bug描述 Describe the Bug 在Aistudio上，我使用paddle2.2.2运行ernie的SKEP（paddlenlp）预训练模型时，模型训练速度为0.8step/s，在同等条件下（V100 32G显存），paddle升级到2.4.1，就会运行非常慢，共运行过四次，有一次是0.01step/s，有三次是直接内存溢出。请帮忙排查一下是否是paddle 2.4.1和paddlenlp 2.5.0之间的...

问题描述 Please describe your issue Commit: 7f0eb2e causes a significant performance drop in BF16 models: mobilenetV1 mobilenetV2 mobilenetV3 googlenet resnet50 Tests was performed on two machines: CPX: Turbo: OFF Name: Intel(R) d on(R) Gold 6348H CPU @ 2.30GHz Mo...

文档链接&描述 Document Links & Description 选择2.4Stable Linux Pip CUDA11.6时生成了以下命令： python -m pip install paddlepaddle-gpu==2.4.1.post116 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html 实际上不存在2.4.1.post116 应为 2.4.1，即： python -m pip install paddlepaddle-g...

需求描述 Feature Description ubuntu 22.04 python3.10, pip3 only has paddlepaddle-gpu version 2, should has version 1 also 替代实现 Alternatives No response

问题描述 Issue Description During build in Debug build type I got an assertion error: [ 96%] Built target jit_layer [ 96%] Built target kernel_signature_generator [ 96%] Built target eager_legacy_op_function_generator [ 96%] Generating libdnnl.so.0 [ 96%] Generating libiomp5.so [ 96%] copy_if_differ...

bug描述 Describe the Bug paddle_paddle/build/python/paddle/tests/hapi_mnist_bf16_static.py test application is not working properly for fp32 mode. Returns an error: W0125 07:46:14.564448 53298 interpreter_util.cc:886] pool2d_grad raises an exception dnnl::error, could not create a primitive descript...

bug描述 Describe the Bug paddle_paddle/build/python/paddle/tests/hapi_mnist_bf16_static.py test application is not working properly. Returns an error: I0116 23:08:01.824071 51418 interpretercore.cc:258] New Executor is Running. I0116 23:08:01.824502 51418 onednn_context.cc:80] oneDNN v2.6.0 W0116 23...

问题描述 Issue Description Paddle安装成功，但运行检查脚本时出错 报错信息 Running verify PaddlePaddle program ... W0124 21:56:10.481693 1080 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 6.1, Driver API Version: 12.0, Runtime API Version: 11.6 W0124 21:56:10.499675 1080 gpu_...

bug描述 Describe the Bug paddle.distribution.Categorical的log_prob和probs函数与torch不符，甚至出现prob为负、log_prob为nan的情况。 版本：2.4.1.post112 import paddle import torch logit = [-0.00045110, 0.00965287, 0.00661326] dist = paddle.distribution.Categorical(paddle.to_tensor(logit)) log_prob =...

请提出你的问题 Please ask your question 载入静态模型到 translated_layer后无法使用。代码如下： import paddle translated_layer = paddle.jit.load("inference_model/yolov3_darknet53_270e_coco/model")##已经载入成功 program = translated_layer.program() print((program)) translated_layer.eval() x = paddle.ze...