Upload LoRA adapter (README written by author)

README.md

@@ -10,33 +10,51 @@ pipeline_tag: text-generation
 tags:
 - qlora
 - lora
+- unsloth
 - structured-output
+- structeval
 ---
 
-qwen3-4b-structured-output-lora
+# qwen3-4b-structured-output-lora
 
 This repository provides a **LoRA adapter** fine-tuned from
-**Qwen/Qwen3-4B-Instruct-2507** using **QLoRA (4-bit, Unsloth)**.
+**Qwen/Qwen3-4B-Instruct-2507** using **QLoRA (4-bit) with Unsloth**.
 
-This repository contains **LoRA adapter weights only**.
+⚠️ This repository contains **LoRA adapter weights only**.
 The base model must be loaded separately.
 
 ## Training Objective
 
-This adapter is trained to improve **structured output accuracy**
+This adapter is trained to improve structured output accuracy
 (JSON / YAML / XML / TOML / CSV).
 
-Loss is applied only to the final assistant output,
-while intermediate reasoning (Chain-of-Thought) is masked.
+Loss is applied only to the final assistant output (**assistant-only loss**).
+
+Chain-of-Thought masking: Enabled  
+Learning mode: after_marker
+
+## Data Preprocessing
+
+Rule-based normalization was applied before training:
+
+- Extracting content after output markers
+- Removing code fences (```json / ```yaml / ```xml / ```toml)
+- Removing leading boilerplate and trailing notes
+- Recursive JSON exact-match deduplication
+
+Dedupe enabled: Yes
 
 ## Training Configuration
 
 - Base model: Qwen/Qwen3-4B-Instruct-2507
-- Method: QLoRA (4-bit)
+- Method: QLoRA (4-bit) + Unsloth
 - Max sequence length: 1024
 - Epochs: 1
 - Learning rate: 3e-05
-- LoRA: r=48, alpha=96
+- Warmup ratio: 0.06
+- Weight decay: 0.02
+- LoRA: r=48, alpha=96, dropout=0.06
+- Target modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj
 
 ## Usage
 
@@ -46,7 +64,7 @@ from peft import PeftModel
 import torch
 
 base = "Qwen/Qwen3-4B-Instruct-2507"
-adapter = "your_id/your-repo"
+adapter = "tropico0313/my-lora-test"
 
 tokenizer = AutoTokenizer.from_pretrained(base)
 model = AutoModelForCausalLM.from_pretrained(
@@ -55,11 +73,9 @@ model = AutoModelForCausalLM.from_pretrained(
     device_map="auto",
 )
 model = PeftModel.from_pretrained(model, adapter)
-```
-
-## Sources & Terms (IMPORTANT)
-
-Training data: u-10bei/structured_data_with_cot_dataset_512_v2
 
-Dataset License: MIT License. This dataset is used and distributed under the terms of the MIT License.
-Compliance: Users must comply with the MIT license (including copyright notice) and the base model's original terms of use.
+Sources & Terms (IMPORTANT)
+Training dataset: u-10bei/structured_data_with_cot_dataset_512_v2
+Dataset License: MIT License.
+Users must comply with the MIT license (including copyright notice)
+and the base model's original terms of use.

adapter_config.json

@@ -33,13 +33,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "gate_proj",
-    "v_proj",
     "k_proj",
-    "q_proj",
-    "up_proj",
     "o_proj",
-    "down_proj"
+    "gate_proj",
+    "down_proj",
+    "q_proj",
+    "v_proj",
+    "up_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:577a4ee44834ef2178475819f5e1648eae5c9eecaf92799b7d82c1e473ab22f9
+oid sha256:8508d63599150669c60e6d2767b972bf0657a0a474494331767cdcee3f52734d
 size 396429608