| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 6.289308176100629, | |
| "eval_steps": 500, | |
| "global_step": 2000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6216129032258064, | |
| "eval_loss": 0.230796217918396, | |
| "eval_runtime": 2.0292, | |
| "eval_samples_per_second": 1527.677, | |
| "eval_steps_per_second": 32.032, | |
| "step": 318 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "grad_norm": 0.5734981894493103, | |
| "learning_rate": 1.606918238993711e-05, | |
| "loss": 0.3689, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8338709677419355, | |
| "eval_loss": 0.10718880593776703, | |
| "eval_runtime": 2.0538, | |
| "eval_samples_per_second": 1509.399, | |
| "eval_steps_per_second": 31.649, | |
| "step": 636 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8806451612903226, | |
| "eval_loss": 0.068777896463871, | |
| "eval_runtime": 2.2761, | |
| "eval_samples_per_second": 1362.004, | |
| "eval_steps_per_second": 28.558, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 3.14, | |
| "grad_norm": 0.44710573554039, | |
| "learning_rate": 1.2138364779874214e-05, | |
| "loss": 0.1243, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9012903225806451, | |
| "eval_loss": 0.051503442227840424, | |
| "eval_runtime": 2.2457, | |
| "eval_samples_per_second": 1380.387, | |
| "eval_steps_per_second": 28.944, | |
| "step": 1272 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "grad_norm": 0.3705664277076721, | |
| "learning_rate": 8.207547169811321e-06, | |
| "loss": 0.0768, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9135483870967742, | |
| "eval_loss": 0.04221797734498978, | |
| "eval_runtime": 2.1276, | |
| "eval_samples_per_second": 1457.071, | |
| "eval_steps_per_second": 30.551, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9167741935483871, | |
| "eval_loss": 0.03726252540946007, | |
| "eval_runtime": 2.1432, | |
| "eval_samples_per_second": 1446.46, | |
| "eval_steps_per_second": 30.329, | |
| "step": 1908 | |
| }, | |
| { | |
| "epoch": 6.29, | |
| "grad_norm": 0.28841668367385864, | |
| "learning_rate": 4.276729559748428e-06, | |
| "loss": 0.0606, | |
| "step": 2000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2544, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "total_flos": 520991326672152.0, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.5475107121282025, | |
| "num_train_epochs": 8, | |
| "temperature": 5 | |
| } | |
| } | |