serialize_module_with_resources not found

[HemKava]

When generating Torch MLIR, is there a way to write weights and model in separate file? My model weights are huge for Llama3 layers, and this is leading to giant mlir files.

I did some digging and it feels like there a way using serialize_module_with_resources based on google search, but I can't find it in the source code.
Is it something deprecated?

[vfdff]

I’m also interested in getting a tool to do such issue, and the deepseek give a sample

import onnx

# 加载大型 ONNX 文件
model = onnx.load("llama3_huge.onnx")

# 转换为外部数据格式
onnx.save_model(
    model,
    "llama3_external.onnx",
    save_as_external_data=True,
    all_tensors_to_one_file=True,  # 所有张量保存到一个文件
    location="llama3_weights.bin",  # 权重文件名
    size_threshold=1024,  # 超过 1KB 的张量就外部化
    convert_attribute=False
)

(py311-source) root@998ee80b761b:~/test/model# pip list | grep onnx
onnx 1.17.0