-
Notifications
You must be signed in to change notification settings - Fork 5.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
【PaddlePaddle Hackathon 2】33、为Paddle修复ROCM(HIP)算子 #40300
Comments
您好,我们已经收到了您的问题,会安排技术人员尽快解答您的问题,请耐心等待。请您再次检查是否提供了清晰的问题描述、复现代码、环境&版本、报错信息等。同时,您也可以通过查看官网API文档、常见问题、历史Issue、AI社区来寻求解答。祝您生活愉快~ Hi! We've received your issue and please be patient to get responded. We will arrange technicians to answer your questions as soon as possible. Please make sure that you have posted enough message to demo your request. You may also check out the API,FAQ,Github Issue and AI community to get the answer.Have a nice day! |
【Hackathon 33】PR提交vim /public/home/acv2h3zdq0/Paddle/build/python/paddle/fluid/tests/unittests/test_trunc_op.py
42 self.check_output()
45 self.check_grad(['X'], 'Out', numeric_grad_delta=1e-5)
vim /public/home/acv2h3zdq0/Paddle/build/python/paddle/fluid/tests/unittests/test_where_op.py
251 assert np.array_equal(out.cpu().numpy(), np.where(cond_i, x_i, y_i))
269 result = result.cpu().numpy()
270 expect = np.where(cond.cpu(), a.cpu(), b.cpu()) 没有修改直接测试通过 vim /public/home/acv2h3zdq0/Paddle/build/python/paddle/fluid/tests/unittests/test_zeros_like_op.py
73 if core.is_compiled_with_cuda():
74 self.assertEqual((out.cpu().numpy() == np.zeros(shape, dtype)).all(),True)
75 else:
76 self.assertEqual((out.numpy() == np.zeros(shape, dtype)).all(),True)
78 if core.is_compiled_with_cuda():
79 self.assertEqual((out.cpu().numpy() == np.zeros(shape, dtype)).all(), True)
80 else:
81 self.assertEqual((out.numpy() == np.zeros(shape, dtype)).all(), True)
83 if core.is_compiled_with_cuda():
84 self.assertEqual((out.cpu().numpy() == np.zeros(shape, dtype)).all(), True)
85 else:
86 self.assertEqual((out.numpy() == np.zeros(shape, dtype)).all(), True)
87 paddle.enable_static() |
Since you haven't replied for more than a year, we have closed this issue/pr. |
(此 ISSUE 为 PaddlePaddle Hackathon 第二期活动的任务 ISSUE,更多详见 【PaddlePaddle Hackathon 第二期】任务总览)
【任务说明】
任务标题:为Paddle修复ROCM(HIP)算子
技术标签:深度学习框架,C++,Python,HIP
任务难度:简单
详细描述:飞桨支持在C86加速卡上运行,但是存在部分算子在C86加速卡下还无法通过算子单测。下表中为当前在C86加速卡下算子单测失败的用例,请根据单测错误提示,对列表中的算子C++文件和单测Python文件进行相应修复,在不删减单测Python文件中的单测用例的前提下,使得算子单测列表通过。
【修复建议】
【提交流程】
【提交内容】
算子修复代码,在Paddle repo的 paddle/fluid/operators 目录
单测修复代码,在Paddle repo的 python/paddle/fluid/tests/unittests 目录
单测修复之后,在C86加速卡环境下的成功运行结果截图
单测修复之后,每个算子的问题定位及修复思路的简单描述(1-2句话即可)
【合入标准】
【技术要求】
【参考内容】
【答疑交流】
在开发中对于上述任务有任何问题:请登录光合开发者社区 进行交流,可至 光合开发者社区-论坛-AI应用-paddle专栏 进行发帖。
更多任务详情、技术内容可在光合开发者社区【揭榜挂帅】、【资源工具】等更多页面查看。
The text was updated successfully, but these errors were encountered: