Text Generation
Transformers
Safetensors
GGUF
llama-cpp-python
MLX
Korean
English
qwen2
finance
korean
stock-analysis
reasoning
dpo
llama-cpp
apple-silicon
4bit
quantized
vllm
ollama
conversational
text-generation-inference
Instructions to use intrect/VELA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use intrect/VELA with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="intrect/VELA") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("intrect/VELA") model = AutoModelForCausalLM.from_pretrained("intrect/VELA") messages = [ {"role": "user", "content": "Who are you?"}, ] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=40) print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - llama-cpp-python
How to use intrect/VELA with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="intrect/VELA", filename="vela-dpo-v6-q4_k_m.gguf", )
llm.create_chat_completion( messages = [ { "role": "user", "content": "What is the capital of France?" } ] ) - MLX
How to use intrect/VELA with MLX:
# Make sure mlx-lm is installed # pip install --upgrade mlx-lm # Generate text with mlx-lm from mlx_lm import load, generate model, tokenizer = load("intrect/VELA") prompt = "Write a story about Einstein" messages = [{"role": "user", "content": prompt}] prompt = tokenizer.apply_chat_template( messages, add_generation_prompt=True ) text = generate(model, tokenizer, prompt=prompt, verbose=True) - Notebooks
- Google Colab
- Kaggle
- Local Apps
- llama.cpp
How to use intrect/VELA with llama.cpp:
Install from brew
brew install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf intrect/VELA:Q4_K_M # Run inference directly in the terminal: llama-cli -hf intrect/VELA:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf intrect/VELA:Q4_K_M # Run inference directly in the terminal: llama-cli -hf intrect/VELA:Q4_K_M
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf intrect/VELA:Q4_K_M # Run inference directly in the terminal: ./llama-cli -hf intrect/VELA:Q4_K_M
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf intrect/VELA:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf intrect/VELA:Q4_K_M
Use Docker
docker model run hf.co/intrect/VELA:Q4_K_M
- LM Studio
- Jan
- vLLM
How to use intrect/VELA with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "intrect/VELA" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "intrect/VELA", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/intrect/VELA:Q4_K_M
- SGLang
How to use intrect/VELA with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "intrect/VELA" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "intrect/VELA", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "intrect/VELA" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "intrect/VELA", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Ollama
How to use intrect/VELA with Ollama:
ollama run hf.co/intrect/VELA:Q4_K_M
- Unsloth Studio new
How to use intrect/VELA with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for intrect/VELA to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for intrect/VELA to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for intrect/VELA to start chatting
- Pi new
How to use intrect/VELA with Pi:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "intrect/VELA"
Configure the model in Pi
# Install Pi: npm install -g @mariozechner/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "mlx-lm": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "intrect/VELA" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent new
How to use intrect/VELA with Hermes Agent:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "intrect/VELA"
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default intrect/VELA
Run Hermes
hermes
- MLX LM
How to use intrect/VELA with MLX LM:
Generate or start a chat session
# Install MLX LM uv tool install mlx-lm # Interactive chat REPL mlx_lm.chat --model "intrect/VELA"
Run an OpenAI-compatible server
# Install MLX LM uv tool install mlx-lm # Start the server mlx_lm.server --model "intrect/VELA" # Calling the OpenAI-compatible server with curl curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "intrect/VELA", "messages": [ {"role": "user", "content": "Hello"} ] }' - Docker Model Runner
How to use intrect/VELA with Docker Model Runner:
docker model run hf.co/intrect/VELA:Q4_K_M
- Lemonade
How to use intrect/VELA with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull intrect/VELA:Q4_K_M
Run and chat with the model
lemonade run user.VELA-Q4_K_M
List all available models
lemonade list
docs: update README to v1.3 (DPO v6)
Browse files
README.md
CHANGED
|
@@ -23,23 +23,12 @@ pipeline_tag: text-generation
|
|
| 23 |
[](https://huggingface.co/intrect/VELA)
|
| 24 |
[](https://huggingface.co/spaces/intrect/vela-demo)
|
| 25 |
|
| 26 |
-
## Version History
|
| 27 |
-
|
| 28 |
-
| Version | Date | Changes |
|
| 29 |
-
|---------|------|---------|
|
| 30 |
-
| **v1.2** | **2026-02-16** | **SFT v3 (58K) Gap Fill 12카테고리, Markdown RT, 벤치마크 추가** |
|
| 31 |
-
| v1.1 | 2026-02-12 | GGUF 양자화 모델 추가 (Q4_K_M, Q8_0) |
|
| 32 |
-
| v1.0 | 2026-01-28 | DPO 병합, 중국어/영어 leak 해결 |
|
| 33 |
-
| v0.9 | 2026-01-15 | SFT 베이스 모델 공개 |
|
| 34 |
-
|
| 35 |
**한국 주식시장 전문 AI 애널리스트**
|
| 36 |
|
| 37 |
VELA는 한국 주식시장 뉴스 분석 및 투자 리서치를 위해 특화된 7B 파라미터 언어 모델입니다.
|
| 38 |
KOSPI/KOSDAQ 2,135개 종목에 대한 뉴스 영향 분석, 증권사 리포트 해석,
|
| 39 |
Reasoning Trace 기반 구조화된 투자 분석을 수행합니다.
|
| 40 |
|
| 41 |
-
지속적인 개발과 학습으로 개선해나가고 있습니다. 사용상의 이
|
| 42 |
-
|
| 43 |
> **58K+ SFT 샘플**과 **26K+ DPO 페어**로 학습하여,
|
| 44 |
> 한국어 금융 도메인에서 정확하고 구조화된 분석을 제공합니다.
|
| 45 |
>
|
|
@@ -63,68 +52,10 @@ Reasoning Trace 기반 구조화된 투자 분석을 수행합니다.
|
|
| 63 |
|
| 64 |
| Format | File | Size | Use Case |
|
| 65 |
|--------|------|------|----------|
|
| 66 |
-
| **BF16** (safetensors) | `model.safetensors` |
|
| 67 |
-
| **GGUF
|
| 68 |
-
| **GGUF Q4_K_M** | `vela-q4_k_m.gguf` | 4.4 GB | Fast & lightweight, GPU/CPU |
|
| 69 |
-
|
| 70 |
-
> MLX 4-bit 양자화 모델도 별도 제공 예정 (Apple Silicon 최적화)
|
| 71 |
-
|
| 72 |
-
### Recommended Inference Settings
|
| 73 |
-
|
| 74 |
-
VELA의 `generation_config.json`은 llama-cpp-python 서버와 동일한 샘플링 파라미터를 사용합니다.
|
| 75 |
-
모든 백엔드에서 동일한 출력 품질을 보장하기 위해 아래 설정을 권장합니다:
|
| 76 |
-
|
| 77 |
-
| Parameter | Value | Note |
|
| 78 |
-
|-----------|-------|------|
|
| 79 |
-
| `temperature` | 0.7 | 창의성-일관성 균형점 |
|
| 80 |
-
| `top_k` | 40 | 상위 40개 토큰 후보 (llama.cpp default) |
|
| 81 |
-
| `top_p` | 0.95 | 누적 확률 95% 이내 토큰 사용 |
|
| 82 |
-
| `repetition_penalty` | 1.0 | **비활성화** (후처리로 반복 제어) |
|
| 83 |
-
| `max_tokens` | 1024-2048 | 분석 복잡도에 따라 조절 |
|
| 84 |
-
|
| 85 |
-
> **주의사항**:
|
| 86 |
-
> - `repetition_penalty ≥ 1.2`는 **사용 금지** — Qwen 7B 기반 모델에서 중국어 text leak 및 환각(hallucination)을 유발합니다
|
| 87 |
-
> - `top_k < 20` 또는 `top_p < 0.8`은 출력 다양성을 과도하게 제한하여 confidence 고정(50%) 현상을 발생시킵니다
|
| 88 |
-
> - 반복 제어는 `repetition_penalty` 대신 **후처리 파이프라인**으로 처리하는 것이 안정적입니다
|
| 89 |
-
|
| 90 |
-
#### Backend별 설정 가이드
|
| 91 |
-
|
| 92 |
-
**llama-cpp-python / Ollama** — 별도 설정 불필요 (기본값이 권장값과 동일):
|
| 93 |
-
```python
|
| 94 |
-
model.create_chat_completion(
|
| 95 |
-
messages=messages,
|
| 96 |
-
max_tokens=1024,
|
| 97 |
-
temperature=0.7,
|
| 98 |
-
# top_k, top_p, repeat_penalty는 서버 기본값 사용
|
| 99 |
-
)
|
| 100 |
-
```
|
| 101 |
-
|
| 102 |
-
**HuggingFace Transformers** — `generation_config.json`이 자동 로드됨. 명시적 파라미터 최소화:
|
| 103 |
-
```python
|
| 104 |
-
outputs = model.generate(
|
| 105 |
-
**inputs,
|
| 106 |
-
max_new_tokens=1024,
|
| 107 |
-
temperature=0.7,
|
| 108 |
-
do_sample=True,
|
| 109 |
-
# top_k, top_p, repetition_penalty는 generation_config.json에서 로드
|
| 110 |
-
)
|
| 111 |
-
```
|
| 112 |
-
|
| 113 |
-
**vLLM** — 명시적 설정 권장:
|
| 114 |
-
```python
|
| 115 |
-
params = SamplingParams(
|
| 116 |
-
temperature=0.7,
|
| 117 |
-
top_k=40,
|
| 118 |
-
top_p=0.95,
|
| 119 |
-
repetition_penalty=1.0,
|
| 120 |
-
max_tokens=1024,
|
| 121 |
-
)
|
| 122 |
-
```
|
| 123 |
|
| 124 |
-
|
| 125 |
-
```python
|
| 126 |
-
generate(model, tokenizer, prompt=prompt, max_tokens=1024, temp=0.7)
|
| 127 |
-
```
|
| 128 |
|
| 129 |
---
|
| 130 |
|
|
@@ -197,7 +128,7 @@ Qwen/Qwen2.5-7B-Instruct
|
|
| 197 |
└── Reasoning Trace 정렬 566 (2.1%)
|
| 198 |
|
|
| 199 |
v
|
| 200 |
-
VELA v1.
|
| 201 |
```
|
| 202 |
|
| 203 |
---
|
|
@@ -264,8 +195,7 @@ RTX 3060 12GB, llama-cpp-python, `n_gpu_layers=-1`, `n_ctx=4096`
|
|
| 264 |
|
| 265 |
| Format | Speed | Chinese Leak | Quality |
|
| 266 |
|--------|-------|--------------|---------|
|
| 267 |
-
| **Q4_K_M** | **36 tok/s** | 0/5 CLEAN | RT + Report OK |
|
| 268 |
-
| **Q8_0** | 25 tok/s | 0/5 CLEAN | RT + Report OK |
|
| 269 |
|
| 270 |
> Stress test 5회: Synthesis + 3K Reasoning Trace 교대 — 양쪽 모두 **Chinese leak 제로**
|
| 271 |
|
|
@@ -303,7 +233,7 @@ MLX 4-bit vs PyTorch CPU:
|
|
| 303 |
from llama_cpp import Llama
|
| 304 |
|
| 305 |
model = Llama(
|
| 306 |
-
model_path="vela-q4_k_m.gguf",
|
| 307 |
n_ctx=4096,
|
| 308 |
n_gpu_layers=-1,
|
| 309 |
chat_format="chatml",
|
|
@@ -598,4 +528,16 @@ VELA는 단독 LLM으로도 동작하지만, 에이전트 시스템과 결합하
|
|
| 598 |
|
| 599 |
---
|
| 600 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 601 |
**Disclaimer**: 이 모델의 출력은 투자 조언이 아닙니다. 모든 투자 결정은 본인의 판단과 책임 하에 이루어져야 합니다.
|
|
|
|
| 23 |
[](https://huggingface.co/intrect/VELA)
|
| 24 |
[](https://huggingface.co/spaces/intrect/vela-demo)
|
| 25 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
**한국 주식시장 전문 AI 애널리스트**
|
| 27 |
|
| 28 |
VELA는 한국 주식시장 뉴스 분석 및 투자 리서치를 위해 특화된 7B 파라미터 언어 모델입니다.
