LyricZhao edited a comment on issue #10369:
URL: https://github.com/apache/tvm/issues/10369#issuecomment-1052914658
I'm using commit id `cb7f773218c26c3092cc275ef28e71839ebd18e3` or TVM 0.8
release tag (both have this problem).
The logging information is in
[https://pastebin.ubuntu.
LyricZhao commented on issue #10369:
URL: https://github.com/apache/tvm/issues/10369#issuecomment-1052914658
I'm using commit id `cb7f773218c26c3092cc275ef28e71839ebd18e3` or TVM 0.8
release tag (both have this problem).
The logging information is in
[https://pastebin.ubuntu.com/p/Q
This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch last-successful
in repository https://gitbox.apache.org/repos/asf/tvm.git.
from 4a9db23 meta schedule misc update (#10389)
add 01f306f Fix tvmc run error message when inputs aren't
masahi commented on issue #10223:
URL: https://github.com/apache/tvm/issues/10223#issuecomment-1052837280
https://github.com/apache/tvm/tree/main/gallery/how_to/tune_with_autoscheduler
has e2e examples of using the auto scheduler. But yeah, I don't expect it to
beat cuDNN, unless cuDNN im
masahi edited a comment on issue #10397:
URL: https://github.com/apache/tvm/issues/10397#issuecomment-1052831955
You can use this change
https://github.com/apache/tvm/commit/c3c57c60a97761719d43781ae61bcc85ce6d11c0
I'm not sure if we are supporting fp16 cuda math function properly, b
masahi commented on issue #10397:
URL: https://github.com/apache/tvm/issues/10397#issuecomment-1052831955
You can use this change
https://github.com/apache/tvm/commit/c3c57c60a97761719d43781ae61bcc85ce6d11c0
I'm not sure if we are supporting fp16 cuda math function properly, but for
ganler commented on issue #10369:
URL: https://github.com/apache/tvm/issues/10369#issuecomment-1052650805
@LyricZhao Thanks for reporting this bug. In the meantime, could you also
provide more information? e.g.,
1. which version (or git hash) you are using?
2. full error message;
This is an automated email from the ASF dual-hosted git repository.
areusch pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git.
from 4a9db23 meta schedule misc update (#10389)
add 01f306f Fix tvmc run error message when inputs aren't found. (#1001
areusch merged pull request #10017:
URL: https://github.com/apache/tvm/pull/10017
--
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: commits-unsubscr
masahi commented on a change in pull request #10396:
URL: https://github.com/apache/tvm/pull/10396#discussion_r815345574
##
File path: tests/python/frontend/pytorch/test_forward.py
##
@@ -4198,4 +4205,5 @@ def forward(self, x):
if __name__ == "__main__":
-pytest.main([
masahi commented on a change in pull request #10396:
URL: https://github.com/apache/tvm/pull/10396#discussion_r815345554
##
File path: tests/python/frontend/pytorch/test_forward.py
##
@@ -1090,7 +1092,12 @@ def forward(self, x, w, s):
styles = torch.rand(b)
# cuda n
JCBrouwer commented on issue #10223:
URL: https://github.com/apache/tvm/issues/10223#issuecomment-1052408332
> How TVM + cuDNN compares to PT? Since you are running on fp16, I'd hope
that we can use tensorcore. But I've never seen grouped convolution running on
tensorcore. Also cutlass is
JCBrouwer opened a new issue #10397:
URL: https://github.com/apache/tvm/issues/10397
### Expected behavior
It should be possible to take the absolute value of a half precision tensor.
### Actual behavior
The kernel fails to compile due to hfabs not being found:
This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch last-successful
in repository https://gitbox.apache.org/repos/asf/tvm.git.
from 37f7e57 Fix plint error. (#10394)
add 4a9db23 meta schedule misc update (#10389)
No new revisions w
JCBrouwer opened a new pull request #10396:
URL: https://github.com/apache/tvm/pull/10396
Some discussion here: #10223
@masahi
--
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 specifi
leeexyz commented on issue #10324:
URL: https://github.com/apache/tvm/issues/10324#issuecomment-1052135454
@wrongtest Thanks for your fix! Actually, there are two locations that need
to assert the empty, `Line 75` and `Line 143`, as I mentioned in this issue.
But, still, the issue is
This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch last-successful
in repository https://gitbox.apache.org/repos/asf/tvm.git.
from 920654c [Bugfix][TVMScript] Convert BufferSlice to BufferLoad when
used as range/loop start and end (#103
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git.
from 37f7e57 Fix plint error. (#10394)
add 4a9db23 meta schedule misc update (#10389)
No new revisions were added by th
masahi merged pull request #10389:
URL: https://github.com/apache/tvm/pull/10389
--
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: commits-unsubscr.
manupa-arm commented on issue #10300:
URL: https://github.com/apache/tvm/issues/10300#issuecomment-1051989970
@ekalda @lhutton1 , so it seems the flaky ness happens to come from the run,
until we know better, I would suggest adding a retry to the FVP run. WDYT?
--
This is an automated me
masahi commented on issue #10223:
URL: https://github.com/apache/tvm/issues/10223#issuecomment-1051988687
oops good find! Can you send a PR? I can quickly merge it (if I do it you
need to wait until next week).
> Sadly I'm still just a few FPS shy of my performance target so I'll hav
JCBrouwer edited a comment on issue #10223:
URL: https://github.com/apache/tvm/issues/10223#issuecomment-1051938367
Hello @masahi , sorry for the slow response, I somehow missed the
notification on this one, thanks for enabling the op!
I took a look at running on your branch and was
JCBrouwer edited a comment on issue #10223:
URL: https://github.com/apache/tvm/issues/10223#issuecomment-1051938367
Hello @masahi , sorry for the slow response, I somehow missed the
notification on this one, thanks for enabling the op!
I took a look at running on your branch and was
JCBrouwer commented on issue #10223:
URL: https://github.com/apache/tvm/issues/10223#issuecomment-1051938367
Hello @masahi , sorry for the slow response, I somehow missed the
notification on this one, thanks for enabling the op!
I took a look at running on your branch and was also ge
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git.
from 8b30198 [FIX,PROFILING] Add extra precision to numbers when
serializing to json (#10392)
add 37f7e57 Fix plint err
masahi merged pull request #10394:
URL: https://github.com/apache/tvm/pull/10394
--
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: commits-unsubscr.
26 matches
Mail list logo