Upload LoRA adapter (README written by author)

README.md

@@ -34,9 +34,9 @@ while intermediate reasoning (Chain-of-Thought) is masked.
 - Base model: Qwen/Qwen3-4B-Instruct-2507
 - Method: QLoRA (4-bit)
 - Max sequence length: 512
-- Epochs: 3
-- Learning rate: 1e-05
-- LoRA: r=64, alpha=64
+- Epochs: 2
+- Learning rate: 2e-05
+- LoRA: r=64, alpha=128
 
 ## Usage
 

adapter_config.json

@@ -20,7 +20,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 64,
+  "lora_alpha": 128,
   "lora_bias": false,
   "lora_dropout": 0.0,
   "megatron_config": null,
@@ -33,13 +33,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
     "k_proj",
+    "up_proj",
     "o_proj",
-    "gate_proj",
+    "q_proj",
+    "down_proj",
     "v_proj",
-    "up_proj",
-    "down_proj"
+    "gate_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:73b7735c2ec1c4c6e11e2595074e04628b34112c639d28addf946762dbb0d27c
+oid sha256:ed019f79a1b0c0c26a7f923a9d33a6d6b9dae671eb8ee45934c2ae5c64129995
 size 528550256