bug描述 Describe the Bug
环境：python3.7 ,paddlepaddle-gpu 2.4.2.post117,cuda 11.7,paddlenlp 2.5.2
模型推理时报错OSError: (External) CUBLAS error(1).
[Hint: Please search for the error code(1) on website (https://docs.nvidia.com/cuda/cublas/index.html#cublasstatus_t) to get Nvidia's official solution and advice about CUBLAS Error.] (at /paddle/paddle/phi/backends/gpu/gpu_resources.cc:156)
[operator < linear > error]
其他补充信息 Additional Supplementary Information
No response
4条答案
按热度按时间0md85ypi1#
你好，根据你提供的信息，不足以判断和定位问题。请补充问题背景和细节描述，或提供一份最小可复现代码，以方便我们分析问题。
fdbelqdn2#
您好，我也遇到了相同的问题，在使用paddlescience时报错：
File "PaddleScience/ppsci/arch/mlp.py", line 142, in forward_tensor
y = linear(y)
File "paddle/lib/python3.8/site-packages/paddle/nn/layer/layers.py", line 1254, in call
return self.forward(*inputs, **kwargs)
File "PaddleScience/ppsci/arch/mlp.py", line 48, in forward
return nn.functional.linear(input, weight, self.bias)
File "paddle/lib/python3.8/site-packages/paddle/nn/functional/common.py", line 1842, in linear
return _C_ops.linear(x, weight, bias)
OSError: (External) CUBLAS error(1).
[Hint: 'CUBLAS_STATUS_NOT_INITIALIZED'. The cuBLAS library was not initialized. This is usually caused by the lack of a prior cublasCreate() call, an error in the CUDA Runtime API called by the cuBLAS routine, or an error in the hardware setup. To correct: call cublasCreate() prior to the function call; and check that the hardware, an appropriate version of the driver, and the cuBLAS library are correctly installed. ] (at ../paddle/phi/backends/gpu/gpu_resources.cc:214)
[operator < linear > error]
同样的代码在使用AIstudio时没有报错
cetgtptt3#
你好，请问这个问题得到解决了吗
41ik7eoe4#
我也遇到了同样的问题，去年还能跑的，包版本和环境都没有变，但是运行报错了。