(tvm) branch nightly updated (268d15c987 -> 1698793226)
This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch nightly in repository https://gitbox.apache.org/repos/asf/tvm.git from 268d15c987 [CI] Fix CI Script and Broken Tests (#16521) add 1698793226 [CI] Disable testing that downloads from mxnet (#16546) No new revisions were added by this update. Summary of changes: gallery/how_to/compile_models/from_mxnet.py | 8 +++- gallery/how_to/deploy_models/deploy_model_on_nano.py | 8 +++- gallery/how_to/deploy_models/deploy_model_on_rasp.py | 8 +++- gallery/how_to/deploy_models/deploy_quantized.py | 12 +--- .../how_to/extend_tvm/bring_your_own_datatypes.py| 8 +++- .../tune_with_autoscheduler/tune_network_arm.py | 11 --- .../tune_with_autoscheduler/tune_network_cuda.py | 8 ++-- .../tune_with_autoscheduler/tune_network_mali.py | 10 -- .../tune_with_autoscheduler/tune_network_x86.py | 20 +--- rust/tvm/examples/resnet/build.rs| 6 ++ rust/tvm/examples/resnet/src/main.rs | 5 + tests/python/frontend/mxnet/test_forward.py | 5 - .../quantization/test_quantization_accuracy.py | 19 +-- vta/scripts/tune_resnet.py | 8 ++-- vta/tutorials/autotvm/tune_alu_vta.py| 7 ++- vta/tutorials/autotvm/tune_relay_vta.py | 8 +++- vta/tutorials/frontend/deploy_classification.py | 7 ++- 17 files changed, 125 insertions(+), 33 deletions(-)
[PR] [KVCache] Fix attention kernel for ROCm [tvm]
MasterJH5574 opened a new pull request, #16551: URL: https://github.com/apache/tvm/pull/16551 When compiling a TensorIR function to ROCm backends, we need to be careful to store and load local registers in a consistent approach (either using vector access or scalar access), or otherwise there will be correctness issue. This PR fixes the attention kernel for ROCm. Meanwhile, it adds tests for float32 dtype and head dim other than 128. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [Unity][TVMScript] Optionally hide StructInfo that can be inferred [tvm]
Lunderberg commented on PR #16356: URL: https://github.com/apache/tvm/pull/16356#issuecomment-1936593774 Merged from main to PR branch, to resolve CI breakage that was fixed in https://github.com/apache/tvm/pull/16546. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [Lint] Add check to prevent usage of #include [tvm]
Lunderberg commented on PR #16412: URL: https://github.com/apache/tvm/pull/16412#issuecomment-1936592785 Merged from main to PR branch, to resolve CI breakage that was fixed in https://github.com/apache/tvm/pull/16546. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [Relax] Support callback as argument [tvm]
Lunderberg commented on PR #16542: URL: https://github.com/apache/tvm/pull/16542#issuecomment-1936581062 Merged from main to PR branch, to resolve CI breakage that was fixed in https://github.com/apache/tvm/pull/16546. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [TIR] Fix segfaults from ordering of Let/Assert in MakePackedAPI [tvm]
Lunderberg commented on PR #16543: URL: https://github.com/apache/tvm/pull/16543#issuecomment-1936580514 Merged from main to PR branch, to resolve CI breakage that was fixed in https://github.com/apache/tvm/pull/16546. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [TIR] Expand debug symbol output for CodeGenLLVM [tvm]
Lunderberg commented on PR #16544: URL: https://github.com/apache/tvm/pull/16544#issuecomment-1936579862 Merged from main to PR branch, to resolve CI breakage that was fixed in https://github.com/apache/tvm/pull/16546. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
(tvm) branch main updated: [CI] Disable testing that downloads from mxnet (#16546)
This is an automated email from the ASF dual-hosted git repository.
lunderberg pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new 1698793226 [CI] Disable testing that downloads from mxnet (#16546)
1698793226 is described below
commit 1698793226ef326af2a8559918b6dac20a2e0932
Author: Eric Lunderberg
AuthorDate: Fri Feb 9 14:42:25 2024 -0600
[CI] Disable testing that downloads from mxnet (#16546)
* [Testing] Disable testing that downloads from mxnet
The mxnet project was archived in
2023-09 ([link](https://attic.apache.org/projects/mxnet.html)).
Downloads from the mxnet AWS
links (e.g.
`https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip`)
no longer can be performed. This commit disables TVM CI steps that
depend on downloading models from these locations.
