metadata
language:
- en
library_name: mlx
tags:
- glm
- MOE
- pruning
- compression
- mlx
license: mit
name: cerebras/Kimi-Linear-REAP-35B-A3B-Instruct
description: >
This model was obtained by uniformly pruning 30% of experts in
Kimi-Linear-48B-A3B-Instruct using the REAP method.
readme: >
https://huggingface.co/cerebras/Kimi-Linear-REAP-35B-A3B-Instruct/main/README.md
pipeline_tag: text-generation
base_model: cerebras/Kimi-Linear-REAP-35B-A3B-Instruct
NexVeridian/Kimi-Linear-REAP-35B-A3B-Instruct-4bit
This model NexVeridian/Kimi-Linear-REAP-35B-A3B-Instruct-4bit was converted to MLX format from cerebras/Kimi-Linear-REAP-35B-A3B-Instruct using mlx-lm version 0.28.4.
Use with mlx
pip install mlx-lm
from mlx_lm import load, generate
model, tokenizer = load("NexVeridian/Kimi-Linear-REAP-35B-A3B-Instruct-4bit")
prompt = "hello"
if tokenizer.chat_template is not None:
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
messages, add_generation_prompt=True
)
response = generate(model, tokenizer, prompt=prompt, verbose=True)