This is an automated email from the ASF dual-hosted git repository.
masahi pushed a commit to branch unity
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/unity by this push:
new 225d067fb8 [Unity] Support Padding Reversal in
masahi merged PR #15679:
URL: https://github.com/apache/tvm/pull/15679
--
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:
srkreddy1238 commented on code in PR #15833:
URL: https://github.com/apache/tvm/pull/15833#discussion_r1340435814
##
include/tvm/runtime/memory/memory_manager.h:
##
@@ -37,15 +37,15 @@
namespace tvm {
namespace runtime {
-namespace vm {
+namespace memory {
struct Buffer
srkreddy1238 commented on PR #15818:
URL: https://github.com/apache/tvm/pull/15818#issuecomment-1739147760
Never mind. Wile I was relooking into the code I found a reference the
deleted object and fixed it.
--
This is an automated message from the Apache Git Service.
To respond to the
Lunderberg commented on issue #15148:
URL: https://github.com/apache/tvm/issues/15148#issuecomment-1739257241
> I see what you're saying--maybe we can add min/max overrides for
`_test_forward_elemwise_quantized`.
That's what I'm thinking as well. It looks like it currently uses the
This is an automated email from the ASF dual-hosted git repository.
lunderberg pushed a commit to branch unity
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/unity by this push:
new aa4587feb5 [Unity] Implement
Lunderberg merged PR #15810:
URL: https://github.com/apache/tvm/pull/15810
--
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:
tlopex commented on issue #15148:
URL: https://github.com/apache/tvm/issues/15148#issuecomment-1739727222
@p3achyjr
```# calculate the mean and quantization scale for every input
tensor,
# with respect to its fp32 input range, defined in fake_quant.
Lunderberg commented on PR #15839:
URL: https://github.com/apache/tvm/pull/15839#issuecomment-1739748683
As part of this implementation, moving the `R.memory.kill_tensor` and
`R.memory.kill_storage` to be generated in a later `KillAfterLastUse` pass
meant that the `RemoveCUDAGraph` pass
tlopex commented on issue #15148:
URL: https://github.com/apache/tvm/issues/15148#issuecomment-1739729430
@p3achyjr
```
# with respect to its fp32 input range, defined in fake_quant.
# s = 255/(fmax-fmin); m = -fmin*s (the zero point)
for i
Lunderberg opened a new pull request, #15839:
URL: https://github.com/apache/tvm/pull/15839
Prior to this commit, the last kernel launch would not be included in
a captured CUDA graph. This commit updates `RewriteCUDAGraph` to
include the last kernel launch.
The previous
yongwww commented on code in PR #15833:
URL: https://github.com/apache/tvm/pull/15833#discussion_r1340399667
##
include/tvm/runtime/memory/memory_manager.h:
##
@@ -37,15 +37,15 @@
namespace tvm {
namespace runtime {
-namespace vm {
+namespace memory {
struct Buffer {
vinx13 commented on PR #15839:
URL: https://github.com/apache/tvm/pull/15839#issuecomment-1739959284
I didn't consider `R.builtin.alloc_tensor` previously because it can be
either dynamic allocation or memory visible to external (tensors for
input/output). The implication here is running
vinx13 commented on PR #15839:
URL: https://github.com/apache/tvm/pull/15839#issuecomment-1740116164
That's correct. It's relying on the current behavior of
`StaticBlockPlanMemory` that doesn't rewrite the input/output tensors. So only
tensors from `R.builtin.alloc_tensor` can be returned
masahi commented on PR #15839:
URL: https://github.com/apache/tvm/pull/15839#issuecomment-1739872085
cc @vinx13
--
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
zxybazh commented on code in PR #15823:
URL: https://github.com/apache/tvm/pull/15823#discussion_r1340704043
##
src/tir/transforms/default_gpu_schedule.cc:
##
@@ -98,24 +98,48 @@ IRModule MarkScheduled(const IRModule& mod) {
mod->type_definitions, //
Lunderberg commented on PR #15839:
URL: https://github.com/apache/tvm/pull/15839#issuecomment-1740113779
I'm afraid I don't quite understand. Either `R.memory.alloc_storage` or
`R.builtin.alloc_tensor` could be independent allocations that could be
returned as output. (e.g. A static
yongwww commented on PR #15686:
URL: https://github.com/apache/tvm/pull/15686#issuecomment-1739990584
@ibsidorenko thanks for the effort! looks we might not need use these two
ops, shall we close this pr?
--
This is an automated message from the Apache Git Service.
To respond to the
yongwww commented on PR #15772:
URL: https://github.com/apache/tvm/pull/15772#issuecomment-1740031754
cc: @slyubomirsky @psrivas2
--
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
This is an automated email from the ASF dual-hosted git repository.
srk 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 8b40f5d028 [FRONTEND] Fix unnecessary pylint errors
srkreddy1238 merged PR #15838:
URL: https://github.com/apache/tvm/pull/15838
--
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:
This is an automated email from the ASF dual-hosted git repository.
srk 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 def551dfd5 [CLI TOOLS][RTVM] Improve rtvm tool with new
srkreddy1238 merged PR #15818:
URL: https://github.com/apache/tvm/pull/15818
--
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:
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 cf081d9929 [BugFix][CPP] Fix cpp deploy bug (#15773)
add 9d8e6fda50 [ADRENO] Minor changes for Adreno docs and
lhutton1 opened a new pull request, #15836:
URL: https://github.com/apache/tvm/pull/15836
The ci-gpu is updated to make changes in #15819 reflect in CI.
cc @ashutosh-arm
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
srkreddy1238 commented on code in PR #15833:
URL: https://github.com/apache/tvm/pull/15833#discussion_r1339827581
##
include/tvm/runtime/memory_manager.h:
##
@@ -130,7 +131,7 @@ class MemoryManager {
class StorageObj : public Object {
public:
/*! \brief The index into the
masahi opened a new pull request, #15837:
URL: https://github.com/apache/tvm/pull/15837
(no comment)
--
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,
echuraev commented on code in PR #15818:
URL: https://github.com/apache/tvm/pull/15818#discussion_r1339863992
##
apps/cpp_rtvm/main.cc:
##
@@ -350,43 +350,46 @@ int ExecuteModel(ToolArgs& args) {
} else if (!args.input.empty() && !args.output.empty()) {
LOG(INFO) <<
srkreddy1238 commented on code in PR #15818:
URL: https://github.com/apache/tvm/pull/15818#discussion_r1339876548
##
apps/cpp_rtvm/main.cc:
##
@@ -350,43 +350,46 @@ int ExecuteModel(ToolArgs& args) {
} else if (!args.input.empty() && !args.output.empty()) {
LOG(INFO) <<
srkreddy1238 commented on PR #15818:
URL: https://github.com/apache/tvm/pull/15818#issuecomment-1738870461
@echuraev There are some requirements and concerns on high load and unload
timings with TVM in general. With PreCompiled bin we could reduce the majority
and there exist some more
echuraev commented on code in PR #15818:
URL: https://github.com/apache/tvm/pull/15818#discussion_r1339925369
##
apps/cpp_rtvm/main.cc:
##
@@ -350,43 +350,46 @@ int ExecuteModel(ToolArgs& args) {
} else if (!args.input.empty() && !args.output.empty()) {
LOG(INFO) <<
srkreddy1238 opened a new pull request, #15838:
URL: https://github.com/apache/tvm/pull/15838
Handle unnecessary pylint errors from these frontends
--
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
echuraev merged PR #15830:
URL: https://github.com/apache/tvm/pull/15830
--
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:
This is an automated email from the ASF dual-hosted git repository.
echuraev 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 9d8e6fda50 [ADRENO] Minor changes for Adreno docs
34 matches
Mail list logo
