Darwin-2B-Opus: V3 merged (Qwen3.5-2B + Opus/Sonnet/Korean reasoning LoRA)

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,183 @@
+---
+license: apache-2.0
+base_model: Qwen/Qwen3.5-2B
+tags:
+- darwin
+- darwin-v8
+- opus-distilled
+- qwen3.5
+- reasoning
+- korean
+- claude-opus
+- lora-merged
+language:
+- en
+- ko
+- zh
+- ja
+pipeline_tag: text-generation
+library_name: transformers
+---
+
+# 🧠 Darwin-2B-Opus
+
+**Darwin V8 시리즈의 2B 경량 모델**
+Claude Opus 4.5/4.6 및 Sonnet 4.6의 추론 스타일을 주입한 Qwen3.5-2B 기반 모델.
+
+---
+
+## 🧬 가계도 (Pedigree)
+
+- 👨 **Father (Base)**: [`Qwen/Qwen3.5-2B`](https://huggingface.co/Qwen/Qwen3.5-2B)
+- 👩 **Mother (LoRA Adapter)**: [`FINAL-Bench/Darwin-2B-Opus-LoRA`](https://huggingface.co/FINAL-Bench/Darwin-2B-Opus-LoRA)
+- 👶 **Child (This model)**: `FINAL-Bench/Darwin-2B-Opus` — merged full-weight standalone
+
+---
+
+## 🏆 Darwin V8 시리즈 정보
+
+| 항목 | 값 |
+|------|-----|
+| 모델 크기 | 2.3B 파라미터 |
+| 아키텍처 | Qwen3.5 (hybrid attention) |
+| 학습 방식 | SFT with LoRA (all-linear, rank=16) |
+| 학습 데이터 | 9,762 샘플 (Claude Opus/Sonnet + 한국어 reasoning) |
+| 학습 시간 | 29분 (8×B200 GPU) |
+| 최종 Loss | 0.837 |
+| Token Accuracy | 76.6% |
+
+### 📊 벤치마크 (GPQA Diamond 198)
+
+- **정확도**: 37.37% (74/198)
+- **답변 추출 성공률 기준 정답률**: 50.7%
+
+---
+
+## 🚀 빠른 사용법
+
+```python
+from transformers import AutoTokenizer, AutoModelForCausalLM
+import torch
+
+model_id = "FINAL-Bench/Darwin-2B-Opus"
+tok = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id, torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True
+)
+
+messages = [
+    {"role": "user", "content": "2024년 한국 최저시급 9,860원이다. 주 40시간 × 4주 임금은?"}
+]
+prompt = tok.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+inputs = tok(prompt, return_tensors="pt").to(model.device)
+
+with torch.no_grad():
+    outputs = model.generate(
+        **inputs,
+        max_new_tokens=800,
+        do_sample=False,
+        pad_token_id=tok.eos_token_id,
+    )
+print(tok.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))
+```
+
+---
+
+## 🧬 Darwin V8 학습 파이프라인
+
+```
+[Qwen/Qwen3.5-2B] ──── Base 모델 (동결)
+        +
+[9,762 Claude Opus/Sonnet + 한국어 Reasoning 샘플]
+        ↓
+[SFT Training]
+  - LoRA (all-linear, r=16, α=32)
+  - Learning rate: 2e-4 (V8 rule: ×10 FullFT)
+  - 2 epochs, bf16, 8×B200 DDP
+  - Loss: 0.991 → 0.837 (-15%)
+  - Token accuracy: 73.9% → 76.6% (+2.7%p)
+        ↓
+[LoRA merge into base weights]
+        ↓
+[Darwin-2B-Opus] ← 이 모델
+```
+
+---
+
+## 📊 학습 데이터 구성
+
+| 카테고리 | 샘플 수 | % | 출처 |
+|---------|--------|---|-----|
+| General Reasoning | 4,422 | 45% | Opus 4.5/4.6, Sonnet 4.6 |
+| Math (English) | 1,960 | 20% | DeepSeek-v3.2 OpenR1-Math |
+| Code (English) | 1,680 | 17% | DeepSeek-v3.2 CodeReasoning + GPT-5 Codex |
+| Korean Thinking | 200 | 2% | Multilingual-Thinking-Korean |
+| **Korean Math** | **1,500** | **15%** | orca-math-word-problems-korean |
+| **합계 (필터 후)** | **9,762** | 100% | - |
+
+---
+
+## 🎯 Darwin V8 설계 철학
+
+1. **LoRA Without Regret** — `all-linear` target, LR × 10, rank=16으로 충분
+2. **Response Distillation** — Pre-generated Opus traces로 비용 효율적 증류
+3. **한국어 Reasoning 강화** — KoAlpaca 간단 QA 대신 Claude 추론 궤적 사용
+4. **Merge-and-Deploy** — LoRA 어댑터 통합 후 추가 의존성 없이 배포
+
+---
+
+## 📝 샘플 테스트 결과 (5문제)
+
+| 유형 | 정답 | 비고 |
+|-----|:---:|-----|
+| 영어 수학 (기차 속도) | ✅ 80 km/h | LaTeX 단계별 풀이 |
+| 영어 논리 (키 비교) | ✅ Carol | 추이율 명시 |
+| 영어 코드 (소수 판별) | ✅ 정확 | docstring + 복잡도 분석 |
+| **한국어 시급 계산** | ✅ **1,577,600원** | 단계별 한국어 설명 |
+| **한국어 연립방정식** | ✅ **1,200원** | 정석 풀이 + 검증 |
+
+**5/5 정답** — 영어+한국어 모두 완벽 ⭐
+
+---
+
+## ⚠️ 제한 사항
+
+- **규모**: 2.3B 파라미터 (Darwin 시리즈 최소)
+- **GPQA Diamond**: 37.37% (대형 모델 대비 낮지만 2B 중 최고 수준)
+- **긴 컨텍스트**: 학습 시 `max_length=4,096`로 학습됨
+- **지식 한계**: 2B 모델은 백과사전적 지식 한계 있음
+
+---
+
+## 🔗 관련 모델
+
+- 🧩 [`FINAL-Bench/Darwin-2B-Opus-LoRA`](https://huggingface.co/FINAL-Bench/Darwin-2B-Opus-LoRA) — 이 모델의 **LoRA 어댑터 단독 버전** (67MB)
+- ⚡ [`FINAL-Bench/Darwin-2B-Opus-ONNX`](https://huggingface.co/FINAL-Bench/Darwin-2B-Opus-ONNX) — **브라우저/WebGPU용 ONNX 양자화 버전** (예정)
+
+### 🏆 Darwin 시리즈
+- [`Darwin-31B-Opus`](https://huggingface.co/FINAL-Bench/Darwin-31B-Opus) — GPQA 85.9%
+- [`Darwin-27B-Opus`](https://huggingface.co/FINAL-Bench/Darwin-27B-Opus) — GPQA 86.9%
+- [`Darwin-9B-Opus`](https://huggingface.co/FINAL-Bench/Darwin-9B-Opus)
+- [`Darwin-4B-Opus`](https://huggingface.co/FINAL-Bench/Darwin-4B-Opus)
+- **Darwin-2B-Opus** (이 모델) ⭐ 최경량
+
+---
+
+## 🪪 라이선스
+
+- Base model: Apache 2.0 (Qwen)
+- 학습 데이터: 각 데이터셋 개별 라이선스 참조
+- 이 모델: Apache 2.0
+
+---
+
+## 🙏 크레딧
+
+- **Base**: Qwen team (Alibaba)
+- **Teacher**: Anthropic (Claude Opus 4.5/4.6, Sonnet 4.6)
+- **데이터 공개**: nohurry, TeichAI, kuotient, PoSTMEDIA
+- **Training & Release**: **FINAL-Bench / VIDRAFT_LAB**
+
+---
+
+*Darwin V8 · Part of the evolutionary model series by FINAL-Bench*

chat_template.jinja

@@ -0,0 +1,154 @@
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- macro render_content(content, do_vision_count, is_system_content=false) %}
+    {%- if content is string %}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping %}
+        {%- for item in content %}
+            {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain images.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Picture ' ~ image_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|image_pad|><|vision_end|>' }}
+            {%- elif 'video' in item or item.type == 'video' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain videos.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Video ' ~ video_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|video_pad|><|vision_end|>' }}
+            {%- elif 'text' in item %}
+                {{- item.text }}
+            {%- else %}
+                {{- raise_exception('Unexpected item type in content.') }}
+            {%- endif %}
+        {%- endfor %}
+    {%- elif content is none or content is undefined %}
+        {{- '' }}
+    {%- else %}
+        {{- raise_exception('Unexpected content type.') }}
+    {%- endif %}
+{%- endmacro %}
+{%- if not messages %}
+    {{- raise_exception('No messages provided.') }}
+{%- endif %}
+{%- if tools and tools is iterable and tools is not mapping %}
+    {{- '<|im_start|>system\n' }}
+    {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>" }}
+    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {%- if content %}
+            {{- '\n\n' + content }}
+        {%- endif %}
+    {%- endif %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" %}
+        {%- set content = render_content(message.content, false)|trim %}
+        {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
+            {%- set ns.multi_step_tool = false %}
+            {%- set ns.last_query_index = index %}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if ns.multi_step_tool %}
+    {{- raise_exception('No user query found in messages.') }}
+{%- endif %}
+{%- for message in messages %}
+    {%- set content = render_content(message.content, true)|trim %}
+    {%- if message.role == "system" %}
+        {%- if not loop.first %}
+            {{- raise_exception('System message must be at the beginning.') }}
+        {%- endif %}
+    {%- elif message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- set reasoning_content = reasoning_content|trim %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if tool_call.function is defined %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {%- if loop.first %}
+                    {%- if content|trim %}
+                        {{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- else %}
+                        {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- endif %}
+                {%- else %}
+                    {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                {%- endif %}
+                {%- if tool_call.arguments is defined %}
+                    {%- for args_name, args_value in tool_call.arguments|items %}
+                        {{- '<parameter=' + args_name + '>\n' }}
+                        {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
+                        {{- args_value }}
+                        {{- '\n</parameter>\n' }}
+                    {%- endfor %}
+                {%- endif %}
+                {{- '</function>\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.previtem and loop.previtem.role != "tool" %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if not loop.last and loop.nextitem.role != "tool" %}
+            {{- '<|im_end|>\n' }}
+        {%- elif loop.last %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- else %}
+        {{- raise_exception('Unexpected message role.') }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is true %}
+        {{- '<think>\n' }}
+    {%- else %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}

config.json

@@ -0,0 +1,75 @@
+{
+  "architectures": [
+    "Qwen3_5ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "attn_output_gate": true,
+  "bos_token_id": null,
+  "dtype": "bfloat16",
+  "eos_token_id": 248044,
+  "full_attention_interval": 4,
+  "head_dim": 256,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "layer_types": [
+    "linear_attention",
+    "linear_attention",
+    "linear_attention",
+    "full_attention",
+    "linear_attention",
+    "linear_attention",
+    "linear_attention",
+    "full_attention",
+    "linear_attention",
+    "linear_attention",
+    "linear_attention",
+    "full_attention",
+    "linear_attention",
+    "linear_attention",
+    "linear_attention",
+    "full_attention",
+    "linear_attention",
+    "linear_attention",
+    "linear_attention",
+    "full_attention",
+    "linear_attention",
+    "linear_attention",
+    "linear_attention",
+    "full_attention"
+  ],
+  "linear_conv_kernel_dim": 4,
+  "linear_key_head_dim": 128,
+  "linear_num_key_heads": 16,
+  "linear_num_value_heads": 16,
+  "linear_value_head_dim": 128,
+  "mamba_ssm_dtype": "float32",
+  "max_position_embeddings": 262144,
+  "mlp_only_layers": [],
+  "model_type": "qwen3_5_text",
+  "mtp_num_hidden_layers": 1,
+  "mtp_use_dedicated_embeddings": false,
+  "num_attention_heads": 8,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 2,
+  "pad_token_id": null,
+  "partial_rotary_factor": 0.25,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "mrope_interleaved": true,
+    "mrope_section": [
+      11,
+      11,
+      10
+    ],
+    "partial_rotary_factor": 0.25,
+    "rope_theta": 10000000,
+    "rope_type": "default"
+  },
+  "tie_word_embeddings": true,
+  "transformers_version": "5.5.4",
+  "use_cache": true,
+  "vocab_size": 248320
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": 248044,
+  "transformers_version": "5.5.4",
+  "use_cache": true
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09213b86c44787505321403648e53655f07f323cd70fb5872bcad9ad5123341e
+size 3763692048

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4
+size 19989343

tokenizer_config.json

@@ -0,0 +1,31 @@
+{
+  "add_prefix_space": false,
+  "audio_bos_token": "<|audio_start|>",
+  "audio_eos_token": "<|audio_end|>",
+  "audio_token": "<|audio_pad|>",
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "image_token": "<|image_pad|>",
+  "is_local": false,
+  "model_max_length": 262144,
+  "model_specific_special_tokens": {
+    "audio_bos_token": "<|audio_start|>",
+    "audio_eos_token": "<|audio_end|>",
+    "audio_token": "<|audio_pad|>",
+    "image_token": "<|image_pad|>",
+    "video_token": "<|video_pad|>",
+    "vision_bos_token": "<|vision_start|>",
+    "vision_eos_token": "<|vision_end|>"
+  },
+  "pad_token": "<|endoftext|>",
+  "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
+  "split_special_tokens": false,
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": null,
+  "video_token": "<|video_pad|>",
+  "vision_bos_token": "<|vision_start|>",
+  "vision_eos_token": "<|vision_end|>"
+}