Views
No views yet
| Parameter | Value |
|---|---|
| Method | AWQ (Activation-aware Weight Quantization) |
| Bits | 4 |
| Group Size | 128 |
| Original Size | ~8 GB |
| Quantized Size | ~2.5 GB |
| VRAM Required | ~3 GB |
1python -m vllm.entrypoints.openai.api_server \
2 --model tellang/yeji-4b-instruct-v9-AWQ \
3 --quantization awq \
4 --max-model-len 40961from transformers import AutoModelForCausalLM, AutoTokenizer
2
3model_name = "tellang/yeji-4b-instruct-v9-AWQ"
4tokenizer = AutoTokenizer.from_pretrained(model_name)
5model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")
6
7messages = [
8 {"role": "system", "content": "당신은 전문 타로 상담사입니다."},
9 {"role": "user", "content": "오늘의 원카드 타로 리딩을 해주세요."}
10]
11
12text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
13inputs = tokenizer(text, return_tensors="pt").to(model.device)
14outputs = model.generate(**inputs, max_new_tokens=512, temperature=0.7)
15print(tokenizer.decode(outputs[0], skip_special_tokens=True))| Format | Model | Use Case |
|---|---|---|
| Full (FP16) | yeji-4b-instruct-v9 | 학습, 고정밀 추론 |
| MLX 4-bit | yeji-4b-instruct-v9-mlx-4bit | Apple Silicon |
| GGUF | yeji-4b-gguf | llama.cpp, Ollama |