Thrsu opened a new pull request, #15935:
URL: https://github.com/apache/tvm/pull/15935
This pull request addresses a TypeError that occurs in the interpolate
function when the scale_factor parameter is of type tuple. The current
implementation fails to handle cases where scale_factor is a tuple, resulting
in the error message: "TypeError: int() argument must be a string, a bytes-like
object or a number, not 'list'."
This fix addresses the underlying problem and allows the interpolate
function to handle both scalar and tuple values for the scale_factor parameter
without raising a TypeError.
And this bug can be reproduced by the following script:
```python
import torch
from torch import fx
from torch.nn import Module
import tvm
from tvm import relax
from tvm.relax.frontend.torch import from_fx
input_data = torch.randn([1, 2, 4, 4], dtype=torch.float32)
class interpolate(Module):
def forward(self, input):
return torch.nn.functional.interpolate(input, size=None,
scale_factor=(2.0, 1.0), mode='bilinear', align_corners=False,)
model = interpolate().float()
input_data = [input_data]
input_info = list(zip([list(inp.shape) for inp in input_data],
[str(inp.dtype) for inp in input_data]))
fx_model : torch.fx.GraphModule = fx.symbolic_trace(model)
with torch.no_grad():
mod = from_fx(fx_model, input_info)
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