|
| 29 |
KOSPI/KOSDAQ 2,135개 종목에 대한 뉴스 영향 분석, 증권사 리포트 해석,
|
| 30 |
Reasoning Trace 기반 구조화된 투자 분석을 수행합니다.
|
| 31 |
|
|
|
|
|
|
|
| 32 |
> **58K+ SFT 샘플**과 **26K+ DPO 페어**로 학습하여,
|
| 33 |
> 한국어 금융 도메인에서 정확하고 구조화된 분석을 제공합니다.
|
| 34 |
>
|
|
|
|
| 52 |
|
| 53 |
| Format | File | Size | Use Case |
|
| 54 |
|--------|------|------|----------|
|
| 55 |
+
| **BF16** (safetensors) | `model-*.safetensors` | 14.5 GB | Full precision, GPU inference |
|
| 56 |
+
| **GGUF Q4_K_M** | `vela-dpo-v6-q4_k_m.gguf` | 4.4 GB | Fast & lightweight, GPU/CPU |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 57 |
|
| 58 |
+
> MLX 4-bit 양자화 모델은 별도 레포에서 제공 예정 (Apple Silicon 최적화)
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
---
|
| 61 |
|
|
|
|
| 128 |
└── Reasoning Trace 정렬 566 (2.1%)
|
| 129 |
|
|
| 130 |
v
|
| 131 |
+
VELA v1.3
|
| 132 |
```
|
| 133 |
|
| 134 |
---
|
|
|
|
| 195 |
|
| 196 |
| Format | Speed | Chinese Leak | Quality |
|
| 197 |
|--------|-------|--------------|---------|
|
| 198 |
+
| **Q4_K_M (v6)** | **36 tok/s** | 0/5 CLEAN | RT + Report OK |
|
|
|
|
| 199 |
|
| 200 |
> Stress test 5회: Synthesis + 3K Reasoning Trace 교대 — 양쪽 모두 **Chinese leak 제로**
|
| 201 |
|
|
|
|
| 233 |
from llama_cpp import Llama
|
| 234 |
|
| 235 |
model = Llama(
|
| 236 |
+
model_path="vela-dpo-v6-q4_k_m.gguf",
|
| 237 |
n_ctx=4096,
|
| 238 |
n_gpu_layers=-1,
|
| 239 |
chat_format="chatml",
|
|
|
|
| 528 |
|
| 529 |
---
|
| 530 |
|
| 531 |
+
## Version History
|
| 532 |
+
|
| 533 |
+
| Version | Date | Changes |
|
| 534 |
+
|---------|------|---------|
|
| 535 |
+
| **v1.3** | **2026-03-31** | **DPO v6 모델 업데이트 (Unsloth 학습, SFT v6 + DPO v6 merged)** |
|
| 536 |
+
| v1.2 | 2026-02-16 | SFT v3 (58K) Gap Fill 12카테고리, Markdown RT, 벤치마크 추가 |
|
| 537 |
+
| v1.1 | 2026-02-12 | GGUF 양자화 모델 추가 (Q4_K_M, Q8_0) |
|
| 538 |
+
| v1.0 | 2026-01-28 | DPO 병합, 중국어/영어 leak 해결 |
|
| 539 |
+
| v0.9 | 2026-01-15 | SFT 베이스 모델 공개 |
|
| 540 |
+
|
| 541 |
+
---
|
| 542 |
+
|
| 543 |
**Disclaimer**: 이 모델의 출력은 투자 조언이 아닙니다. 모든 투자 결정은 본인의 판단과 책임 하에 이루어져야 합니다.
|