I think mxnet still depends on it. @tqchen @junrushao1994
---
[Visit Topic](https://discuss.tvm.apache.org/t/should-we-delete-nnvm/12330/2)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
here](https://discuss.tvm.apache.
Halide people did something like that https://dl.acm.org/doi/10.1145/3428234
---
[Visit
Topic](https://discuss.tvm.apache.org/t/rfc-verification-and-possibly-program-synthesis-of-expression-rewriting-rules/12319/2)
to respond.
You are receiving this because you enabled mailing list mode.
You need to use the `MergeComposite` pass. You can search its usage in our repo.
---
[Visit Topic](https://discuss.tvm.apache.org/t/pattern-matching-pass/12249/2)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
here](http
You can take a look at how we integrate DNNL (aka oneDNN). I think you need to
link your `*.so` file to `libtvm.so` or `libtvm_runtime.so`. cc @comaniac
---
[Visit
Topic](https://discuss.tvm.apache.org/t/byoc-how-to-link-extern-so-files/12186/2)
to respond.
You are receiving this because
I wonder why TIR constants doesn't support fp16? Because of the need for
c-codegen? @manupa-arm
---
[Visit
Topic](https://discuss.tvm.apache.org/t/problem-with-fuseops-and-embedded-constants-in-tir/12165/3)
to respond.
You are receiving this because you enabled mailing list mode.
To uns
[quote="lemo, post:2, topic:11968"]
can someone point me to the code responsible for calculating kernel launch
parameters
[/quote]
This is done by each op separately, most of them written in python, for example
*
https://github.com/apache/tvm/blob/bef7bf9b2b326488bc2bcc039710b2723023aaa0/test
First, you shouldn't use `stackvm`. Use llvm for host codegen.
`unknown function call Op(tir.call_pure_extern)` means you have some
backend-specific intrinsic functions our codegen doesn't know about. You can
dump the intrinsic name.
---
[Visit
Topic](https://discuss.tvm.apache.org/t/byo
You shouldn't need to deal with strategy stuff for operators offloaded to BYOC.
---
[Visit
Topic](https://discuss.tvm.apache.org/t/byoc-codegen-for-customized-call-node-operator/11863/7)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these e
Yes, if you don't intend to compile your model with native tvm, you only need
relay op so can skip that steps.
---
[Visit
Topic](https://discuss.tvm.apache.org/t/byoc-codegen-for-customized-call-node-operator/11863/2)
to respond.
You are receiving this because you enabled mailing list mo
[quote="manupa-arm, post:6, topic:11362"]
In the partition_for_* function where the full IRModule is visible (along with
@main and external functions) you could actually mutate the constants within
external function and hoist them out of the external function prior to calling
the relay.build(…
[quote="comaniac, post:3, topic:11362"]
This mechanism is used for the case that a BYOC backend attempts to manage the
constant values with certain processes, such as layout transform
[/quote]
CUTLASS does seem to support specialized layouts for gemm / conv2d. If we want
to make use of them an
UPDATE: For the particular case I've been working with, replacing one
`is_constant()` in my pattern with `wildcard()` allowed me to avoid the need
for running constant folding before pattern matching. So for now, I'm unblocked.
But I still wonder if is realistic not to support Constant at all
The recently merged [CUTLASS BYOC](https://github.com/apache/tvm/pull/9261)
relies on C-codegen based BYOC infra to JIT generate and compile C++ template
classes.
Currently it doesn't support Constants embedded in an external function and
instead requires all weight and bias parameters etc t
Currently, VM `PooledAllocator` releases its memory only when the underlying
device fails to allocate more memory:
https://github.com/apache/tvm/blob/553778885388a9eff4d611e1022baecd75c69088/src/runtime/vm/pooled_allocator.h#L60-L65.
This causes a program crash when doing repeated inferences
@junrushao1994 If by "meta programming" you mean an ability to call python
function from script to generate other code, and embed the generated code into
the calling context, then YES, we absolutely need this! I think it's called
"splicing" or "unquote" in the literature.
The lack of such fe
A PR posted https://github.com/apache/tvm/pull/7425
---
[Visit
Topic](https://discuss.tvm.apache.org/t/rfc-add-while-loop-node-to-tir/9028/15)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
here](https://discuss.tvm.apa
Thanks, yeah I also remember `te.scan` only supports scanning along the first
axis. I think `te.extern` is good as a first step.
---
[Visit
Topic](https://discuss.tvm.apache.org/t/add-some-new-tensorflow-ops/8217/7) to
respond.
You are receiving this because you enabled mailing list mode
I'm curious, how did you implement cumsum? I also wanted cumsum op a while
back, but for me it was not clear how it can be implemented efficiently in TVM.
---
[Visit
Topic](https://discuss.tvm.apache.org/t/add-some-new-tensorflow-ops/8217/4) to
respond.
You are receiving this because you
Interestingly, compiling faster rcnn and mask rcnn from PyTorch, enabled by the
PR https://github.com/apache/incubator-tvm/pull/6449, takes less than 3 min on
my laptop. I wonder where the difference in compilation time between TF and
PyTorch comes from.
---
[Visit
Topic](https://discuss
Hi, I've just came across a model that requires support for ONNX CumSum op
https://github.com/onnx/onnx/blob/master/docs/Operators.md#CumSum. The model
comes from DETR object detection model
https://github.com/facebookresearch/detr. Since this model doesn't need ad hoc
object detection ops th
20 matches
Mail list logo