* Disable build.rs in resnet Rust example
---
gallery/how_to/compile_models/from_mxnet.py | 8 +++-
gallery/how_to/deploy_models/deploy_model_on_nano.py | 8 +++-
gallery/how_to/deploy_models/deploy_model_on_rasp.py | 8 +++-
gallery/how_to/deploy_models/deploy_quantized.py | 12 +---
.../how_to/extend_tvm/bring_your_own_datatypes.py| 8 +++-
.../tune_with_autoscheduler/tune_network_arm.py | 11 ---
.../tune_with_autoscheduler/tune_network_cuda.py | 8 ++--
.../tune_with_autoscheduler/tune_network_mali.py | 10 --
.../tune_with_autoscheduler/tune_network_x86.py | 20 +---
rust/tvm/examples/resnet/build.rs| 6 ++
rust/tvm/examples/resnet/src/main.rs | 5 +
tests/python/frontend/mxnet/test_forward.py | 5 -
.../quantization/test_quantization_accuracy.py | 19 +--
vta/scripts/tune_resnet.py | 8 ++--
vta/tutorials/autotvm/tune_alu_vta.py| 7 ++-
vta/tutorials/autotvm/tune_relay_vta.py | 8 +++-
vta/tutorials/frontend/deploy_classification.py | 7 ++-
17 files changed, 125 insertions(+), 33 deletions(-)
diff --git a/gallery/how_to/compile_models/from_mxnet.py
b/gallery/how_to/compile_models/from_mxnet.py
index 0694d2aed0..132f098d92 100644
--- a/gallery/how_to/compile_models/from_mxnet.py
+++ b/gallery/how_to/compile_models/from_mxnet.py
@@ -37,6 +37,7 @@ https://mxnet.apache.org/versions/master/install/index.html
# sphinx_gallery_start_ignore
# sphinx_gallery_requires_cuda = True
# sphinx_gallery_end_ignore
+import sys
import mxnet as mx
import tvm
import tvm.relay as relay
@@ -51,7 +52,12 @@ from mxnet.gluon.model_zoo.vision import get_model
from PIL import Image
from matplotlib import pyplot as plt
-block = get_model("resnet18_v1", pretrained=True)
+try:
+block = get_model("resnet18_v1", pretrained=True)
+except RuntimeError:
+print("Downloads from mxnet no longer supported", file=sys.stderr)
+sys.exit(0)
+
img_url = "https://github.com/dmlc/mxnet.js/blob/main/data/cat.png?raw=true";
img_name = "cat.png"
synset_url = "".join(
diff --git a/gallery/how_to/deploy_models/deploy_model_on_nano.py
b/gallery/how_to/deploy_models/deploy_model_on_nano.py
index abd0b3fab6..761187e2d7 100644
--- a/gallery/how_to/deploy_models/deploy_model_on_nano.py
+++ b/gallery/how_to/deploy_models/deploy_model_on_nano.py
@@ -106,12 +106,18 @@ from tvm.contrib.download import download_testdata
# `MXNet Gluon model zoo
`_.
# You can found more details about this part at tutorial
:ref:`tutorial-from-mxnet`.
+import sys
+
from mxnet.gluon.model_zoo.vision import get_model
from PIL import Image
import numpy as np
# one line to get the model
-block = get_model("resnet18_v1", pretrained=True)
+try:
+block = get_model("resnet18_v1", pretrained=True)
+except RuntimeError:
+print("Downloads from mxnet no longer supported", file=sys.stderr)
+sys.exit(0)
##
# In order to test our model, here we download an image of cat and
diff --git a/gallery/how_to/deploy_models/deploy_model_on_rasp.py
b/gallery/how_to/deploy_models/deploy_model_on_rasp.py
index de4ed9aff0..5196ae9ce1 100644
--- a/gallery/how_to/deploy_models/deploy_model_on_rasp.py
+++ b/gallery/how_to/deploy_models/deploy_model_on_rasp.py
@@ -99,12 +99,18 @@ from tvm.contrib.download import download_testdata
# `MXNet Gluon model zoo
`_.
# You can found more details about this part at tutorial
:ref:`tutorial-from-mxnet`.
