| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 70, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.14285714285714285, | |
| "grad_norm": 183.0, | |
| "learning_rate": 2.857142857142857e-05, | |
| "loss": 47.2142, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.7142857142857143, | |
| "grad_norm": 97.5, | |
| "learning_rate": 0.00014285714285714287, | |
| "loss": 44.3098, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 13.720385551452637, | |
| "eval_runtime": 0.2481, | |
| "eval_samples_per_second": 40.305, | |
| "eval_steps_per_second": 4.031, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 1.4285714285714286, | |
| "grad_norm": 18.375, | |
| "learning_rate": 0.00019888308262251285, | |
| "loss": 25.7366, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 8.691579818725586, | |
| "eval_runtime": 0.2437, | |
| "eval_samples_per_second": 41.04, | |
| "eval_steps_per_second": 4.104, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.142857142857143, | |
| "grad_norm": 4.53125, | |
| "learning_rate": 0.00019214762118704076, | |
| "loss": 20.6774, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 2.857142857142857, | |
| "grad_norm": 4.46875, | |
| "learning_rate": 0.00017971325072229226, | |
| "loss": 19.0375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 7.630770683288574, | |
| "eval_runtime": 0.2338, | |
| "eval_samples_per_second": 42.776, | |
| "eval_steps_per_second": 4.278, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 3.571428571428571, | |
| "grad_norm": 8.375, | |
| "learning_rate": 0.00016234898018587337, | |
| "loss": 18.2973, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 7.119815826416016, | |
| "eval_runtime": 0.2413, | |
| "eval_samples_per_second": 41.435, | |
| "eval_steps_per_second": 4.143, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 4.285714285714286, | |
| "grad_norm": 12.3125, | |
| "learning_rate": 0.00014112871031306119, | |
| "loss": 16.6166, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 16.625, | |
| "learning_rate": 0.00011736481776669306, | |
| "loss": 14.8387, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 6.846951484680176, | |
| "eval_runtime": 0.2339, | |
| "eval_samples_per_second": 42.753, | |
| "eval_steps_per_second": 4.275, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 5.714285714285714, | |
| "grad_norm": 23.5, | |
| "learning_rate": 9.252699064135758e-05, | |
| "loss": 12.5684, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 6.8495354652404785, | |
| "eval_runtime": 0.2342, | |
| "eval_samples_per_second": 42.698, | |
| "eval_steps_per_second": 4.27, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 6.428571428571429, | |
| "grad_norm": 26.0, | |
| "learning_rate": 6.815133497483157e-05, | |
| "loss": 9.6308, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 6.679884910583496, | |
| "eval_runtime": 0.2349, | |
| "eval_samples_per_second": 42.574, | |
| "eval_steps_per_second": 4.257, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 7.142857142857143, | |
| "grad_norm": 26.25, | |
| "learning_rate": 4.574537361342407e-05, | |
| "loss": 7.3589, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 7.857142857142857, | |
| "grad_norm": 23.125, | |
| "learning_rate": 2.669481281701739e-05, | |
| "loss": 5.7187, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 6.081761360168457, | |
| "eval_runtime": 0.2356, | |
| "eval_samples_per_second": 42.45, | |
| "eval_steps_per_second": 4.245, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 8.571428571428571, | |
| "grad_norm": 22.625, | |
| "learning_rate": 1.2177842662977135e-05, | |
| "loss": 4.8487, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 6.031832695007324, | |
| "eval_runtime": 0.2458, | |
| "eval_samples_per_second": 40.691, | |
| "eval_steps_per_second": 4.069, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 9.285714285714286, | |
| "grad_norm": 22.25, | |
| "learning_rate": 3.092271377092215e-06, | |
| "loss": 4.4729, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 21.75, | |
| "learning_rate": 0.0, | |
| "loss": 4.4303, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 6.032225608825684, | |
| "eval_runtime": 0.2341, | |
| "eval_samples_per_second": 42.713, | |
| "eval_steps_per_second": 4.271, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 70, | |
| "total_flos": 2.1344245100262195e+17, | |
| "train_loss": 14.937398610796247, | |
| "train_runtime": 171.2768, | |
| "train_samples_per_second": 25.573, | |
| "train_steps_per_second": 0.409 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 70, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 100, | |
| "total_flos": 2.1344245100262195e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |