fine-tune RuntimeError: expected dtype Float but got dtype Long

When fine-tune on my own data, error happen when run model.fit

Traceback (most recent call last):
  File "fine_tune_sbert.py", line 109, in <module>
    output_path=model_save_path)
  File "/opt/conda/lib/python3.7/site-packages/sentence_transformers/SentenceTransformer.py", line 402, in fit
    loss_value.backward()
  File "/opt/conda/lib/python3.7/site-packages/torch/tensor.py", line 198, in backward
    torch.autograd.backward(self, gradient, retain_graph, create_graph)
  File "/opt/conda/lib/python3.7/site-packages/torch/autograd/__init__.py", line 100, in backward
    allow_unreachable=True)  # allow_unreachable flag
RuntimeError: expected dtype Float but got dtype Long (validate_dtype at /pytorch/aten/src/ATen/native/TensorIterator.cpp:143)
frame #0: c10::Error::Error(c10::SourceLocation, std::string const&) + 0x46 (0x7f4a9deda536 in /opt/conda/lib/python3.7/site-packages/torch/lib/libc10.so)
frame #1: at::TensorIterator::compute_types() + 0xce3 (0x7f4adb981a23 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #2: at::TensorIterator::build() + 0x44 (0x7f4adb984404 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #3: at::native::mse_loss_backward_out(at::Tensor&, at::Tensor const&, at::Tensor const&, at::Tensor const&, long) + 0x193 (0x7f4adb7d1953 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #4: <unknown function> + 0xf903d7 (0x7f4a9f2c63d7 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #5: at::native::mse_loss_backward(at::Tensor const&, at::Tensor const&, at::Tensor const&, long) + 0x172 (0x7f4adb7da092 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #6: <unknown function> + 0xf9068f (0x7f4a9f2c668f in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #7: <unknown function> + 0x10c2536 (0x7f4adbc0a536 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #8: <unknown function> + 0x2a9ecdb (0x7f4add5e6cdb in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #9: <unknown function> + 0x10c2536 (0x7f4adbc0a536 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #10: torch::autograd::generated::MseLossBackward::apply(std::vector<at::Tensor, std::allocator<at::Tensor> >&&) + 0x1f7 (0x7f4add3ee777 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #11: <unknown function> + 0x2d89705 (0x7f4add8d1705 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #12: torch::autograd::Engine::evaluate_function(std::shared_ptr<torch::autograd::GraphTask>&, torch::autograd::Node*, torch::autograd::InputBuffer&) + 0x16f3 (0x7f4add8cea03 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #13: torch::autograd::Engine::thread_main(std::shared_ptr<torch::autograd::GraphTask> const&, bool) + 0x3d2 (0x7f4add8cf7e2 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #14: torch::autograd::Engine::thread_init(int) + 0x39 (0x7f4add8c7e59 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #15: torch::autograd::python::PythonEngine::thread_init(int) + 0x38 (0x7f4aea20fac8 in /opt/conda/lib/python3.7/site-packages/torch/lib/libtorch_python.so)
frame #16: <unknown function> + 0xbd66f (0x7f4aeb0bf66f in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
frame #17: <unknown function> + 0x76db (0x7f4aece476db in /lib/x86_64-linux-gnu/libpthread.so.0)
frame #18: clone + 0x3f (0x7f4aecb7088f in /lib/x86_64-linux-gnu/libc.so.6)

My torch version is 1.15

[nreimers]

@naty88 Appears your labels are ints. But you need to pass floats

[chintanckg]

You are doing supervised training without labels, this will not work. Please try un-supervised training (well documented on the sentence-transformers site).

Issue can be marked closed.