-
def fuse_conv(conv, norm):
fused_conv = torch.nn.Conv2d(conv.in_channels,
conv.out_channels,
kernel_size=conv.kernel_size,
…
T-wow updated
1 month ago
-
Hi! Thank you for your amazing work!
I'm having some trouble on comparing the fused swiglu kernel with its plain pytorch version.
I checked the wall clock time with code below, and it gives me l…
-
## Description
Consider adding additional FusedCrossEntropyLoss kernel to FOAK set of kernels given the additional improvement seen using it in earlier tests (See Background below).
Considerati…
-
Hello author, thank you for your contribution to the pansharpening deep learning toolbox. When I was running run_test_manshharpening. py, the saved test result (i.e. the fused image file) was empty. H…
-
I have followed this program since Aug, 2023. (Many thanks to the relevant workers!) Recently, I noticed the program updated a lot. Therefore, I also updated the codes in my computer. However, when I …
-
When I tried to download datasets and models, I found that my account did not have the necessary permissions. When I attempted to apply for permissions, I did not receive approval. My site is in mainl…
-
Hello, in your paper, you mentioned some comparison algorithms, such as YDTR, but I noticed that their fused results are all grayscale images. Could you please elaborate on the method you used to conv…
-
The dependencies listed in the docs at https://docs.fused.io/python-sdk/dependencies/
are not the same as those found by running a UDF that inspects the installed packages:
```
@fused.udf
def ud…
-
### System Info
Debian 11
`nvidia-smi`
```
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.161.07 Driver Version: 535.16…
-
Here is an example traceback of me trying to use `heavyball` in TorchTitan. It seems that `heavyball` is not yet compatible with FSDP2 due to some of the utilities operating on tensors instead of dten…