RichardErkhov/TheBlueObserver_-_SmolLM2-135M-Instruct-MLX-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

SmolLM2-135M-Instruct-MLX - GGUF

• Model creator: https://huggingface.co/TheBlueObserver/
• Original model: https://huggingface.co/TheBlueObserver/SmolLM2-135M-Instruct-MLX/

Name	Quant method	Size
SmolLM2-135M-Instruct-MLX.Q2_K.gguf	Q2_K	0.08GB
SmolLM2-135M-Instruct-MLX.IQ3_XS.gguf	IQ3_XS	0.08GB
SmolLM2-135M-Instruct-MLX.IQ3_S.gguf	IQ3_S	0.08GB
SmolLM2-135M-Instruct-MLX.Q3_K_S.gguf	Q3_K_S	0.08GB
SmolLM2-135M-Instruct-MLX.IQ3_M.gguf	IQ3_M	0.08GB
SmolLM2-135M-Instruct-MLX.Q3_K.gguf	Q3_K	0.09GB
SmolLM2-135M-Instruct-MLX.Q3_K_M.gguf	Q3_K_M	0.09GB
SmolLM2-135M-Instruct-MLX.Q3_K_L.gguf	Q3_K_L	0.09GB
SmolLM2-135M-Instruct-MLX.IQ4_XS.gguf	IQ4_XS	0.09GB
SmolLM2-135M-Instruct-MLX.Q4_0.gguf	Q4_0	0.09GB
SmolLM2-135M-Instruct-MLX.IQ4_NL.gguf	IQ4_NL	0.09GB
SmolLM2-135M-Instruct-MLX.Q4_K_S.gguf	Q4_K_S	0.1GB
SmolLM2-135M-Instruct-MLX.Q4_K.gguf	Q4_K	0.1GB
SmolLM2-135M-Instruct-MLX.Q4_K_M.gguf	Q4_K_M	0.1GB
SmolLM2-135M-Instruct-MLX.Q4_1.gguf	Q4_1	0.09GB
SmolLM2-135M-Instruct-MLX.Q5_0.gguf	Q5_0	0.1GB
SmolLM2-135M-Instruct-MLX.Q5_K_S.gguf	Q5_K_S	0.1GB
SmolLM2-135M-Instruct-MLX.Q5_K.gguf	Q5_K	0.1GB
SmolLM2-135M-Instruct-MLX.Q5_K_M.gguf	Q5_K_M	0.1GB
SmolLM2-135M-Instruct-MLX.Q5_1.gguf	Q5_1	0.1GB
SmolLM2-135M-Instruct-MLX.Q6_K.gguf	Q6_K	0.13GB
SmolLM2-135M-Instruct-MLX.Q8_0.gguf	Q8_0	0.13GB

Original model description:

base_model: HuggingFaceTB/SmolLM2-135M-Instruct language: - en library_name: transformers license: apache-2.0 pipeline_tag: text-generation tags: - safetensors - onnx - transformers.js - mlx

TheBlueObserver/SmolLM2-135M-Instruct-MLX

The Model TheBlueObserver/SmolLM2-135M-Instruct-MLX was converted to MLX format from HuggingFaceTB/SmolLM2-135M-Instruct using mlx-lm version 0.20.2.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("TheBlueObserver/SmolLM2-135M-Instruct-MLX")

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)