System Info

Copy-and-paste the text below in your GitHub issue and FILL OUT the two last points.

• transformers version: 4.57.0.dev0
• Platform: Linux-6.8.0-79-generic-x86_64-with-glibc2.39
• Python version: 3.12.11
• Huggingface_hub version: 0.35.0
• Safetensors version: 0.6.2
• Accelerate version: 1.10.1
• Accelerate config: not found
• DeepSpeed version: not installed
• PyTorch version (accelerator?): 2.8.0+cu128 (CUDA)
• Tensorflow version (GPU?): not installed (NA)
• Flax version (CPU?/GPU?/TPU?): not installed (NA)
• Jax version: not installed
• JaxLib version: not installed
• Using distributed or parallel set-up in script?:
• Using GPU in script?:
• GPU type: NVIDIA GeForce RTX 4090

Who can help?

@ArthurZucker @Cyrilvallez @gante

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

1、running with test.py given by Qwen official website.

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen/Qwen3-Next-80B-A3B-Instruct"
custom_cache_path = "/mnt/m2_4/models/Qwen3_Next_80B_A3B"
# load the tokenizer and the model from the local directory
tokenizer = AutoTokenizer.from_pretrained(model_name,cache_dir=custom_cache_path)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    dtype="auto",
    device_map="auto",
    cache_dir = custom_cache_path,
    
)

# prepare the model input
prompt = "Give me a short introduction to large language model."
messages = [
    {"role": "user", "content": prompt},
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=16384,
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 

content = tokenizer.decode(output_ids, skip_special_tokens=True)

print("content:", content)

2、the error log as follows, error when loading "mtp.***" tensor weights.

Expected behavior

output correct answers