Upload LoRA adapter (README written by author)
Browse files- README.md +3 -3
- adapter_config.json +5 -5
- adapter_model.safetensors +2 -2
README.md
CHANGED
|
@@ -34,9 +34,9 @@ while intermediate reasoning (Chain-of-Thought) is masked.
|
|
| 34 |
- Base model: Qwen/Qwen3-4B-Instruct-2507
|
| 35 |
- Method: QLoRA (4-bit)
|
| 36 |
- Max sequence length: 512
|
| 37 |
-
- Epochs:
|
| 38 |
-
- Learning rate:
|
| 39 |
-
- LoRA: r=
|
| 40 |
|
| 41 |
## Usage
|
| 42 |
|
|
|
|
| 34 |
- Base model: Qwen/Qwen3-4B-Instruct-2507
|
| 35 |
- Method: QLoRA (4-bit)
|
| 36 |
- Max sequence length: 512
|
| 37 |
+
- Epochs: 2
|
| 38 |
+
- Learning rate: 2e-05
|
| 39 |
+
- LoRA: r=32, alpha=128
|
| 40 |
|
| 41 |
## Usage
|
| 42 |
|
adapter_config.json
CHANGED
|
@@ -29,17 +29,17 @@
|
|
| 29 |
"peft_type": "LORA",
|
| 30 |
"peft_version": "0.18.1",
|
| 31 |
"qalora_group_size": 16,
|
| 32 |
-
"r":
|
| 33 |
"rank_pattern": {},
|
| 34 |
"revision": null,
|
| 35 |
"target_modules": [
|
| 36 |
"q_proj",
|
|
|
|
|
|
|
| 37 |
"up_proj",
|
| 38 |
-
"o_proj",
|
| 39 |
"gate_proj",
|
| 40 |
-
"
|
| 41 |
-
"
|
| 42 |
-
"v_proj"
|
| 43 |
],
|
| 44 |
"target_parameters": null,
|
| 45 |
"task_type": "CAUSAL_LM",
|
|
|
|
| 29 |
"peft_type": "LORA",
|
| 30 |
"peft_version": "0.18.1",
|
| 31 |
"qalora_group_size": 16,
|
| 32 |
+
"r": 32,
|
| 33 |
"rank_pattern": {},
|
| 34 |
"revision": null,
|
| 35 |
"target_modules": [
|
| 36 |
"q_proj",
|
| 37 |
+
"v_proj",
|
| 38 |
+
"down_proj",
|
| 39 |
"up_proj",
|
|
|
|
| 40 |
"gate_proj",
|
| 41 |
+
"o_proj",
|
| 42 |
+
"k_proj"
|
|
|
|
| 43 |
],
|
| 44 |
"target_parameters": null,
|
| 45 |
"task_type": "CAUSAL_LM",
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6499c34c1078d2e150980a90c68c9dceb202674e0e23be9a4f229db22f66b9e9
|
| 3 |
+
size 264308896
|