Try to load pytorch model from python training save for javacpp meet error

[mullerhai]

Hi, @saudet
when I use javacpp-pytorch traning pytorch model and save as net.pt , I could reload it with javacpp-pytorch too. I use zip viewer saw the pt is really zip dir file and have constants.pkl and some other file.
but when I try to reload pytorch model from python ,it meet error , yeah , I found these model pt file not have constants.pkl ,but have data.pkl file .

  val modelPath = "D:\\data\\git\\storch-transformers\\src\\main\\resources\\llama4_moe_model.pt" //lstm-net.pt"//pytorch_model.bin"
  // 初始化 InputArchive
  val inputArchive = new InputArchive()
  try {
    // 加载模型文件
    val nativeModule = torch.load( modelPath)
    nativeModule.named_modules()
    inputArchive.load_from(modelPath)
//    val nativeModule = new Module()
//    nativeModule.load(inputArchive)

console
An error occurred while loading the model: PytorchStreamReader failed locating file constants.pkl: file not found
Exception raised from valid at D:\a\javacpp-presets\javacpp-presets\pytorch\cppbuild\windows-x86_64\pytorch\caffe2\serialize\inline_container.cc:242 (most recent call first):
00007FF991F983C9 <unknown symbol address> c10.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF991F96C5A <unknown symbol address> c10.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B79BD663 <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B79BAA2D <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B79B9FFD <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B8B1062A <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B8B0E276 <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B8B0AEAE <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B8B0C508 <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B8B0C74F <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF8B8B0D68F <unknown symbol address> torch_cpu.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
00007FF92262C063 <unknown symbol address> jnitorch.dll!<unknown symbol> [<unknown file> @ <unknown line number>]
000002164BB0D207 <unknown symbol address> !<unknown symbol> [<unknown file> @ <unknown line number>]

so how to deal with python pytorch model load?