+import sys
+
from mxnet.gluon.model_zoo.vision import get_model
from PIL import Image
import numpy as np
# one line to get the model
-block = get_model("resnet18_v1", pretrained=True)
+try:
+block = g
Re: [PR] [CI] Disable testing that downloads from mxnet [tvm]
Lunderberg merged PR #16546: URL: https://github.com/apache/tvm/pull/16546 -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [CI] Disable testing that downloads from mxnet [tvm]
Lunderberg commented on PR #16546: URL: https://github.com/apache/tvm/pull/16546#issuecomment-1936576065 Thank you for mentioning that. Looks like the ARM CI was related to another of my PRs, which I recently rebased onto main. I've unstuck the ARM tests, and CI has now passed here. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[PR] [RUNTIME][ROCM] Properly align rocm parameter buffer [tvm]
MasterJH5574 opened a new pull request, #16550:
URL: https://github.com/apache/tvm/pull/16550
This PR properly aligns rocm param buffer fields to avoid bug in cases like
```c
struct Param {
void *arg0;
int32_t arg1;
int64_t arg2;
};
```
In this case, we need to align arg2 by padding a i32 after arg1.
---
Co-authored-by: Tianqi Chen
--
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...@tvm.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Re: [I] [CI Problem] CI-arm build queue blocked [tvm]
Lunderberg commented on issue #16549: URL: https://github.com/apache/tvm/issues/16549#issuecomment-1936553425 Short-term, I've stopped all tasks related to #16183, and the ARM queue is recovering. All remaining tasks are related to other PRs, as it works through the backlog.  Long-term, it looks like there's a Jenkins option `disableConcurrentBuilds(abortPrevious: true)` ([stack overflow link](https://stackoverflow.com/questions/46343348/build-only-latest-commit-on-branch-with-jenkins), [GH link](https://github.com/jenkinsci/workflow-job-plugin/pull/200))) that we should enable. If there's two concurrent builds for the same PR, it would cancel the previous one. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [I] [CI Problem] CI-arm build queue blocked [tvm]
Lunderberg commented on issue #16549: URL: https://github.com/apache/tvm/issues/16549#issuecomment-1936489408 This PR is also one that impacts `packed_func.h`, which is included by practically everything in TVM. So, not only is it re-building TVM once for each rebased commit, but they're also full rebuilds that can't benefit from `ccache`. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [I] [CI Problem] CI-arm build queue blocked [tvm]
Lunderberg commented on issue #16549: URL: https://github.com/apache/tvm/issues/16549#issuecomment-1936484947 Huh, that definitely wasn't intended with that PR. The large number of commits was due to rebasing onto main, where I had expected that only the last commit on the rebased branch would be built, as the intermediate commits don't require testing. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [CI] Disable testing that downloads from mxnet [tvm]
yongwww commented on PR #16546: URL: https://github.com/apache/tvm/pull/16546#issuecomment-1936444704 arm tests are stuck https://github.com/apache/tvm/issues/16549 -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [CI] Disable testing that downloads from mxnet [tvm]
Lunderberg commented on PR #16546: URL: https://github.com/apache/tvm/pull/16546#issuecomment-1936329639 Looks like CI is passing, except for the ARM tests, which are stalled waiting for `ARM-GRAVITON3` instances. 1/8 has run since I checked this morning, so hopefully they'll get queued up sooner rather than later. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [microNPU][ETHOSU] Add fixed point for matmul [tvm]
Aleksei-grovety commented on PR #16401: URL: https://github.com/apache/tvm/pull/16401#issuecomment-1936187639 > For example, is it possible to insert `relay.fixed_point_multiply` operators on the inputs/outputs of the graph in between the matmul Yes, it's possible, I've updated the code. -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[I] [CI Problem] CI-arm build queue blocked [tvm]
Liam-Sturge opened a new issue, #16549: URL: https://github.com/apache/tvm/issues/16549 Jobs that require an arm agent are struggling to find an Graviton-3 executor in Jenkins. There are hundreds of builds queued for ci-arm. Some of the builds queued are stuck. This appears to have started on 8th February. ### Branch/PR Failing https://github.com/apache/tvm/pull/16183 may have caused the backlog of jobs. Recently a large amount of commits were pushed simultaneously for this PR and builds have been scheduled for each commit. I have attached a graph of the agent queue and allocation, which shows a big spike in queued jobs.  ### Jenkins Link Build logs on these jobs indicate that some of the builds are stuck. They have ended with an unfinished status, but remain in the queue. See this console log as one example of many: https://ci.tlcpack.ai/job/tvm-arm/job/PR-16183/495/console ### Triage Please refer to the list of label tags [here](https://github.com/apache/tvm/wiki/Issue-Triage-Labels) to find the relevant tags and add them below in a bullet format (example below). * needs-triage CC @lhutton1 @konturn @tqchen @yongwww -- 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...@tvm.apache.org.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
Re: [PR] [Build] Improving debug and build-dir options [tvm]
lhutton1 commented on PR #16524: URL: https://github.com/apache/tvm/pull/16524#issuecomment-1935497678 cc @Liam-Sturge -- 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...@tvm.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
