| { | |
| "best_metric": 0.8431240916252136, | |
| "best_model_checkpoint": "./output/training_results/C019_random_sample_llama3-8b-base_instruct_20240504_182259/checkpoint-20", | |
| "epoch": 4.0, | |
| "eval_steps": 20, | |
| "global_step": 192, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.020833333333333332, | |
| "grad_norm": 0.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.9787, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.10416666666666667, | |
| "grad_norm": 12.192626949795681, | |
| "learning_rate": 2.25e-06, | |
| "loss": 0.9575, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.20833333333333334, | |
| "grad_norm": 4.52126946078145, | |
| "learning_rate": 5.25e-06, | |
| "loss": 0.8647, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3125, | |
| "grad_norm": 4.831880513298797, | |
| "learning_rate": 9e-06, | |
| "loss": 0.8335, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.4166666666666667, | |
| "grad_norm": 4.892917165423346, | |
| "learning_rate": 1.275e-05, | |
| "loss": 0.8164, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.4166666666666667, | |
| "eval_loss": 0.8431240916252136, | |
| "eval_runtime": 1.9991, | |
| "eval_samples_per_second": 170.072, | |
| "eval_steps_per_second": 1.501, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.5208333333333334, | |
| "grad_norm": 4.415048447330865, | |
| "learning_rate": 1.3195176200175283e-05, | |
| "loss": 0.8222, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.625, | |
| "grad_norm": 4.0612946994016434, | |
| "learning_rate": 9.515676612044427e-06, | |
| "loss": 0.8995, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.7291666666666666, | |
| "grad_norm": 3.749755477511079, | |
| "learning_rate": 6.797580677308734e-06, | |
| "loss": 0.8188, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.8333333333333334, | |
| "grad_norm": 4.222046474278687, | |
| "learning_rate": 4.808575415542887e-06, | |
| "loss": 0.7962, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.8333333333333334, | |
| "eval_loss": 0.8461718559265137, | |
| "eval_runtime": 1.9687, | |
| "eval_samples_per_second": 172.702, | |
| "eval_steps_per_second": 1.524, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.9375, | |
| "grad_norm": 3.9382293828818082, | |
| "learning_rate": 3.3676619069852654e-06, | |
| "loss": 0.8768, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 1.0416666666666667, | |
| "grad_norm": 3.5738168321090593, | |
| "learning_rate": 2.334947896124909e-06, | |
| "loss": 0.721, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.1458333333333333, | |
| "grad_norm": 3.06331450723627, | |
| "learning_rate": 1.603233215095547e-06, | |
| "loss": 0.4634, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 3.6197487832931996, | |
| "learning_rate": 1.0911174606561334e-06, | |
| "loss": 0.4335, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "eval_loss": 0.8650219440460205, | |
| "eval_runtime": 1.9704, | |
| "eval_samples_per_second": 172.55, | |
| "eval_steps_per_second": 1.523, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.3541666666666667, | |
| "grad_norm": 3.7661003279348537, | |
| "learning_rate": 7.373930741131784e-07, | |
| "loss": 0.3981, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 1.4583333333333333, | |
| "grad_norm": 4.160189703076145, | |
| "learning_rate": 4.965174334325768e-07, | |
| "loss": 0.4411, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.5625, | |
| "grad_norm": 4.5757896800217885, | |
| "learning_rate": 3.349849877937343e-07, | |
| "loss": 0.4387, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "grad_norm": 3.832613476181348, | |
| "learning_rate": 2.2844505627726646e-07, | |
| "loss": 0.4578, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "eval_loss": 0.8532621264457703, | |
| "eval_runtime": 1.9686, | |
| "eval_samples_per_second": 172.711, | |
| "eval_steps_per_second": 1.524, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.7708333333333335, | |
| "grad_norm": 3.3221659720718155, | |
| "learning_rate": 1.594328760942437e-07, | |
| "loss": 0.4262, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 1.875, | |
| "grad_norm": 3.4150483727266727, | |
| "learning_rate": 1.156010161291434e-07, | |
| "loss": 0.4247, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.9791666666666665, | |
| "grad_norm": 3.7582137935644253, | |
| "learning_rate": 8.835555547373544e-08, | |
| "loss": 0.4835, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 2.0833333333333335, | |
| "grad_norm": 3.4652201585485707, | |
| "learning_rate": 7.181664349277562e-08, | |
| "loss": 0.3944, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.0833333333333335, | |
| "eval_loss": 0.848356306552887, | |
| "eval_runtime": 1.9641, | |
| "eval_samples_per_second": 173.109, | |
| "eval_steps_per_second": 1.527, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.1875, | |
| "grad_norm": 3.1934106643776103, | |
| "learning_rate": 6.203637972657601e-08, | |
| "loss": 0.3698, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 2.2916666666666665, | |
| "grad_norm": 3.1313040719559755, | |
| "learning_rate": 5.6418543066491835e-08, | |
| "loss": 0.3999, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.3958333333333335, | |
| "grad_norm": 3.455998545535417, | |
| "learning_rate": 5.329471712759216e-08, | |
| "loss": 0.3575, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 4.316899239508259, | |
| "learning_rate": 5.161995210302015e-08, | |
| "loss": 0.3997, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_loss": 0.8528212904930115, | |
| "eval_runtime": 1.9732, | |
| "eval_samples_per_second": 172.312, | |
| "eval_steps_per_second": 1.52, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.6041666666666665, | |
| "grad_norm": 3.301008618352482, | |
| "learning_rate": 5.075841465580837e-08, | |
| "loss": 0.3796, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 2.7083333333333335, | |
| "grad_norm": 3.631871358764962, | |
| "learning_rate": 5.033564114946932e-08, | |
| "loss": 0.3768, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.8125, | |
| "grad_norm": 3.5129100349792806, | |
| "learning_rate": 5.013915282607116e-08, | |
| "loss": 0.3762, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 2.9166666666666665, | |
| "grad_norm": 3.374052247691012, | |
| "learning_rate": 5.005343402153039e-08, | |
| "loss": 0.3752, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.9166666666666665, | |
| "eval_loss": 0.8572859764099121, | |
| "eval_runtime": 1.9683, | |
| "eval_samples_per_second": 172.738, | |
| "eval_steps_per_second": 1.524, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 3.0208333333333335, | |
| "grad_norm": 2.957006248440721, | |
| "learning_rate": 5.001872829857116e-08, | |
| "loss": 0.3768, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 3.125, | |
| "grad_norm": 3.215743121265744, | |
| "learning_rate": 5.000587713853837e-08, | |
| "loss": 0.3765, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 3.2291666666666665, | |
| "grad_norm": 3.202764952665669, | |
| "learning_rate": 5.0001608748597456e-08, | |
| "loss": 0.3702, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 3.3333333333333335, | |
| "grad_norm": 4.415632412034816, | |
| "learning_rate": 5.0000370319656156e-08, | |
| "loss": 0.3697, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 3.3333333333333335, | |
| "eval_loss": 0.8608274459838867, | |
| "eval_runtime": 1.9662, | |
| "eval_samples_per_second": 172.924, | |
| "eval_steps_per_second": 1.526, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 3.4375, | |
| "grad_norm": 3.3333903437391057, | |
| "learning_rate": 5.0000067945715855e-08, | |
| "loss": 0.3654, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 3.5416666666666665, | |
| "grad_norm": 3.495136224595858, | |
| "learning_rate": 5.0000009144677036e-08, | |
| "loss": 0.3523, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 3.6458333333333335, | |
| "grad_norm": 3.291266939205713, | |
| "learning_rate": 5.0000000785521776e-08, | |
| "loss": 0.3668, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "grad_norm": 3.4466892020449005, | |
| "learning_rate": 5.000000003317662e-08, | |
| "loss": 0.3636, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "eval_loss": 0.8633963465690613, | |
| "eval_runtime": 1.9654, | |
| "eval_samples_per_second": 172.993, | |
| "eval_steps_per_second": 1.526, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 3.8541666666666665, | |
| "grad_norm": 3.7460542764286364, | |
| "learning_rate": 5.000000000038355e-08, | |
| "loss": 0.3717, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 3.9583333333333335, | |
| "grad_norm": 3.076543710036104, | |
| "learning_rate": 5.000000000000018e-08, | |
| "loss": 0.3687, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 192, | |
| "total_flos": 5362900008960.0, | |
| "train_loss": 0.5116761994237701, | |
| "train_runtime": 1046.9703, | |
| "train_samples_per_second": 11.668, | |
| "train_steps_per_second": 0.183 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 192, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 20, | |
| "total_flos": 5362900008960.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |