, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 3.28, | |
| "learning_rate": 3.3595881595881596e-05, | |
| "loss": 1.2314, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 3.35, | |
| "learning_rate": 3.3274131274131275e-05, | |
| "loss": 1.2284, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 3.35, | |
| "eval_accuracy": 0.7501192662767082, | |
| "eval_loss": 1.1552441120147705, | |
| "eval_runtime": 176.5773, | |
| "eval_samples_per_second": 155.756, | |
| "eval_steps_per_second": 19.47, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 3.41, | |
| "learning_rate": 3.2952380952380954e-05, | |
| "loss": 1.2244, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "learning_rate": 3.263063063063063e-05, | |
| "loss": 1.2217, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "learning_rate": 3.230888030888031e-05, | |
| "loss": 1.2204, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "learning_rate": 3.198712998712999e-05, | |
| "loss": 1.2177, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "eval_accuracy": 0.7529280009402969, | |
| "eval_loss": 1.140738606452942, | |
| "eval_runtime": 175.8979, | |
| "eval_samples_per_second": 156.358, | |
| "eval_steps_per_second": 19.545, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "learning_rate": 3.166537966537967e-05, | |
| "loss": 1.2126, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 3.73, | |
| "learning_rate": 3.134427284427285e-05, | |
| "loss": 1.2123, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "learning_rate": 3.1022522522522526e-05, | |
| "loss": 1.2015, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 3.86, | |
| "learning_rate": 3.07014157014157e-05, | |
| "loss": 1.2045, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 3.86, | |
| "eval_accuracy": 0.7548470260547614, | |
| "eval_loss": 1.1305773258209229, | |
| "eval_runtime": 177.469, | |
| "eval_samples_per_second": 154.974, | |
| "eval_steps_per_second": 19.372, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 3.93, | |
| "learning_rate": 3.037966537966538e-05, | |
| "loss": 1.2081, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 3.99, | |
| "learning_rate": 3.0057915057915058e-05, | |
| "loss": 1.2009, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 4.05, | |
| "learning_rate": 2.973616473616474e-05, | |
| "loss": 1.191, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "learning_rate": 2.9414414414414416e-05, | |
| "loss": 1.1901, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "eval_accuracy": 0.7579209971336668, | |
| "eval_loss": 1.1146818399429321, | |
| "eval_runtime": 176.2162, | |
| "eval_samples_per_second": 156.075, | |
| "eval_steps_per_second": 19.51, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "learning_rate": 2.9092664092664095e-05, | |
| "loss": 1.1939, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "learning_rate": 2.877091377091377e-05, | |
| "loss": 1.179, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 4.31, | |
| "learning_rate": 2.8449163449163453e-05, | |
| "loss": 1.1831, | |
| "step": 33500 | |
| }, | |
| { | |
| "epoch": 4.38, | |
| "learning_rate": 2.812741312741313e-05, | |
| "loss": 1.1801, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 4.38, | |
| "eval_accuracy": 0.7587203295541968, | |
| "eval_loss": 1.110528588294983, | |
| "eval_runtime": 177.1827, | |
| "eval_samples_per_second": 155.224, | |
| "eval_steps_per_second": 19.404, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "learning_rate": 2.7805662805662808e-05, | |
| "loss": 1.1761, | |
| "step": 34500 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "learning_rate": 2.7483912483912483e-05, | |
| "loss": 1.1767, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 4.57, | |
| "learning_rate": 2.7162805662805664e-05, | |
| "loss": 1.166, | |
| "step": 35500 | |
| }, | |
| { | |
| "epoch": 4.63, | |
| "learning_rate": 2.6841055341055343e-05, | |
| "loss": 1.167, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 4.63, | |
| "eval_accuracy": 0.7609539556365092, | |
| "eval_loss": 1.1016037464141846, | |
| "eval_runtime": 176.8657, | |
| "eval_samples_per_second": 155.502, | |
| "eval_steps_per_second": 19.438, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 4.7, | |
| "learning_rate": 2.6519305019305018e-05, | |
| "loss": 1.161, | |
| "step": 36500 | |
| }, | |
| { | |
| "epoch": 4.76, | |
| "learning_rate": 2.61975546975547e-05, | |
| "loss": 1.1658, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 2.5876447876447878e-05, | |
| "loss": 1.1656, | |
| "step": 37500 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "learning_rate": 2.5554697554697553e-05, | |
| "loss": 1.1543, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "eval_accuracy": 0.7620338199796781, | |
| "eval_loss": 1.0933340787887573, | |
| "eval_runtime": 176.4351, | |
| "eval_samples_per_second": 155.882, | |
| "eval_steps_per_second": 19.486, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "learning_rate": 2.5233590733590734e-05, | |
| "loss": 1.1568, | |
| "step": 38500 | |
| }, | |
| { | |
| "epoch": 5.02, | |
| "learning_rate": 2.4911840411840413e-05, | |
| "loss": 1.1591, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 5.08, | |
| "learning_rate": 2.459009009009009e-05, | |
| "loss": 1.1485, | |
| "step": 39500 | |
| }, | |
| { | |
| "epoch": 5.15, | |
| "learning_rate": 2.426833976833977e-05, | |
| "loss": 1.1481, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 5.15, | |
| "eval_accuracy": 0.7639144921830399, | |
| "eval_loss": 1.0837804079055786, | |
| "eval_runtime": 179.9041, | |
| "eval_samples_per_second": 152.876, | |
| "eval_steps_per_second": 19.11, | |
| "step": 40000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 77700, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 2000, | |
| "total_flos": 1.6967761699709952e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |