search

bytedeco / javacpp-presets

The missing Java distribution of native C++ libraries
Other
2.65k stars 738 forks source link

pytorch tensor mul api bug? #1271

Closed lzmchina closed 10 months ago

lzmchina commented 1 year ago

I realize a network in java like this:

public Tensor[] forward(Tensor... inputs) {
        Tensor x = inputs[0];
        Tensor hidden_state = inputs[1];
        // [r, u] = sigmoid(A[x, h]W + b)
        // [r, u] (batch_size, num_nodes * (2 * num_gru_units))
        Tensor concatenation = torch.sigmoid(this.graph_conv1.forward(x, hidden_state)[0]);
        // r (batch_size, num_nodes*num_gru_units)
        // u (batch_size, num_nodes* num_gru_units)
        TensorVector tensors = torch.chunk(concatenation, 2, 1);
        Tensor r = tensors.get(0);
        Tensor u = tensors.get(1);
        // c = tanh(A[x, (r * h)W + b])
        // c (batch_size, num_nodes * num_gru_units)
        Tensor c = torch.tanh(this.graph_conv2.forward(x, r.mul(hidden_state))[0]);
        // h := u * h + (1 - u) * c
        // h (batch_size, num_nodes * num_gru_units)
        Tensor tmp = this.constantOne.sub(u);
        Tensor new_hidden_state = u.mul(hidden_state).add(c.mul(tmp));
        return new Tensor[] { new_hidden_state, new_hidden_state };
    }

The problem is that after training thousands of steps the program will crash. The log shows:

Caused by: java.lang.RuntimeException: Could not run 'aten::mul.Tensor' with arguments from the 'FuncTorchGradWrapper' 
backend. This could be because the operator doesn't exist for this backend, or was omitted during the selective/custom build 
process (if using custom build). If you are a Facebook employee using PyTorch on mobile, please visit https://fburl.com/ptmfixes 
for possible resolutions. 'aten::mul.Tensor' is only available for these backends: [Dense, FPGA, ORT, Vulkan, Metal, Meta, 
Quantized, CustomRNGKeyId, MkldnnCPU, Sparse, SparseCsrCPU, SparseCsrCUDA, NestedTensor, BackendSelect, Python, Fake, 
Named, Conjugate, Negative, ZeroTensor, FuncTorchDynamicLayerBackMode, ADInplaceOrView, AutogradOther, AutogradFunctionality, AutogradNestedTensor, Tracer, AutocastCPU, AutocastXPU, Autocast, FuncTorchBatched, FuncTorchVmapMode, Batched, VmapMode, FuncTorchGradWrapper, Functionalize, DeferredInit, PythonTLSSnapshot, FuncTorchDynamicLayerFrontMode, TESTING_ONLY_GenericWrapper, TESTING_ONLY_GenericMode, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, CPU, CUDA, HIP, XLA, MPS, IPU, XPU, HPU, UNKNOWN_TENSOR_TYPE_ID, QuantizedXPU, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, SparseCPU, SparseCUDA, SparseHIP, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, SparseXPU, UNKNOWN_TENSOR_TYPE_ID, SparseVE, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, NestedTensorCUDA, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID, UNKNOWN_TENSOR_TYPE_ID].

Undefined: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
CPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCPU.cpp:37386 [kernel]
CUDA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64-gpu/pytorch/build/aten/src/ATen/RegisterCUDA.cpp:51977 [kernel]
HIP: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
XLA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
MPS: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
IPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
XPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
HPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
VE: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
Lazy: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
PrivateUse1: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
PrivateUse2: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
PrivateUse3: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
FPGA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
ORT: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
Vulkan: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
Metal: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
Meta: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterMeta.cpp:31637 [kernel]
QuantizedCPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
QuantizedCUDA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
QuantizedXPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
CustomRNGKeyId: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
MkldnnCPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterMkldnnCPU.cpp:690 [kernel]
SparseCPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterSparseCPU.cpp:1858 [kernel]
SparseCUDA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64-gpu/pytorch/build/aten/src/ATen/RegisterSparseCUDA.cpp:2018 [kernel]
SparseHIP: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
SparseXPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
SparseVE: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterCompositeExplicitAutograd.cpp:29545 [default backend kernel]
SparseCsrCPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterSparseCsrCPU.cpp:1507 [kernel]
SparseCsrCUDA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64-gpu/pytorch/build/aten/src/ATen/RegisterSparseCsrCUDA.cpp:1657 [kernel]
NestedTensorCPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterNestedTensorCPU.cpp:290 [kernel]
NestedTensorCUDA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64-gpu/pytorch/build/aten/src/ATen/RegisterNestedTensorCUDA.cpp:334 [kernel]
BackendSelect: fallthrough registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/core/BackendSelectFallbackKernel.cpp:3 [backend fallback]
Python: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/core/PythonFallbackKernel.cpp:133 [backend fallback]
Named: fallthrough registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/core/NamedRegistrations.cpp:11 [kernel]
Conjugate: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/ConjugateFallback.cpp:18 [backend fallback]
Negative: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/native/NegateFallback.cpp:18 [backend fallback]
ZeroTensor: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/build/aten/src/ATen/RegisterZeroTensor.cpp:198 [kernel]
ADInplaceOrView: fallthrough registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/core/VariableFallbackKernel.cpp:64 [backend fallback]
AutogradOther: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradCPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradCUDA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradXLA: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradMPS: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradIPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradXPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradHPU: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradLazy: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradPrivateUse1: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradPrivateUse2: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
AutogradPrivateUse3: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/VariableType_0.cpp:11935 [autograd kernel]
Tracer: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/torch/csrc/autograd/generated/TraceType_0.cpp:13506 [kernel]
AutocastCPU: fallthrough registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/autocast_mode.cpp:481 [backend fallback]
Autocast: fallthrough registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/autocast_mode.cpp:324 [backend fallback]
Batched: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/BatchingRegistrations.cpp:1068 [kernel]
VmapMode: fallthrough registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/VmapModeRegistrations.cpp:33 [backend fallback]
Functionalize: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/FunctionalizeFallbackKernel.cpp:89 [backend fallback]
PythonTLSSnapshot: registered at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/core/PythonFallbackKernel.cpp:137 [backend fallback]

Exception raised from reportError at /__w/javacpp-presets/javacpp-presets/pytorch/cppbuild/linux-x86_64/pytorch/aten/src/ATen/core/dispatch/OperatorEntry.cpp:440 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f29c8081477 in /home/lzm/.javacpp/cache/pytorch-1.12.1-1.5.8-linux-x86_64-gpu.jar/org/bytedeco/pytorch/linux-x86_64-gpu/libc10.so)
frame #1: <unknown function> + 0xa7947a (0x7f285ace747a in /home/lzm/.javacpp/cache/pytorch-1.12.1-1.5.8-linux-x86_64.jar/org/bytedeco/pytorch/linux-x86_64/libtorch_cpu.so)
frame #2: <unknown function> + 0x19ce8b3 (0x7f285bc3c8b3 in /home/lzm/.javacpp/cache/pytorch-1.12.1-1.5.8-linux-x86_64.jar/org/bytedeco/pytorch/linux-x86_64/libtorch_cpu.so)
frame #3: at::_ops::mul_Tensor::call(at::Tensor const&, at::Tensor const&) + 0x87 (0x7f285bbdb2e7 in /home/lzm/.javacpp/cache/pytorch-1.12.1-1.5.8-linux-x86_64.jar/org/bytedeco/pytorch/linux-x86_64/libtorch_cpu.so)
frame #4: Java_org_bytedeco_pytorch_Tensor_mul__Lorg_bytedeco_pytorch_Tensor_2 + 0xa6 (0x7f2848e17bb6 in /home/lzm/.javacpp/cache/pytorch-1.12.1-1.5.8-linux-x86_64-gpu.jar/org/bytedeco/pytorch/linux-x86_64-gpu/libjnitorch.so)
frame #5: [0x7f2a0476a497]

        at org.bytedeco.pytorch.Tensor.mul(Native Method)

Java thread heap info shows the line Tensor new_hidden_state = u.mul(hidden_state).add(c.mul(tmp)); crashes.

So, what's wrong with it?Thanks!

saudet commented 1 year ago

It's possible that it's running out of GPU memory. Make sure to deallocate any objects you do not need anymore as soon as possible. We can easily do that with PointerScope: http://bytedeco.org/news/2018/07/17/bytedeco-as-distribution/

lzmchina commented 1 year ago

Actually,it runs in CPU mode.I use gpu version of java-pytorch,but,the nvidia driver is too old, so it still use CPU

lzmchina commented 1 year ago

Actually,it runs in CPU mode.I use gpu version of java-pytorch,but,the nvidia driver is too old, so it still use CPU

I will try to use PointerScope right now to see whether it can fix the problem.

HGuillemet commented 11 months ago

Is it fixed ? Can you close the issue ?