Sanitize adapter_config: keep only LoRA-loading fields, drop training-process params
Browse files- adapters/adapter_config.json +3 -35
adapters/adapter_config.json
CHANGED
|
@@ -1,41 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"adapter_path": "adapters/gemma4-31b-v6",
|
| 3 |
-
"batch_size": 2,
|
| 4 |
-
"clear_cache_threshold": 0,
|
| 5 |
-
"config": "/tmp/v6_31b_train_config.yaml",
|
| 6 |
-
"data": "benchmarks/dataset_v6",
|
| 7 |
"fine_tune_type": "lora",
|
| 8 |
-
"
|
| 9 |
-
"grad_checkpoint": true,
|
| 10 |
-
"iters": 1000,
|
| 11 |
-
"learning_rate": 3e-05,
|
| 12 |
"lora_parameters": {
|
| 13 |
"rank": 8,
|
| 14 |
"dropout": 0.0,
|
| 15 |
"scale": 20.0
|
| 16 |
-
}
|
| 17 |
-
|
| 18 |
-
"mask_prompt": true,
|
| 19 |
-
"max_seq_length": 2048,
|
| 20 |
-
"model": "mlx/gemma4-31b-q4",
|
| 21 |
-
"num_layers": 16,
|
| 22 |
-
"optimizer": "adam",
|
| 23 |
-
"optimizer_config": {
|
| 24 |
-
"adam": {},
|
| 25 |
-
"adamw": {},
|
| 26 |
-
"muon": {},
|
| 27 |
-
"sgd": {},
|
| 28 |
-
"adafactor": {}
|
| 29 |
-
},
|
| 30 |
-
"project_name": null,
|
| 31 |
-
"report_to": null,
|
| 32 |
-
"resume_adapter_file": null,
|
| 33 |
-
"save_every": 250,
|
| 34 |
-
"seed": 0,
|
| 35 |
-
"steps_per_eval": 250,
|
| 36 |
-
"steps_per_report": 25,
|
| 37 |
-
"test": false,
|
| 38 |
-
"test_batches": 500,
|
| 39 |
-
"train": true,
|
| 40 |
-
"val_batches": 25
|
| 41 |
-
}
|
|
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
"fine_tune_type": "lora",
|
| 3 |
+
"num_layers": 16,
|
|
|
|
|
|
|
|
|
|
| 4 |
"lora_parameters": {
|
| 5 |
"rank": 8,
|
| 6 |
"dropout": 0.0,
|
| 7 |
"scale": 20.0
|
| 8 |
+
}
|
| 9 |
+
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|