mlx-community/Dolphin-Mistral-24B-Venice-Edition-mlx-8Bit

MLX 8bit quant of cognitivecomputations/Dolphin-Mistral-24B-Venice-Edition released 2025-06-12.

"This is an updated version based on feedback we received on v1", see discussion at original repo.

The Model mlx-community/Dolphin-Mistral-24B-Venice-Edition-mlx-8Bit was converted to MLX format from cognitivecomputations/Dolphin-Mistral-24B-Venice-Edition using mlx-lm version 0.22.3.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Dolphin-Mistral-24B-Venice-Edition-mlx-8Bit")

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)