Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
```
|
| 2 |
+
{
|
| 3 |
+
"cache_dir": "/leonardo_work/EUHPC_E03_068/.cache",
|
| 4 |
+
"method": "orpo",
|
| 5 |
+
"dataset": "autoredteam",
|
| 6 |
+
"model": "togethercomputer/RedPajama-INCITE-Base-3B-v1",
|
| 7 |
+
"tokenizer": "togethercomputer/RedPajama-INCITE-Instruct-3B-v1",
|
| 8 |
+
"lr": 0.01,
|
| 9 |
+
"train_batch_size": 5,
|
| 10 |
+
"eval_batch_size": 5,
|
| 11 |
+
"num_epochs": 3,
|
| 12 |
+
"seed": 42,
|
| 13 |
+
"eval_only": false,
|
| 14 |
+
"evaluation_size": 2000,
|
| 15 |
+
"checkpoint_path": null,
|
| 16 |
+
"experiment_name": "RedPajama3b_v1-autoredteam_v2-train-3",
|
| 17 |
+
"experiment_group": "results",
|
| 18 |
+
"reference_model": "togethercomputer/RedPajama-INCITE-Instruct-3B-v1",
|
| 19 |
+
"context_length": 1024,
|
| 20 |
+
"train_summarization": "",
|
| 21 |
+
"dpo_beta": 0.1,
|
| 22 |
+
"orpo_beta": 0.1,
|
| 23 |
+
"kl_coef": 0.0,
|
| 24 |
+
"reward_model": "",
|
| 25 |
+
"bestofn_size": 4,
|
| 26 |
+
"train_reward_model": ""
|
| 27 |
+
}
|
| 28 |
+
```
|