| { | |
| "best_metric": 73.3085, | |
| "best_model_checkpoint": "bin/indosum-lora-0/checkpoint-4460", | |
| "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": 0.5802992582321167, | |
| "learning_rate": 0.0008, | |
| "loss": 0.8234, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 100.72533333333334, | |
| "eval_loss": 0.5383104681968689, | |
| "eval_rouge1": 70.3236, | |
| "eval_rouge2": 62.968, | |
| "eval_rougeL": 67.3562, | |
| "eval_rougeLsum": 69.3577, | |
| "eval_runtime": 218.3908, | |
| "eval_samples_per_second": 3.434, | |
| "eval_steps_per_second": 0.11, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.8689345717430115, | |
| "learning_rate": 0.0006, | |
| "loss": 0.6236, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 106.88, | |
| "eval_loss": 0.5276467800140381, | |
| "eval_rouge1": 70.7232, | |
| "eval_rouge2": 63.3489, | |
| "eval_rougeL": 67.5777, | |
| "eval_rougeLsum": 69.7735, | |
| "eval_runtime": 220.8428, | |
| "eval_samples_per_second": 3.396, | |
| "eval_steps_per_second": 0.109, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.7641115784645081, | |
| "learning_rate": 0.0004, | |
| "loss": 0.5819, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 103.876, | |
| "eval_loss": 0.501513659954071, | |
| "eval_rouge1": 72.5246, | |
| "eval_rouge2": 65.3573, | |
| "eval_rougeL": 69.5275, | |
| "eval_rougeLsum": 71.631, | |
| "eval_runtime": 219.8404, | |
| "eval_samples_per_second": 3.412, | |
| "eval_steps_per_second": 0.109, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.5158720016479492, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5563, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 104.65333333333334, | |
| "eval_loss": 0.5031823515892029, | |
| "eval_rouge1": 72.7472, | |
| "eval_rouge2": 65.6552, | |
| "eval_rougeL": 69.7436, | |
| "eval_rougeLsum": 71.8704, | |
| "eval_runtime": 218.7145, | |
| "eval_samples_per_second": 3.429, | |
| "eval_steps_per_second": 0.11, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.7064367532730103, | |
| "learning_rate": 0.0, | |
| "loss": 0.5381, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 103.344, | |
| "eval_loss": 0.4996706545352936, | |
| "eval_rouge1": 73.3085, | |
| "eval_rouge2": 66.3297, | |
| "eval_rougeL": 70.3711, | |
| "eval_rougeLsum": 72.4621, | |
| "eval_runtime": 218.4594, | |
| "eval_samples_per_second": 3.433, | |
| "eval_steps_per_second": 0.11, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.90238608932864e+16, | |
| "train_loss": 0.6246773758276696, | |
| "train_runtime": 4840.7009, | |
| "train_samples_per_second": 14.731, | |
| "train_steps_per_second": 0.921 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.90238608932864e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |