Upload LoRA adapter (README written by author)

README.md

@@ -36,7 +36,7 @@ while intermediate reasoning (Chain-of-Thought) is masked.
 - Max sequence length: 512
 - Epochs: 2
 - Learning rate: 2e-05
-- LoRA: r=128, alpha=256
+- LoRA: r=32, alpha=64
 
 ## Usage
 

adapter_config.json

@@ -20,7 +20,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 256,
+  "lora_alpha": 64,
   "lora_bias": false,
   "lora_dropout": 0.0,
   "megatron_config": null,
@@ -29,17 +29,17 @@
   "peft_type": "LORA",
   "peft_version": "0.18.1",
   "qalora_group_size": 16,
-  "r": 128,
+  "r": 32,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "o_proj",
-    "k_proj",
-    "q_proj",
-    "down_proj",
     "gate_proj",
     "v_proj",
-    "up_proj"
+    "k_proj",
+    "q_proj",
+    "o_proj",
+    "up_proj",
+    "down_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3fcabee6445ffd79f7bc5dbfdde3c599b18577ab53137a7f52e6001992ee477e
-size 1057033224
+oid sha256:0c237e01f6eb66e7c247cc78a45a3290b43c04b76c0ce909f4282fd64fe72316
+size 264308896