You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The operator should support all optional arguments defined in the interface.
算子应支持接口中定义的所有参数选项。
DDL 提交时间
Please submit a Pull Request within 1 week after accepting the assignment.
请于接取任务一周内提交PR。
Please provide both accuracy test and performance test code.
请同时提供实现正确性测试与性能测试代码。
The text was updated successfully, but these errors were encountered:
the forward function of diagonal is actually a view of input tensor, and it doesn't launch cuda kernel. this issue calls for implementation of diagonal_backward, which could be registered independently into aten library.
Description 任务介绍
Develop backward function for diagonal operator.
开发diagonal算子的反向功能。
Requirements 任务要求
Interface 接口
diagonal_backward(Tensor grad_output, SymInt[] input_sizes, int offset, int dim1, int dim2) -> Tensor
Function reference 功能参考
https://pytorch.org/docs/stable/generated/torch.diagonal.html#torch-diagonal
Implementation reference 实现参考
https://github.com/FlagOpen/FlagGems/blob/master/src/flag_gems/ops/diag.py
The operator should support all optional arguments defined in the interface.
算子应支持接口中定义的所有参数选项。
DDL 提交时间
Please submit a Pull Request within 1 week after accepting the assignment.
请于接取任务一周内提交PR。
Please provide both accuracy test and performance test code.
请同时提供实现正确性测试与性能测试代码。
The text was updated successfully, but these errors were encountered: