README.md

---
base_model: Qwen/Qwen3-4B-Instruct-2507
library_name: peft
pipeline_tag: text-generation
license: apache-2.0
tags:
  - lora
  - metabolomics
---

# MetaboLLM-Qwen3-4B

PEFT LoRA adapter for metabolomics and biochemical knowledge tasks, trained from
`Qwen/Qwen3-4B-Instruct-2507`. The base-model weights are not included in this repository.

Requires `transformers>=4.51`, since the chat template ships as a standalone
`chat_template.jinja` file.

```bash
pip install -U "transformers>=4.51" peft accelerate torch
```

## Usage

```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

BASE = "Qwen/Qwen3-4B-Instruct-2507"
ADAPTER = "MetaboLLM/MetaboLLM-Qwen3-4B"

tokenizer = AutoTokenizer.from_pretrained(ADAPTER)
model = AutoModelForCausalLM.from_pretrained(
    BASE,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)
model = PeftModel.from_pretrained(model, ADAPTER)
model.eval()

messages = [
    {"role": "system", "content": "You are a metabolomics expert."},
    {"role": "user", "content": "What is the biological role of L-Alanine?"},
]
text = tokenizer.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
inputs = tokenizer(text, return_tensors="pt").to(model.device)

outputs = model.generate(**inputs, max_new_tokens=512, do_sample=False)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:],
                       skip_special_tokens=True))
```

Call `model.merge_and_unload()` after loading to fold the adapter into the base
weights for faster repeated inference.