| { | |
| "best_metric": 0.8956840634346008, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-2_60-samples_config-2_full/checkpoint-48", | |
| "epoch": 24.0, | |
| "eval_steps": 500, | |
| "global_step": 69, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.34782608695652173, | |
| "grad_norm": 0.4618525207042694, | |
| "learning_rate": 1e-05, | |
| "loss": 1.3363, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.6956521739130435, | |
| "grad_norm": 0.4205852746963501, | |
| "learning_rate": 2e-05, | |
| "loss": 1.2929, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.6956521739130435, | |
| "eval_loss": 1.3116449117660522, | |
| "eval_runtime": 5.2851, | |
| "eval_samples_per_second": 2.271, | |
| "eval_steps_per_second": 2.271, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 1.391304347826087, | |
| "grad_norm": 0.4501768946647644, | |
| "learning_rate": 4e-05, | |
| "loss": 1.3056, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 1.7391304347826086, | |
| "eval_loss": 1.2842429876327515, | |
| "eval_runtime": 5.0224, | |
| "eval_samples_per_second": 2.389, | |
| "eval_steps_per_second": 2.389, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 2.0869565217391304, | |
| "grad_norm": 0.44193580746650696, | |
| "learning_rate": 6e-05, | |
| "loss": 1.3142, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 2.782608695652174, | |
| "grad_norm": 0.38786908984184265, | |
| "learning_rate": 8e-05, | |
| "loss": 1.2411, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 2.782608695652174, | |
| "eval_loss": 1.1889811754226685, | |
| "eval_runtime": 5.0179, | |
| "eval_samples_per_second": 2.391, | |
| "eval_steps_per_second": 2.391, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 3.4782608695652173, | |
| "grad_norm": 0.253694087266922, | |
| "learning_rate": 0.0001, | |
| "loss": 1.1401, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 3.8260869565217392, | |
| "eval_loss": 1.1171661615371704, | |
| "eval_runtime": 5.0143, | |
| "eval_samples_per_second": 2.393, | |
| "eval_steps_per_second": 2.393, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 4.173913043478261, | |
| "grad_norm": 0.2503201961517334, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 1.1116, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 4.869565217391305, | |
| "grad_norm": 0.3115910291671753, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 1.0766, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 4.869565217391305, | |
| "eval_loss": 1.073786973953247, | |
| "eval_runtime": 5.0168, | |
| "eval_samples_per_second": 2.392, | |
| "eval_steps_per_second": 2.392, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 5.565217391304348, | |
| "grad_norm": 0.24022558331489563, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 1.024, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 5.913043478260869, | |
| "eval_loss": 1.0269761085510254, | |
| "eval_runtime": 5.0203, | |
| "eval_samples_per_second": 2.39, | |
| "eval_steps_per_second": 2.39, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 6.260869565217392, | |
| "grad_norm": 0.2337518185377121, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 0.9874, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 6.956521739130435, | |
| "grad_norm": 0.2258654534816742, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.9524, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 6.956521739130435, | |
| "eval_loss": 0.9850594401359558, | |
| "eval_runtime": 5.0158, | |
| "eval_samples_per_second": 2.392, | |
| "eval_steps_per_second": 2.392, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 7.6521739130434785, | |
| "grad_norm": 0.19988380372524261, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 0.9285, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.9538821578025818, | |
| "eval_runtime": 5.067, | |
| "eval_samples_per_second": 2.368, | |
| "eval_steps_per_second": 2.368, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 8.347826086956522, | |
| "grad_norm": 0.15236511826515198, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.8884, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 8.695652173913043, | |
| "eval_loss": 0.9410328269004822, | |
| "eval_runtime": 5.0744, | |
| "eval_samples_per_second": 2.365, | |
| "eval_steps_per_second": 2.365, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 9.043478260869565, | |
| "grad_norm": 0.1501324474811554, | |
| "learning_rate": 9.24024048078213e-05, | |
| "loss": 0.8745, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 9.73913043478261, | |
| "grad_norm": 0.14823651313781738, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.8534, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 9.73913043478261, | |
| "eval_loss": 0.9267178177833557, | |
| "eval_runtime": 5.0691, | |
| "eval_samples_per_second": 2.367, | |
| "eval_steps_per_second": 2.367, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 10.434782608695652, | |
| "grad_norm": 0.12540778517723083, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 0.8501, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 10.782608695652174, | |
| "eval_loss": 0.9163870215415955, | |
| "eval_runtime": 5.0719, | |
| "eval_samples_per_second": 2.366, | |
| "eval_steps_per_second": 2.366, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 11.130434782608695, | |
| "grad_norm": 0.28291353583335876, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.8146, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 11.826086956521738, | |
| "grad_norm": 0.12781503796577454, | |
| "learning_rate": 8.345653031794292e-05, | |
| "loss": 0.8295, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 11.826086956521738, | |
| "eval_loss": 0.9090666174888611, | |
| "eval_runtime": 5.0684, | |
| "eval_samples_per_second": 2.368, | |
| "eval_steps_per_second": 2.368, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 12.521739130434783, | |
| "grad_norm": 0.13382355868816376, | |
