Instructions to use huangang/Arch-Agent-3B-mlx-4Bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use huangang/Arch-Agent-3B-mlx-4Bit with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="huangang/Arch-Agent-3B-mlx-4Bit")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("huangang/Arch-Agent-3B-mlx-4Bit")
model = AutoModelForCausalLM.from_pretrained("huangang/Arch-Agent-3B-mlx-4Bit", device_map="auto")

MLX

How to use huangang/Arch-Agent-3B-mlx-4Bit with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm
# if on a CUDA device, also pip install mlx[cuda]

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("huangang/Arch-Agent-3B-mlx-4Bit")

prompt = "Once upon a time in"
text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

vLLM

How to use huangang/Arch-Agent-3B-mlx-4Bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "huangang/Arch-Agent-3B-mlx-4Bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "huangang/Arch-Agent-3B-mlx-4Bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/huangang/Arch-Agent-3B-mlx-4Bit

SGLang

How to use huangang/Arch-Agent-3B-mlx-4Bit 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 "huangang/Arch-Agent-3B-mlx-4Bit" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "huangang/Arch-Agent-3B-mlx-4Bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "huangang/Arch-Agent-3B-mlx-4Bit" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "huangang/Arch-Agent-3B-mlx-4Bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

MLX LM

How to use huangang/Arch-Agent-3B-mlx-4Bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Generate some text
mlx_lm.generate --model "huangang/Arch-Agent-3B-mlx-4Bit" --prompt "Once upon a time"

Docker Model Runner
How to use huangang/Arch-Agent-3B-mlx-4Bit with Docker Model Runner:
```
docker model run hf.co/huangang/Arch-Agent-3B-mlx-4Bit
```
Atomic Chat

Arch-Agent-3B-mlx-4Bit

File size: 1,019 Bytes

55620bc

---
license: other
license_name: katanemo-research
license_link: https://huggingface.co/katanemo/Arch-Agent-3B/blob/main/LICENSE
base_model: katanemo/Arch-Agent-3B
language:
- en
pipeline_tag: text-generation
library_name: transformers
tags:
- mlx
---

# huangang/Arch-Agent-3B-mlx-4Bit

The Model [huangang/Arch-Agent-3B-mlx-4Bit](https://huggingface.co/huangang/Arch-Agent-3B-mlx-4Bit) was converted to MLX format from [katanemo/Arch-Agent-3B](https://huggingface.co/katanemo/Arch-Agent-3B) using mlx-lm version **0.22.3**.

## Use with mlx

```bash
pip install mlx-lm
```

```python
from mlx_lm import load, generate

model, tokenizer = load("huangang/Arch-Agent-3B-mlx-4Bit")

prompt="hello"

if hasattr(tokenizer, "apply_chat_template") and tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, tokenize=False, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)
```