Upload LoRA adapter (README written by author)

README.md

@@ -35,8 +35,8 @@ while intermediate reasoning (Chain-of-Thought) is masked.
 - Method: QLoRA (4-bit)
 - Max sequence length: 2048
 - Epochs: 1
-- Learning rate: 1e-06
-- LoRA: r=64, alpha=128
+- Learning rate: 2e-05
+- LoRA: r=128, alpha=128
 
 ## Usage
 
@@ -46,7 +46,7 @@ from peft import PeftModel
 import torch
 
 base = "Qwen/Qwen3-4B-Instruct-2507"
-adapter = "your_id/your-repo"
+adapter = "saeoshi/m_tune_model3"
 
 tokenizer = AutoTokenizer.from_pretrained(base)
 model = AutoModelForCausalLM.from_pretrained(

adapter_config.json

@@ -22,24 +22,24 @@
   "loftq_config": {},
   "lora_alpha": 128,
   "lora_bias": false,
-  "lora_dropout": 0.0,
+  "lora_dropout": 0.05,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
   "peft_version": "0.18.1",
   "qalora_group_size": 16,
-  "r": 64,
+  "r": 128,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "up_proj",
     "v_proj",
-    "gate_proj",
+    "up_proj",
     "o_proj",
-    "k_proj",
     "down_proj",
-    "q_proj"
+    "q_proj",
+    "k_proj",
+    "gate_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2a080bba910615cc8d95b7bc21e9c8867b82e9c39cb6382738dff2567689efaa
-size 528550256
+oid sha256:93622922fb440b349d72395b05c72e3fbaacf860545f428cc9691aa274f7ec23
+size 1057033224