[inference Zero-Dim][trt]Unary operation support 0d #53506
Conversation
|
你的PR提交成功，感谢你对开源项目的贡献! |
|
❌ The PR is not created using PR's template. You can refer to this Demo. |
| + runtime_version[2] * 10 | ||
| < 8600 | ||
| and self.dims == 0 | ||
| ): |
There was a problem hiding this comment.
- A common function can be extracted to calculate TensorRT version number
- In line 174, use
ver = paddle_infer.get_trt_compile_version()to get version
There was a problem hiding this comment.
- A common function can be extracted to calculate TensorRT version number
- In line 174, use
ver = paddle_infer.get_trt_compile_version()to get version
When the OP behavior is inconsistent among different TensorRT versions, using the runtime version is more appropriate than the compiled version, making it easier to test multiple versions.
| @@ -60,6 +60,7 @@ def generate_int_input(dims, batch, attrs: List[Dict[str, Any]]): | |||
| "sin", | |||
| "cos", | |||
| "tan", | |||
| "tanh", | |||
There was a problem hiding this comment.
If remove tanh, where to test it？
There was a problem hiding this comment.
If remove
tanh, where to test it？
move tanhfrom test_trt_convert_activation.py to test_trt_convert_unary.py.
zhangjun
changed the title
zhangjun
changed the title
zhwesky2010
changed the title
PR types
Others
PR changes
Others
Description