Xonaz81/XortronCriminalComputingConfig-mlx-6Bit

Because this model seems to be promising and there was no 6-bit version to be found, I decided to create one from the full model weights. This is a normal 6-bit MLX quant. No advanced DWQ quants for now but coming in the future! The original model Xonaz81/XortronCriminalComputingConfig-mlx-6Bit was converted to MLX format from darkc0de/XortronCriminalComputingConfig using mlx-lm

Use with mlx or LM-studio

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("Xonaz81/XortronCriminalComputingConfig-mlx-6Bit")

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)