| { |
| "best_global_step": 121, |
| "best_metric": 0.4985087513923645, |
| "best_model_checkpoint": "/workspace/gemma4-26b-securecode/checkpoint-121", |
| "epoch": 1.0, |
| "eval_steps": 500, |
| "global_step": 121, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 1.0907821020111441, |
| "epoch": 0.0827300930713547, |
| "grad_norm": 20.875, |
| "learning_rate": 1.8e-05, |
| "loss": 80.26775512695312, |
| "mean_token_accuracy": 0.4542873948812485, |
| "num_tokens": 326185.0, |
| "step": 10 |
| }, |
| { |
| "entropy": 0.8271314173936843, |
| "epoch": 0.1654601861427094, |
| "grad_norm": 8.75, |
| "learning_rate": 3.8e-05, |
| "loss": 58.08096923828125, |
| "mean_token_accuracy": 0.5611657274886965, |
| "num_tokens": 653865.0, |
| "step": 20 |
| }, |
| { |
| "entropy": 0.4787554959766567, |
| "epoch": 0.2481902792140641, |
| "grad_norm": 1.7109375, |
| "learning_rate": 5.8e-05, |
| "loss": 25.493240356445312, |
| "mean_token_accuracy": 0.7378443486988544, |
| "num_tokens": 981337.0, |
| "step": 30 |
| }, |
| { |
| "entropy": 0.7855595085769892, |
| "epoch": 0.3309203722854188, |
| "grad_norm": 0.8671875, |
| "learning_rate": 7.800000000000001e-05, |
| "loss": 14.629072570800782, |
| "mean_token_accuracy": 0.7917733617126942, |
| "num_tokens": 1308584.0, |
| "step": 40 |
| }, |
| { |
| "entropy": 0.7569877350702882, |
| "epoch": 0.4136504653567735, |
| "grad_norm": 2.109375, |
| "learning_rate": 9.8e-05, |
| "loss": 12.609142303466797, |
| "mean_token_accuracy": 0.8013272784650326, |
| "num_tokens": 1635098.0, |
| "step": 50 |
| }, |
| { |
| "entropy": 0.6735223602503538, |
| "epoch": 0.4963805584281282, |
| "grad_norm": 16.875, |
| "learning_rate": 0.000118, |
| "loss": 10.704925537109375, |
| "mean_token_accuracy": 0.8209844313561916, |
| "num_tokens": 1962302.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 0.6005677949637175, |
| "epoch": 0.5791106514994829, |
| "grad_norm": 1.546875, |
| "learning_rate": 0.000138, |
| "loss": 9.783185577392578, |
| "mean_token_accuracy": 0.8308866504579783, |
| "num_tokens": 2289982.0, |
| "step": 70 |
| }, |
| { |
| "entropy": 0.5877057909965515, |
| "epoch": 0.6618407445708376, |
| "grad_norm": 11.25, |
| "learning_rate": 0.00015800000000000002, |
| "loss": 9.298844909667968, |
| "mean_token_accuracy": 0.8359990835189819, |
| "num_tokens": 2616786.0, |
| "step": 80 |
| }, |
| { |
| "entropy": 0.5447238819673658, |
| "epoch": 0.7445708376421923, |
| "grad_norm": 1.2890625, |
| "learning_rate": 0.00017800000000000002, |
| "loss": 8.777264404296876, |
| "mean_token_accuracy": 0.8440194871276617, |
| "num_tokens": 2941975.0, |
| "step": 90 |
| }, |
| { |
| "entropy": 0.5323287105187774, |
| "epoch": 0.827300930713547, |
| "grad_norm": 0.70703125, |
| "learning_rate": 0.00019800000000000002, |
| "loss": 8.489185333251953, |
| "mean_token_accuracy": 0.8486687760800123, |
| "num_tokens": 3269655.0, |
| "step": 100 |
| }, |
| { |
| "entropy": 0.4949887519702315, |
| "epoch": 0.9100310237849017, |
| "grad_norm": 0.439453125, |
| "learning_rate": 0.00019942266891397815, |
| "loss": 8.192723083496094, |
| "mean_token_accuracy": 0.8528529018163681, |
| "num_tokens": 3595193.0, |
| "step": 110 |
| }, |
| { |
| "entropy": 0.4980895221233368, |
| "epoch": 0.9927611168562565, |
| "grad_norm": 0.921875, |
| "learning_rate": 0.00019743551343638324, |
| "loss": 7.908926391601563, |
| "mean_token_accuracy": 0.8567473825067282, |
| "num_tokens": 3922475.0, |
| "step": 120 |
| }, |
| { |
| "epoch": 1.0, |
| "eval_entropy": 0.5629051625728607, |
| "eval_loss": 0.4985087513923645, |
| "eval_mean_token_accuracy": 0.8571729124978531, |
| "eval_num_tokens": 3949618.0, |
| "eval_runtime": 122.3216, |
| "eval_samples_per_second": 1.758, |
| "eval_steps_per_second": 1.758, |
| "step": 121 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 363, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 5.944883327916494e+17, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
