| { | |
| "best_metric": 72.3075, | |
| "best_model_checkpoint": "bin/indosum-lora-3/checkpoint-2676", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 4460, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.5876893401145935, | |
| "learning_rate": 0.0008, | |
| "loss": 0.8274, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 100.152, | |
| "eval_loss": 0.5684335827827454, | |
| "eval_rouge1": 69.5991, | |
| "eval_rouge2": 62.0811, | |
| "eval_rougeL": 66.2103, | |
| "eval_rougeLsum": 68.5179, | |
| "eval_runtime": 219.0061, | |
| "eval_samples_per_second": 3.425, | |
| "eval_steps_per_second": 0.11, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.4317479431629181, | |
| "learning_rate": 0.0006, | |
| "loss": 0.6237, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 104.78133333333334, | |
| "eval_loss": 0.5486577153205872, | |
| "eval_rouge1": 71.3379, | |
| "eval_rouge2": 64.0461, | |
| "eval_rougeL": 68.1399, | |
| "eval_rougeLsum": 70.3424, | |
| "eval_runtime": 219.5178, | |
| "eval_samples_per_second": 3.417, | |
| "eval_steps_per_second": 0.109, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.5349451303482056, | |
| "learning_rate": 0.0004, | |
| "loss": 0.5824, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 101.356, | |
| "eval_loss": 0.5441932678222656, | |
| "eval_rouge1": 72.3075, | |
| "eval_rouge2": 65.3703, | |
| "eval_rougeL": 69.1638, | |
| "eval_rougeLsum": 71.3776, | |
| "eval_runtime": 219.0561, | |
| "eval_samples_per_second": 3.424, | |
| "eval_steps_per_second": 0.11, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.44907253980636597, | |
| "learning_rate": 0.0002, | |
| "loss": 0.555, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 103.64666666666666, | |
| "eval_loss": 0.5253363251686096, | |
| "eval_rouge1": 72.2139, | |
| "eval_rouge2": 65.2916, | |
| "eval_rougeL": 69.2278, | |
| "eval_rougeLsum": 71.2919, | |
| "eval_runtime": 219.9659, | |
| "eval_samples_per_second": 3.41, | |
| "eval_steps_per_second": 0.109, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.3915654420852661, | |
| "learning_rate": 0.0, | |
| "loss": 0.5391, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 102.11733333333333, | |
| "eval_loss": 0.5204364657402039, | |
| "eval_rouge1": 72.0194, | |
| "eval_rouge2": 64.9755, | |
| "eval_rougeL": 68.9244, | |
| "eval_rougeLsum": 71.0549, | |
| "eval_runtime": 218.95, | |
| "eval_samples_per_second": 3.425, | |
| "eval_steps_per_second": 0.11, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.90582346563584e+16, | |
| "train_loss": 0.6255005293362879, | |
| "train_runtime": 4838.9085, | |
| "train_samples_per_second": 14.747, | |
| "train_steps_per_second": 0.922 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.90582346563584e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |