Unsupported type byte size: ComplexFloat

[cczw2010]

System : Apple (mac pro m2)
Use 12 cpu cores for computing 。

when i use Infer from prompt show this error

To create a public link, set `share=True` in `launch()`.
2023-10-10 17:09:35,552 INFO [launch-ui.py:335] synthesize text: [ZH]请注意，默认情况下,只处理语法转译，且 不包含任何 polyfill[ZH]
Building prefix dict from the default dictionary ...
2023-10-10 17:09:35,557 DEBUG [__init__.py:113] Building prefix dict from the default dictionary ...
Loading model from cache /var/folders/br/tr56lxdj3q3d8d1lkqp_xy680000gn/T/jieba.cache
2023-10-10 17:09:35,557 DEBUG [__init__.py:132] Loading model from cache /var/folders/br/tr56lxdj3q3d8d1lkqp_xy680000gn/T/jieba.cache
Loading model cost 0.351 seconds.
2023-10-10 17:09:35,908 DEBUG [__init__.py:164] Loading model cost 0.351 seconds.
Prefix dict has been built successfully.
2023-10-10 17:09:35,908 DEBUG [__init__.py:166] Prefix dict has been built successfully.


VALL-E EOS [493 -> 585]
libc++abi: terminating due to uncaught exception of type c10::Error: Unsupported type byte size: ComplexFloat
Exception raised from getGatherScatterScalarType at /Users/runner/work/pytorch/pytorch/pytorch/aten/src/ATen/native/mps/operations/View.mm:758 (most recent call first):
frame #0: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>> const&) + 92 (0x13755d2b8 in libc10.dylib)
frame #1: at::native::mps::getGatherScatterScalarType(at::Tensor const&) + 304 (0x2be92313c in libtorch_cpu.dylib)
frame #2: invocation function for block in at::native::mps::gatherViewTensor(at::Tensor const&, at::Tensor&) + 128 (0x2be924c8c in libtorch_cpu.dylib)
frame #3: _dispatch_client_callout + 20 (0x19cdf0400 in libdispatch.dylib)
frame #4: _dispatch_lane_barrier_sync_invoke_and_complete + 56 (0x19cdff97c in libdispatch.dylib)
frame #5: at::native::mps::gatherViewTensor(at::Tensor const&, at::Tensor&) + 888 (0x2be923824 in libtorch_cpu.dylib)
frame #6: at::native::mps::mps_copy_(at::Tensor&, at::Tensor const&, bool) + 3096 (0x2be87ab44 in libtorch_cpu.dylib)
frame #7: at::native::copy_impl(at::Tensor&, at::Tensor const&, bool) + 1944 (0x2ba5f75e4 in libtorch_cpu.dylib)
frame #8: at::native::copy_(at::Tensor&, at::Tensor const&, bool) + 100 (0x2ba5f6d8c in libtorch_cpu.dylib)
frame #9: at::_ops::copy_::call(at::Tensor&, at::Tensor const&, bool) + 288 (0x2bb32d6f8 in libtorch_cpu.dylib)
frame #10: at::native::clone(at::Tensor const&, c10::optional<c10::MemoryFormat>) + 444 (0x2ba981f64 in libtorch_cpu.dylib)
frame #11: at::_ops::clone::call(at::Tensor const&, c10::optional<c10::MemoryFormat>) + 276 (0x2bb03b0a4 in libtorch_cpu.dylib)
frame #12: at::_ops::contiguous::call(at::Tensor const&, c10::MemoryFormat) + 272 (0x2bb45fa40 in libtorch_cpu.dylib)
frame #13: at::TensorBase::__dispatch_contiguous(c10::MemoryFormat) const + 40 (0x2ba447110 in libtorch_cpu.dylib)
frame #14: at::native::mps::binaryOpTensor(at::Tensor const&, at::Tensor const&, c10::Scalar const&, at::Tensor const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>>, MPSGraphTensor* (at::native::mps::BinaryOpCachedGraph*, MPSGraphTensor*, MPSGraphTensor*) block_pointer) + 968 (0x2be86331c in libtorch_cpu.dylib)
frame #15: at::native::structured_mul_out_mps::impl(at::Tensor const&, at::Tensor const&, at::Tensor const&) + 128 (0x2be8673dc in libtorch_cpu.dylib)
frame #16: at::(anonymous namespace)::wrapper_MPS_mul_Tensor(at::Tensor const&, at::Tensor const&) + 140 (0x2bc003e88 in libtorch_cpu.dylib)
frame #17: at::_ops::mul_Tensor::call(at::Tensor const&, at::Tensor const&) + 284 (0x2bae41878 in libtorch_cpu.dylib)
frame #18: torch::autograd::THPVariable_mul(_object*, _object*, _object*) + 396 (0x16e69c9e8 in libtorch_python.dylib)
frame #19: _object* torch::autograd::TypeError_to_NotImplemented_<&torch::autograd::THPVariable_mul(_object*, _object*, _object*)>(_object*, _object*, _object*) + 12 (0x16e5f8a3c in libtorch_python.dylib)
frame #20: method_vectorcall_VARARGS_KEYWORDS + 488 (0x1025cbb3c in python3.10)
frame #21: vectorcall_maybe + 260 (0x102639650 in python3.10)
frame #22: slot_nb_multiply + 160 (0x102635d74 in python3.10)
frame #23: binary_op1 + 316 (0x1025a05a0 in python3.10)
frame #24: PyNumber_Multiply + 36 (0x1025a0b24 in python3.10)
frame #25: _PyEval_EvalFrameDefault + 3920 (0x1026a3318 in python3.10)
frame #26: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #27: method_vectorcall + 516 (0x1025c15f8 in python3.10)
frame #28: _PyEval_EvalFrameDefault + 27276 (0x1026a8e54 in python3.10)
frame #29: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #30: _PyObject_FastCallDictTstate + 320 (0x1025be094 in python3.10)
frame #31: _PyObject_Call_Prepend + 164 (0x1025bec8c in python3.10)
frame #32: slot_tp_call + 116 (0x102633ce8 in python3.10)
frame #33: _PyObject_MakeTpCall + 612 (0x1025bdde0 in python3.10)
frame #34: call_function + 676 (0x1026acde4 in python3.10)
frame #35: _PyEval_EvalFrameDefault + 26388 (0x1026a8adc in python3.10)
frame #36: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #37: PyVectorcall_Call + 156 (0x1025be624 in python3.10)
frame #38: _PyEval_EvalFrameDefault + 27276 (0x1026a8e54 in python3.10)
frame #39: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #40: method_vectorcall + 164 (0x1025c1498 in python3.10)
frame #41: call_function + 524 (0x1026acd4c in python3.10)
frame #42: _PyEval_EvalFrameDefault + 26612 (0x1026a8bbc in python3.10)
frame #43: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #44: _PyEval_EvalFrameDefault + 27276 (0x1026a8e54 in python3.10)
frame #45: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #46: _PyEval_EvalFrameDefault + 27276 (0x1026a8e54 in python3.10)
frame #47: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #48: context_run + 348 (0x1026c587c in python3.10)
frame #49: cfunction_vectorcall_FASTCALL_KEYWORDS + 112 (0x1026114e8 in python3.10)
frame #50: _PyEval_EvalFrameDefault + 27276 (0x1026a8e54 in python3.10)
frame #51: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #52: call_function + 524 (0x1026acd4c in python3.10)
frame #53: _PyEval_EvalFrameDefault + 26348 (0x1026a8ab4 in python3.10)
frame #54: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #55: call_function + 524 (0x1026acd4c in python3.10)
frame #56: _PyEval_EvalFrameDefault + 26348 (0x1026a8ab4 in python3.10)
frame #57: _PyEval_Vector + 2056 (0x1026a1ad0 in python3.10)
frame #58: method_vectorcall + 336 (0x1025c1544 in python3.10)
frame #59: thread_run + 180 (0x102774d98 in python3.10)
frame #60: pythread_wrapper + 48 (0x102710144 in python3.10)
frame #61: _pthread_start + 148 (0x19cf9ffa8 in libsystem_pthread.dylib)
frame #62: thread_start + 8 (0x19cf9ada0 in libsystem_pthread.dylib)

run.sh: line 3: 22444 Abort trap: 6           ./venv/bin/python launch-ui.py

[cryptrr]

Got the same error. On Mac Mini M2

[toby1991]

+1 mbp m1 pro

[hidecloud]

same error on Macbook M2 MAX

[graysonchen]

Same error on mbp m1 pro

[graysonchen]

https://github.com/Plachtaa/VALL-E-X/pull/102/files
I commented out the codes. I am using CPUs that are working now.

[cczw2010]

https://github.com/Plachtaa/VALL-E-X/pull/102/files I commented out the codes. I am using CPUs that are working now.

but the same error. Does the running code need any modifications?

[alexivaner]

https://github.com/Plachtaa/VALL-E-X/pull/102/files I commented out the codes. I am using CPUs that are working now.

If I comment MPS support part, I will get no hardware supported error

[alexivaner]

https://github.com/Plachtaa/VALL-E-X/pull/102/files I commented out the codes. I am using CPUs that are working now.

It is working now to use CPU, is it still not possible today to utilize MPS in Vall-E?

[liucr]

I met the same error. Here is the output of the command line terminal.

/Users//work/study/exercises/VALL-E-X/venv/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:30: UserWarning: torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.
  warnings.warn("torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.")
VALL-E EOS [0 -> 897]
libc++abi: terminating due to uncaught exception of type c10::Error: Unsupported type byte size: ComplexFloat
Exception raised from getGatherScatterScalarType at /Users/runner/work/pytorch/pytorch/pytorch/aten/src/ATen/native/mps/operations/View.mm:744 (most recent call first):
frame #0: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>> const&) + 92 (0x100dd78d0 in libc10.dylib)
frame #1: at::native::mps::getGatherScatterScalarType(at::Tensor const&) + 400 (0x113a2b2e0 in libtorch_cpu.dylib)
frame #2: invocation function for block in at::native::mps::gatherViewTensor(at::Tensor const&, at::Tensor&) + 132 (0x113a2a220 in libtorch_cpu.dylib)
frame #3: _dispatch_client_callout + 20 (0x1873c0910 in libdispatch.dylib)
frame #4: _dispatch_lane_barrier_sync_invoke_and_complete + 56 (0x1873cfcc4 in libdispatch.dylib)
frame #5: at::native::mps::gatherViewTensor(at::Tensor const&, at::Tensor&) + 896 (0x113a28c90 in libtorch_cpu.dylib)
frame #6: at::native::mps::mps_copy_(at::Tensor&, at::Tensor const&, bool) + 3896 (0x113952e58 in libtorch_cpu.dylib)
frame #7: at::native::copy_impl(at::Tensor&, at::Tensor const&, bool) + 2592 (0x10f1ebec0 in libtorch_cpu.dylib)
frame #8: at::native::copy_(at::Tensor&, at::Tensor const&, bool) + 100 (0x10f1eb3e0 in libtorch_cpu.dylib)
frame #9: at::_ops::copy_::call(at::Tensor&, at::Tensor const&, bool) + 292 (0x10ff71960 in libtorch_cpu.dylib)
frame #10: at::native::clone(at::Tensor const&, c10::optional<c10::MemoryFormat>) + 456 (0x10f56f018 in libtorch_cpu.dylib)
frame #11: at::_ops::clone::call(at::Tensor const&, c10::optional<c10::MemoryFormat>) + 280 (0x10fc35bd0 in libtorch_cpu.dylib)
frame #12: at::_ops::contiguous::call(at::Tensor const&, c10::MemoryFormat) + 280 (0x1100b8230 in libtorch_cpu.dylib)
frame #13: at::TensorBase::__dispatch_contiguous(c10::MemoryFormat) const + 40 (0x10f0321ac in libtorch_cpu.dylib)
frame #14: at::native::mps::binaryOpTensor(at::Tensor const&, at::Tensor const&, c10::Scalar const&, at::Tensor const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>>, MPSGraphTensor* (at::native::mps::BinaryOpCachedGraph*, MPSGraphTensor*, MPSGraphTensor*) block_pointer) + 940 (0x1139386b8 in libtorch_cpu.dylib)
frame #15: at::native::structured_mul_out_mps::impl(at::Tensor const&, at::Tensor const&, at::Tensor const&) + 112 (0x11393b3e4 in libtorch_cpu.dylib)
frame #16: c10::impl::wrap_kernel_functor_unboxed_<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor (at::Tensor const&, at::Tensor const&), &at::(anonymous namespace)::wrapper_MPS_mul_Tensor(at::Tensor const&, at::Tensor const&)>, at::Tensor, c10::guts::typelist::typelist<at::Tensor const&, at::Tensor const&>>, at::Tensor (at::Tensor const&, at::Tensor const&)>::call(c10::OperatorKernel*, c10::DispatchKeySet, at::Tensor const&, at::Tensor const&) + 136 (0x110d36ac8 in libtorch_cpu.dylib)
frame #17: at::_ops::mul_Tensor::call(at::Tensor const&, at::Tensor const&) + 288 (0x10fa3336c in libtorch_cpu.dylib)
frame #18: torch::autograd::THPVariable_mul(_object*, _object*, _object*) + 408 (0x102f5cc14 in libtorch_python.dylib)
frame #19: _object* torch::autograd::TypeError_to_NotImplemented_<&torch::autograd::THPVariable_mul(_object*, _object*, _object*)>(_object*, _object*, _object*) + 12 (0x102eb5384 in libtorch_python.dylib)
frame #20: method_vectorcall_VARARGS_KEYWORDS + 156 (0x10132b9c0 in Python)
frame #21: slot_nb_multiply + 236 (0x1013b3430 in Python)
frame #22: binary_op1 + 228 (0x1012f7ab0 in Python)
frame #23: PyNumber_Multiply + 36 (0x1012f80c0 in Python)
frame #24: _PyEval_EvalFrameDefault + 3844 (0x101446698 in Python)
frame #25: _PyEval_Vector + 360 (0x101443f28 in Python)
frame #26: method_vectorcall + 288 (0x101320c64 in Python)
frame #27: _PyEval_EvalFrameDefault + 1472 (0x101445d54 in Python)
frame #28: _PyEval_Vector + 360 (0x101443f28 in Python)
frame #29: method_vectorcall + 288 (0x101320c64 in Python)
frame #30: _PyEval_EvalFrameDefault + 1472 (0x101445d54 in Python)
frame #31: _PyEval_Vector + 360 (0x101443f28 in Python)
frame #32: _PyObject_FastCallDictTstate + 96 (0x10131cfb0 in Python)
frame #33: slot_tp_call + 196 (0x1013afccc in Python)
frame #34: _PyObject_MakeTpCall + 136 (0x10131ccf8 in Python)
frame #35: call_function + 380 (0x101453238 in Python)
frame #36: _PyEval_EvalFrameDefault + 23772 (0x10144b470 in Python)
frame #37: _PyEval_Vector + 360 (0x101443f28 in Python)
frame #38: PyVectorcall_Call + 140 (0x10131d800 in Python)
frame #39: _PyEval_EvalFrameDefault + 1472 (0x101445d54 in Python)
frame #40: _PyEval_Vector + 360 (0x101443f28 in Python)
frame #41: method_vectorcall + 124 (0x101320bc0 in Python)
frame #42: call_function + 132 (0x101453140 in Python)
frame #43: _PyEval_EvalFrameDefault + 17484 (0x101449be0 in Python)
frame #44: _PyEval_Vector + 360 (0x101443f28 in Python)
frame #45: _PyEval_EvalFrameDefault + 1472 (0x101445d54 in Python)
frame #46: _PyEval_Vector + 360 (0x101443f28 in Python)
frame #47: call_function + 132 (0x101453140 in Python)
frame #48: _PyEval_EvalFrameDefault + 17352 (0x101449b5c in Python)
frame #49: _PyEval_Vector + 360 (0x101443f28 in Python)
frame #50: pyrun_file + 308 (0x1014aec54 in Python)
frame #51: _PyRun_SimpleFileObject + 336 (0x1014ae398 in Python)
frame #52: _PyRun_AnyFileObject + 216 (0x1014ad9e4 in Python)
frame #53: pymain_run_file_obj + 180 (0x1014d9dd0 in Python)
frame #54: pymain_run_file + 72 (0x1014d9470 in Python)
frame #55: pymain_run_python + 300 (0x1014d8a58 in Python)
frame #56: Py_RunMain + 24 (0x1014d88ec in Python)
frame #57: pymain_main + 56 (0x1014d9f78 in Python)
frame #58: Py_BytesMain + 40 (0x1014da23c in Python)
frame #59: start + 2360 (0x1871f10e0 in dyld)

