Upload LoRA adapter (README written by author)

README.md

@@ -35,7 +35,7 @@ while intermediate reasoning (Chain-of-Thought) is masked.
 - Method: QLoRA (4-bit)
 - Max sequence length: 512
 - Epochs: 2
-- Learning rate: 3e-06
+- Learning rate: 6e-06
 - LoRA: r=64, alpha=128
 
 ## Usage

adapter_config.json

@@ -24,13 +24,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "k_proj",
+    "gate_proj",
     "q_proj",
-    "o_proj",
+    "k_proj",
     "v_proj",
     "up_proj",
-    "down_proj",
-    "gate_proj"
+    "o_proj",
+    "down_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f8beff377ac91c3168066a3909522b10d03f7b2f1cec23b8e7a1059ff5474ae5
+oid sha256:e0de57e57e082b8585d2b893b2dc1cbe244a950453d32f15999e0e1a1a8daa90
 size 528550256