fix failures- cpu offloading casues device mismatch L2_dynamo_compile…#4089
fix failures- cpu offloading casues device mismatch L2_dynamo_compile…#4089
Conversation
| f"Conversion of module {gm} not currently fully supported or convertible!", | ||
| exc_info=True, | ||
| ) | ||
| raise |
There was a problem hiding this comment.
If we are throwing exceptions, provide an actual error type and explanation
There was a problem hiding this comment.
The bare raise should be able to catch the same exception object or the UnsupportedOperatorException. Do we need another exception message. That would be code repeat. But can add.
| exp_program, | ||
| inputs=[input_tensor], | ||
| enabled_precisions={torch.float8_e4m3fn}, | ||
| use_explicit_typing=False, |
There was a problem hiding this comment.
This is deprecated, this test should be ported to explicit typing
775cd03 to
e33ca90
Compare
|
@narendasan had discussion with @zewenli98. Should we remove this https://github.com/pytorch/TensorRT/blob/main/py/torch_tensorrt/dynamo/_compiler.py#L641-648 from the code, since explicit typing means that we should not provide precision now. |
… tests, set explicit typing false for precisions test L2_torch_tensorrt_py_tests, and add raise exceptions for error logging
794f9d9 to
e06de0f
Compare
No description provided.