zsh: abort      python test1.py

I attached the runtime environment information for you to refer to.

Collecting environment information...
PyTorch version: 2.1.2
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A

OS: macOS 14.2 (arm64)
GCC version: Could not collect
Clang version: 15.0.0 (clang-1500.1.0.2.5)
CMake version: version 3.27.8
Libc version: N/A

Python version: 3.10.11 (v3.10.11:7d4cc5aa85, Apr  4 2023, 19:05:19) [Clang 13.0.0 (clang-1300.0.29.30)] (64-bit runtime)
Python platform: macOS-14.2-arm64-arm-64bit
Is CUDA available: False
CUDA runtime version: No CUDA
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Apple M3 Max

Versions of relevant libraries:
[pip3] numpy==1.26.2
[pip3] torch==2.1.2
[pip3] torchaudio==2.1.2
[pip3] torchvision==0.16.2
[conda] numpy                     1.24.3          py311hb57d4eb_0  
[conda] numpy-base                1.24.3          py311h1d85a46_0  
[conda] numpydoc                  1.5.0           py311hca03da5_0  
[conda] open-clip-torch           2.23.0                   pypi_0    pypi
[conda] pytorch                   2.2.0.dev20231206        py3.11_0    pytorch-nightly
[conda] pytorch-lightning         2.1.2                    pypi_0    pypi
[conda] pytorch-optimizer         2.12.0                   pypi_0    pypi
[conda] torch                     2.1.1                    pypi_0    pypi
[conda] torchaudio                2.1.1                    pypi_0    pypi
[conda] torchdiffeq               0.2.3                    pypi_0    pypi
[conda] torchmetrics              1.2.0                    pypi_0    pypi
[conda] torchsde                  0.2.6                    pypi_0    pypi
[conda] torchvision               0.16.1                   pypi_0    pypi

Here is the test1.py code that I ran:

from utils.generation import SAMPLE_RATE, generate_audio, preload_models
from scipy.io.wavfile import write as write_wav
from IPython.display import Audio

# download and load all models
preload_models()

# generate audio from text
text_prompt = """
Hello, my name is Nose. And uh, and I like hamburger. Hahaha... But I also have other interests such as playing tactic toast.
"""
audio_array = generate_audio(text_prompt)

# save audio to disk
write_wav("vallex_generation.wav", SAMPLE_RATE, audio_array)

# play text in notebook
Audio(audio_array, rate=SAMPLE_RATE)

[KodeurKubik]

Get same error +1

[KodeurKubik]

Just found a way around, it works now for me.
utils/generation.py line 30:

Force the device to CPU

device = torch.device("cpu")
if torch.cuda.is_available():
    device = torch.device("cuda", 0)
# if torch.backends.mps.is_available():
    # device = torch.device("mps")

[graysonchen]

@KodeurKubik yep. But not all the relevant code is commented out. plz check that comment and PR: #109 (comment)