torch_mir_eval:https的反向传播pytorch实现

时间:2021-05-16 17:36:36
【文件属性】:
文件名称:torch_mir_eval:https的反向传播pytorch实现
文件大小:21KB
文件格式:ZIP
更新时间:2021-05-16 17:36:36
Python torch_mir_eval Pytorch实施 。 Nvidia RTX 3090,ADM Threadripper 1920X单次运行 .bss_eval_sources test> permutation: False float32 CPU: 3.004 torch-CPU: 2.512 .bss_eval_sources test> Compute permutation: False float32 CPU: 2.590 GPU: 1.616 .bss_eval_sources test> compute_permutation: True float32 CPU: 13.195 torch-CPU: 12.326 .bss_eval_sources test> compute_permutation: True float32 CPU: 13.407 GPU: 7
【文件预览】:
torch_mir_eval-main
----.github()
--------workflows()
----test.py(237B)
----requirements-dev.txt(44B)
----torch_mir_eval()
--------_version.py(27B)
--------__init__.py(75B)
--------batch_separation.py(15KB)
--------toeplitz.py(3KB)
--------separation.py(14KB)
----LICENSE(1KB)
----test()
--------test_bss_float64.py(4KB)
--------test_bss_float32.py(4KB)
--------utils.py(699B)
--------test_batched_bss_float32.py(4KB)
--------test_batched_bss_float64.py(7KB)
--------test_toeplitz.py(1KB)
----setup.py(765B)
----README.md(3KB)
----.gitignore(2KB)

网友评论