Upload LoRA adapter (README written by author)

README.md

@@ -13,19 +13,7 @@ tags:
 - structured-output
 ---
 
-## Model Description
-
-This adapter was trained for the StructEval competition
-to improve structured data generation accuracy.
-
-## Base Model License
-
-The base model follows the Apache 2.0 license.
-
-
-
-
-
+＜【課題】ここは自分で記入して下さい＞
 
 This repository provides a **LoRA adapter** fine-tuned from
 **Qwen/Qwen3-4B-Instruct-2507** using **QLoRA (4-bit, Unsloth)**.
@@ -45,10 +33,10 @@ while intermediate reasoning (Chain-of-Thought) is masked.
 
 - Base model: Qwen/Qwen3-4B-Instruct-2507
 - Method: QLoRA (4-bit)
-- Max sequence length: 2048
-- Epochs: 3
-- Learning rate: 2e-06
-- LoRA: r=64, alpha=128
+- Max sequence length: 3072
+- Epochs: 6
+- Learning rate: 1e-05
+- LoRA: r=128, alpha=128
 
 ## Usage
 

adapter_config.json

@@ -29,17 +29,17 @@
   "peft_type": "LORA",
   "peft_version": "0.18.1",
   "qalora_group_size": 16,
-  "r": 64,
+  "r": 128,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "k_proj",
-    "q_proj",
     "v_proj",
+    "down_proj",
+    "o_proj",
+    "k_proj",
     "up_proj",
     "gate_proj",
-    "down_proj",
-    "o_proj"
+    "q_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:969c15d2e63ee5b70dfc48fd3777b06a4faa740a63e41cfcdce4ffcd1d50e4bd
-size 528550256
+oid sha256:a69d8a5e5919266f0170395d0285e4a05db652dd4766f153d77c5a92213f1c60
+size 1057033224