---
library_name: peft
base_model: meta-llama/Llama-3.1-8B-Instruct
license: cc-by-nc-4.0
pipeline_tag: text-generation
tags:
  - lora
  - peft
  - molly-os
  - software-engineering
---

# Molly OS - Specialist Adapter: Software Engineering

Frontier-distilled **LoRA specialist** (PEFT, rank 32; target modules
`q_proj`, `k_proj`, `v_proj`, `o_proj`) for the Molly OS model-agnostic
orchestration layer. Base model: **[meta-llama/Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct)**.
Domain: **Software Engineering**.

Adapter weights are released under **CC BY-NC 4.0**. The base model is governed by
its own (Llama 3.1) license.

## Before you run: the base model is gated

This adapter needs the base weights, and the base is **access-gated**. Do this **once**:

1. Open the base page and **accept its license**: <https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct>
2. Create a **read token**: <https://huggingface.co/settings/tokens>
3. Make the token available to your environment:
   - **Google Colab:** open the **Secrets** panel (key icon, left sidebar) -> *Add new secret* -> Name `HF_TOKEN`, paste the value, enable **Notebook access**.
   - **Kaggle:** *Add-ons -> Secrets* -> add `HF_TOKEN`.
   - **Local:** run `huggingface-cli login` or `export HF_TOKEN=...`.

If you skip this you will get `GatedRepoError` / `401 Unauthorized` when the **base** loads.
A stored Colab secret is **not** used automatically - you must authenticate in code (see below).

## Quickstart

```python
# pip install -U transformers peft accelerate
import os
from huggingface_hub import login

# Authenticate (Colab secret -> env var -> interactive prompt)
try:
    from google.colab import userdata
    login(userdata.get("HF_TOKEN"))
except Exception:
    tok = os.environ.get("HF_TOKEN")
    login(tok) if tok else login()

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

BASE = "meta-llama/Llama-3.1-8B-Instruct"
ADAPTER = "BoomJules/molly-software-engineering"

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

msgs = [{"role": "user", "content": "Your question here"}]
ids = tok.apply_chat_template(msgs, add_generation_prompt=True, return_tensors="pt").to(model.device)
out = model.generate(ids, max_new_tokens=300)
print(tok.decode(out[0][ids.shape[1]:], skip_special_tokens=True))
```

## Low-VRAM (4-bit) - fits a free Colab/Kaggle GPU (~6-7 GB)

Use a **GPU runtime** (Colab: *Runtime -> Change runtime type -> T4 GPU*).

```python
# pip install -U transformers peft accelerate bitsandbytes
import os, torch
from huggingface_hub import login
try:
    from google.colab import userdata
    login(userdata.get("HF_TOKEN"))
except Exception:
    login(os.environ.get("HF_TOKEN"))

from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
from peft import PeftModel

BASE = "meta-llama/Llama-3.1-8B-Instruct"
ADAPTER = "BoomJules/molly-software-engineering"

bnb = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_quant_type="nf4",
                         bnb_4bit_compute_dtype=torch.bfloat16, bnb_4bit_use_double_quant=True)
tok = AutoTokenizer.from_pretrained(BASE)
base = AutoModelForCausalLM.from_pretrained(BASE, quantization_config=bnb, device_map="auto")
model = PeftModel.from_pretrained(base, ADAPTER).eval()
```

## Troubleshooting

- **`GatedRepoError` / `401 Unauthorized`** - base license not accepted, or `HF_TOKEN`
  missing/invalid, or you stored the Colab secret but did not call `login(...)` in code.
- **CUDA out of memory** - use the 4-bit snippet and a GPU runtime.
- **Adapter seems to have no effect** - confirm the base id matches `base_model` above.

## License & intended use

Adapter: **CC BY-NC 4.0** (attribution, non-commercial). Base model: Llama 3.1 license.
Intended for research and evaluation in Software Engineering.

(c) 2026 Core Labs R&D.