Upload LoRA adapter (Fixed README metadata)

README.md

@@ -1,9 +1,11 @@
 ---
 base_model: Qwen/Qwen3-4B-Instruct-2507
 datasets:
-- daichira/structured-hard-sft-4k
 - u-10bei/structured_data_with_cot_dataset_512
 - u-10bei/structured_data_with_cot_dataset_512_v2
+- u-10bei/structured_data_with_cot_dataset_512_v4
+- u-10bei/structured_data_with_cot_dataset_512_v5
+- u-10bei/structured_data_with_cot_dataset_v2
 language:
 - en
 license: apache-2.0
@@ -15,7 +17,7 @@ tags:
 - structured-output
 ---
 
-# Qwen3-4B RSLoRA (R=256) for Structured Output
+# Qwen3-4B_R_64_ALPHA_128 for Structured Output
 
 This repository provides a **LoRA adapter** fine-tuned from
 **Qwen/Qwen3-4B-Instruct-2507** using **QLoRA (4-bit, Unsloth)**.
@@ -36,9 +38,9 @@ while intermediate reasoning (Chain-of-Thought) is masked.
 - Base model: Qwen/Qwen3-4B-Instruct-2507
 - Method: QLoRA (4-bit)
 - Max sequence length: 512
-- Epochs: 1
-- Learning rate: 2e-04
-- LoRA: r=128, alpha=24
+- Epochs: 5
+- Learning rate: 1e-06
+- LoRA: r=64, alpha=128
 
 ## Usage
 
@@ -61,7 +63,7 @@ model = PeftModel.from_pretrained(model, adapter)
 
 ## Sources & Terms (IMPORTANT)
 
-Training data: ['daichira/structured-hard-sft-4k', 'u-10bei/structured_data_with_cot_dataset_512', 'u-10bei/structured_data_with_cot_dataset_512_v2']
+Training data: ['u-10bei/structured_data_with_cot_dataset_512', 'u-10bei/structured_data_with_cot_dataset_512_v2u-10bei/structured_data_with_cot_dataset_512_v4', 'u-10bei/structured_data_with_cot_dataset_512_v5', 'u-10bei/structured_data_with_cot_dataset_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.

adapter_config.json

@@ -20,7 +20,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 24,
+  "lora_alpha": 128,
   "lora_bias": false,
   "lora_dropout": 0.0,
   "megatron_config": null,
@@ -29,22 +29,22 @@
   "peft_type": "LORA",
   "peft_version": "0.18.1",
   "qalora_group_size": 16,
-  "r": 128,
+  "r": 64,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
+    "o_proj",
     "down_proj",
-    "k_proj",
     "up_proj",
-    "q_proj",
-    "o_proj",
+    "gate_proj",
     "v_proj",
-    "gate_proj"
+    "k_proj",
+    "q_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",
   "trainable_token_indices": null,
   "use_dora": false,
   "use_qalora": false,
-  "use_rslora": true
+  "use_rslora": false
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b0e60c482cea3edf2d00118f32c08cf82ca46f1acc8cbe48cbb11cd70e1befe
-size 1057033224
+oid sha256:95c8105145a3402f947b1e3d60108cbe616a2a6c57ac68085e7d22be509433ac
+size 528550256