Upload adapter_config.json with huggingface_hub
Browse files- adapter_config.json +23 -0
adapter_config.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"adapter_config": {
|
| 3 |
+
"r": 16,
|
| 4 |
+
"lora_alpha": 32,
|
| 5 |
+
"lora_dropout": 0.05,
|
| 6 |
+
"target_modules": [
|
| 7 |
+
"q_proj",
|
| 8 |
+
"k_proj",
|
| 9 |
+
"v_proj",
|
| 10 |
+
"o_proj"
|
| 11 |
+
],
|
| 12 |
+
"description": "Balanced performance and efficiency"
|
| 13 |
+
},
|
| 14 |
+
"training_config": {
|
| 15 |
+
"dataset_format": "conversations",
|
| 16 |
+
"max_seq_length": 2048,
|
| 17 |
+
"num_epochs": 3,
|
| 18 |
+
"batch_size": 4,
|
| 19 |
+
"learning_rate": 0.0002
|
| 20 |
+
},
|
| 21 |
+
"base_model": "Qwen/Qwen2.5-3B-Instruct",
|
| 22 |
+
"created_by": "MLResearcher"
|
| 23 |
+
}
|