| { |
| "best_global_step": 105, |
| "best_metric": 0.53931481, |
| "best_model_checkpoint": "/workspace/xiruili_temporary/guava/runs/qwen35-4b-train-v6-full-r16-lr2e-5/v1-20260513-234931/checkpoint-105", |
| "epoch": 3.0, |
| "eval_steps": 25, |
| "global_step": 105, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.029197080291970802, |
| "grad_norm": 1.995608925819397, |
| "learning_rate": 3.3333333333333333e-06, |
| "loss": 1.0870462656021118, |
| "step": 1, |
| "token_acc": 0.7222708272577039 |
| }, |
| { |
| "epoch": 0.145985401459854, |
| "grad_norm": 1.630089521408081, |
| "learning_rate": 1.6666666666666667e-05, |
| "loss": 1.0861663818359375, |
| "step": 5, |
| "token_acc": 0.7234353454485584 |
| }, |
| { |
| "epoch": 0.291970802919708, |
| "grad_norm": 0.810945987701416, |
| "learning_rate": 1.9919548128307954e-05, |
| "loss": 1.0100549697875976, |
| "step": 10, |
| "token_acc": 0.7320401812893188 |
| }, |
| { |
| "epoch": 0.43795620437956206, |
| "grad_norm": 0.6168099641799927, |
| "learning_rate": 1.9594929736144978e-05, |
| "loss": 0.8932361602783203, |
| "step": 15, |
| "token_acc": 0.752801364142539 |
| }, |
| { |
| "epoch": 0.583941605839416, |
| "grad_norm": 0.5895264744758606, |
| "learning_rate": 1.9029265382866216e-05, |
| "loss": 0.8415294647216797, |
| "step": 20, |
| "token_acc": 0.7655036474215341 |
| }, |
| { |
| "epoch": 0.7299270072992701, |
| "grad_norm": 0.5221685767173767, |
| "learning_rate": 1.8236765814298328e-05, |
| "loss": 0.7946267127990723, |
| "step": 25, |
| "token_acc": 0.7786007141912445 |
| }, |
| { |
| "epoch": 0.7299270072992701, |
| "eval_loss": 0.7223935127258301, |
| "eval_runtime": 19.6664, |
| "eval_samples_per_second": 1.424, |
| "eval_steps_per_second": 0.712, |
| "eval_token_acc": 0.7937243620478398, |
| "step": 25 |
| }, |
| { |
| "epoch": 0.8759124087591241, |
| "grad_norm": 0.505389392375946, |
| "learning_rate": 1.72373403810507e-05, |
| "loss": 0.727280855178833, |
| "step": 30, |
| "token_acc": 0.7940088439236739 |
| }, |
| { |
| "epoch": 1.0, |
| "grad_norm": 0.8667115569114685, |
| "learning_rate": 1.6056096871376667e-05, |
| "loss": 0.6953943252563477, |
| "step": 35, |
| "token_acc": 0.8008854009457692 |
| }, |
| { |
| "epoch": 1.145985401459854, |
| "grad_norm": 0.4732973575592041, |
| "learning_rate": 1.472271074772683e-05, |
| "loss": 0.6909239768981934, |
| "step": 40, |
| "token_acc": 0.802294099742074 |
| }, |
| { |
| "epoch": 1.2919708029197081, |
| "grad_norm": 0.4912143349647522, |
| "learning_rate": 1.3270679633174219e-05, |
| "loss": 0.6232150554656982, |
| "step": 45, |
| "token_acc": 0.8173081972378302 |
| }, |
| { |
| "epoch": 1.437956204379562, |
| "grad_norm": 0.4965362846851349, |
| "learning_rate": 1.1736481776669307e-05, |
| "loss": 0.6203184127807617, |
| "step": 50, |
| "token_acc": 0.817360592270597 |
| }, |
| { |
| "epoch": 1.437956204379562, |
| "eval_loss": 0.5858569741249084, |
| "eval_runtime": 19.4944, |
| "eval_samples_per_second": 1.436, |
| "eval_steps_per_second": 0.718, |
| "eval_token_acc": 0.8254754674764264, |
| "step": 50 |
| }, |
| { |
| "epoch": 1.583941605839416, |
| "grad_norm": 0.5230202078819275, |
| "learning_rate": 1.015865963834808e-05, |
| "loss": 0.6270622253417969, |
| "step": 55, |
| "token_acc": 0.8175454142667258 |
| }, |
| { |
| "epoch": 1.72992700729927, |
| "grad_norm": 0.5222014784812927, |
| "learning_rate": 8.576851617267151e-06, |
| "loss": 0.5988405227661133, |
| "step": 60, |
| "token_acc": 0.8216454622561493 |
| }, |
| { |
| "epoch": 1.8759124087591241, |
| "grad_norm": 0.5679988265037537, |
| "learning_rate": 7.0307962467172555e-06, |
| "loss": 0.6027673721313477, |
| "step": 65, |
| "token_acc": 0.8212479748637636 |
| }, |
| { |
| "epoch": 2.0, |
| "grad_norm": 0.9797302484512329, |
| "learning_rate": 5.559333873942259e-06, |
| "loss": 0.5929806709289551, |
| "step": 70, |
| "token_acc": 0.828269567391848 |
| }, |
| { |
| "epoch": 2.145985401459854, |
| "grad_norm": 0.5088374018669128, |
| "learning_rate": 4.19943090428802e-06, |
| "loss": 0.5630727767944336, |
| "step": 75, |
| "token_acc": 0.830645705417988 |
| }, |
| { |
| "epoch": 2.145985401459854, |
| "eval_loss": 0.5463234782218933, |
| "eval_runtime": 19.6522, |
| "eval_samples_per_second": 1.425, |
| "eval_steps_per_second": 0.712, |
| "eval_token_acc": 0.8345319908369293, |
| "step": 75 |
| }, |
| { |
| "epoch": 2.291970802919708, |
| "grad_norm": 0.49748751521110535, |
| "learning_rate": 2.9852511229367862e-06, |
| "loss": 0.5831465721130371, |
| "step": 80, |
| "token_acc": 0.8253881640954412 |
| }, |
| { |
| "epoch": 2.437956204379562, |
| "grad_norm": 0.5103219747543335, |
| "learning_rate": 1.947297424689414e-06, |
| "loss": 0.5736560821533203, |
| "step": 85, |
| "token_acc": 0.8275780189959294 |
| }, |
| { |
| "epoch": 2.5839416058394162, |
| "grad_norm": 0.5354544520378113, |
| "learning_rate": 1.1116455134507665e-06, |
| "loss": 0.5565204620361328, |
| "step": 90, |
| "token_acc": 0.8317895383059783 |
| }, |
| { |
| "epoch": 2.72992700729927, |
| "grad_norm": 0.5385181903839111, |
| "learning_rate": 4.992888225905467e-07, |
| "loss": 0.5768057823181152, |
| "step": 95, |
| "token_acc": 0.8267348161960575 |
| }, |
| { |
| "epoch": 2.875912408759124, |
| "grad_norm": 0.5486910343170166, |
| "learning_rate": 1.2561111323605714e-07, |
| "loss": 0.5681517601013184, |
| "step": 100, |
| "token_acc": 0.8298371468398327 |
| }, |
| { |
| "epoch": 2.875912408759124, |
| "eval_loss": 0.5395660400390625, |
| "eval_runtime": 19.736, |
| "eval_samples_per_second": 1.419, |
| "eval_steps_per_second": 0.709, |
| "eval_token_acc": 0.8356507378402855, |
| "step": 100 |
| }, |
| { |
| "epoch": 3.0, |
| "grad_norm": 1.13607656955719, |
| "learning_rate": 0.0, |
| "loss": 0.5709176063537598, |
| "step": 105, |
| "token_acc": 0.8310829760807724 |
| }, |
| { |
| "epoch": 3.0, |
| "eval_loss": 0.5393148064613342, |
| "eval_runtime": 19.8847, |
| "eval_samples_per_second": 1.408, |
| "eval_steps_per_second": 0.704, |
| "eval_token_acc": 0.835810558840765, |
| "step": 105 |
| } |
| ], |
| "logging_steps": 5, |
| "max_steps": 105, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 25, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.994522975314903e+17, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
