Upload training_report.json with huggingface_hub
Browse files- training_report.json +40 -0
training_report.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"version": "BiGRU_T_version",
|
| 3 |
+
"training_params": {
|
| 4 |
+
"epochs": 2,
|
| 5 |
+
"per_device_batch_size": 1,
|
| 6 |
+
"grad_accum": 2,
|
| 7 |
+
"lr": 0.001,
|
| 8 |
+
"weight_decay": 0.01,
|
| 9 |
+
"max_grad_norm": 5.0,
|
| 10 |
+
"max_seq_len": 32,
|
| 11 |
+
"use_hypothesis": true,
|
| 12 |
+
"stop_grad_hyp": true,
|
| 13 |
+
"meta_interval": 5,
|
| 14 |
+
"meta_lr": 0.01,
|
| 15 |
+
"sharpness_lambda": 0.01
|
| 16 |
+
},
|
| 17 |
+
"results": {
|
| 18 |
+
"epochs_completed": 2,
|
| 19 |
+
"killed": false,
|
| 20 |
+
"kill_reason": null,
|
| 21 |
+
"global_step": 7,
|
| 22 |
+
"micro_batches": 14,
|
| 23 |
+
"best_loss": 7.62284517288208,
|
| 24 |
+
"final_loss": 7.702183723449707,
|
| 25 |
+
"final_avg_loss": 7.6625144481658936,
|
| 26 |
+
"final_perplexity": 2127.09919063071,
|
| 27 |
+
"elapsed_s": 45.323012351989746,
|
| 28 |
+
"params_total": 4045188,
|
| 29 |
+
"params_M": 4.045188,
|
| 30 |
+
"monitor_summary": {
|
| 31 |
+
"total_steps": 7,
|
| 32 |
+
"min_loss": 7.702183723449707,
|
| 33 |
+
"max_ram_pct": 41.8,
|
| 34 |
+
"min_disk_free_gb": 4.386983936,
|
| 35 |
+
"max_ppl": 42609.15312801587,
|
| 36 |
+
"killed": false,
|
| 37 |
+
"kill_reason": null
|
| 38 |
+
}
|
| 39 |
+
}
|
| 40 |
+
}
|