| "learning_rate": 8.07830737662829e-05, | |
| "loss": 0.7979, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 12.869565217391305, | |
| "eval_loss": 0.9038975834846497, | |
| "eval_runtime": 5.0709, | |
| "eval_samples_per_second": 2.366, | |
| "eval_steps_per_second": 2.366, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 13.217391304347826, | |
| "grad_norm": 0.13056154549121857, | |
| "learning_rate": 7.795964517353735e-05, | |
| "loss": 0.8035, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 13.91304347826087, | |
| "grad_norm": 0.12095347791910172, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.796, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 13.91304347826087, | |
| "eval_loss": 0.900050163269043, | |
| "eval_runtime": 5.0683, | |
| "eval_samples_per_second": 2.368, | |
| "eval_steps_per_second": 2.368, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 14.608695652173914, | |
| "grad_norm": 0.1225428357720375, | |
| "learning_rate": 7.191855733945387e-05, | |
| "loss": 0.7651, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 14.956521739130435, | |
| "eval_loss": 0.8977439999580383, | |
| "eval_runtime": 5.0675, | |
| "eval_samples_per_second": 2.368, | |
| "eval_steps_per_second": 2.368, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 15.304347826086957, | |
| "grad_norm": 0.11379023641347885, | |
| "learning_rate": 6.873032967079561e-05, | |
| "loss": 0.7741, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.12134936451911926, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.769, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 0.8964765071868896, | |
| "eval_runtime": 5.0682, | |
| "eval_samples_per_second": 2.368, | |
| "eval_steps_per_second": 2.368, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 16.695652173913043, | |
| "grad_norm": 0.1257934421300888, | |
| "learning_rate": 6.209609477998338e-05, | |
| "loss": 0.733, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 16.695652173913043, | |
| "eval_loss": 0.8956840634346008, | |
| "eval_runtime": 5.0155, | |
| "eval_samples_per_second": 2.393, | |
| "eval_steps_per_second": 2.393, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 17.391304347826086, | |
| "grad_norm": 0.1331223100423813, | |
| "learning_rate": 5.868240888334653e-05, | |
| "loss": 0.743, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 17.73913043478261, | |
| "eval_loss": 0.8960962891578674, | |
| "eval_runtime": 5.0171, | |
| "eval_samples_per_second": 2.392, | |
| "eval_steps_per_second": 2.392, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 18.08695652173913, | |
| "grad_norm": 0.12764325737953186, | |
| "learning_rate": 5.522642316338268e-05, | |
| "loss": 0.7531, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 18.782608695652176, | |
| "grad_norm": 0.12851761281490326, | |
| "learning_rate": 5.174497483512506e-05, | |
| "loss": 0.7346, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 18.782608695652176, | |
| "eval_loss": 0.896285355091095, | |
| "eval_runtime": 5.0186, | |
| "eval_samples_per_second": 2.391, | |
| "eval_steps_per_second": 2.391, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 19.47826086956522, | |
| "grad_norm": 0.15407174825668335, | |
| "learning_rate": 4.825502516487497e-05, | |
| "loss": 0.7113, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 19.82608695652174, | |
| "eval_loss": 0.8981779217720032, | |
| "eval_runtime": 5.0167, | |
| "eval_samples_per_second": 2.392, | |
| "eval_steps_per_second": 2.392, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 20.17391304347826, | |
| "grad_norm": 0.1754920780658722, | |
| "learning_rate": 4.477357683661734e-05, | |
| "loss": 0.7159, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 20.869565217391305, | |
| "grad_norm": 0.14167214930057526, | |
| "learning_rate": 4.131759111665349e-05, | |
| "loss": 0.7027, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 20.869565217391305, | |
| "eval_loss": 0.9011878371238708, | |
| "eval_runtime": 5.0182, | |
| "eval_samples_per_second": 2.391, | |
| "eval_steps_per_second": 2.391, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 21.565217391304348, | |
| "grad_norm": 0.14583177864551544, | |
| "learning_rate": 3.790390522001662e-05, | |
| "loss": 0.6939, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 21.91304347826087, | |
| "eval_loss": 0.9043379426002502, | |
| "eval_runtime": 5.0142, | |
| "eval_samples_per_second": 2.393, | |
| "eval_steps_per_second": 2.393, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 22.26086956521739, | |
| "grad_norm": 0.1643177568912506, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.6879, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 22.956521739130434, | |
| "grad_norm": 0.1692320853471756, | |
| "learning_rate": 3.12696703292044e-05, | |
| "loss": 0.6772, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 22.956521739130434, | |
| "eval_loss": 0.9080641269683838, | |
| "eval_runtime": 5.0176, | |
| "eval_samples_per_second": 2.392, | |
| "eval_steps_per_second": 2.392, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 23.652173913043477, | |
| "grad_norm": 0.16349433362483978, | |
| "learning_rate": 2.8081442660546125e-05, | |
| "loss": 0.6706, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_loss": 0.9126665592193604, | |
| "eval_runtime": 5.0185, | |
| "eval_samples_per_second": 2.391, | |
| "eval_steps_per_second": 2.391, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "step": 69, | |
| "total_flos": 3.657773920498483e+16, | |
| "train_loss": 0.8807426310967708, | |
| "train_runtime": 1590.0697, | |
| "train_samples_per_second": 1.446, | |
| "train_steps_per_second": 0.063 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 100, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 7, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.657773920498483e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |