The Model pcuenq/Qwen3-235B-A22B-Instruct-2507-mlx-4Bit was converted to MLX format from Qwen/Qwen3-235B-A22B-Instruct-2507 using mlx-lm version 0.26.3.
pip install mlx-lm
from mlx_lm import load, generate
model, tokenizer = load("pcuenq/Qwen3-235B-A22B-Instruct-2507-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)
model-00025-of-00026.safetensors 5.1 GBmodel-00013-of-00026.safetensors 5.1 GBmodel-00004-of-00026.safetensors 5.1 GBmodel-00022-of-00026.safetensors 5.1 GBmodel-00010-of-00026.safetensors 5.1 GBmodel-00008-of-00026.safetensors 5.1 GBmodel-00019-of-00026.safetensors 5.1 GBmodel-00016-of-00026.safetensors 5.1 GBmodel-00014-of-00026.safetensors 5.1 GBmodel-00007-of-00026.safetensors 5.1 GBmodel-00023-of-00026.safetensors 5.1 GBmodel-00017-of-00026.safetensors 5.1 GBmodel-00020-of-00026.safetensors 5.1 GBmodel-00005-of-00026.safetensors 5.1 GBmodel-00011-of-00026.safetensors 5.1 GBmodel-00002-of-00026.safetensors 5.1 GBmodel-00012-of-00026.safetensors 5.1 GBmodel-00024-of-00026.safetensors 5.1 GBmodel-00009-of-00026.safetensors 5.1 GBmodel-00015-of-00026.safetensors 5.1 GBmodel-00021-of-00026.safetensors 5.1 GBmodel-00006-of-00026.safetensors 5.1 GBmodel-00003-of-00026.safetensors 5.1 GBmodel-00018-of-00026.safetensors 5.1 GBmodel-00001-of-00026.safetensors 5.0 GBmodel-00026-of-00026.safetensors 4.1 GB有文件?提交种子。
提交种子