[GitHub] [incubator-tvm] wrongtest commented on pull request #5751: [RUNTIME] Add compile_shared option to linux compile utility fn

2020-06-10 Thread GitBox
wrongtest commented on pull request #5751: URL: https://github.com/apache/incubator-tvm/pull/5751#issuecomment-642423655 There are two python unittest case aims to create executable binary but finally run into `cc.create_shared()` code path. We add a new utility fn

[GitHub] [incubator-tvm] t-vi commented on pull request #5763: Fix gelu in PyTorch frontend, tighten numerical checks

2020-06-10 Thread GitBox
t-vi commented on pull request #5763: URL: https://github.com/apache/incubator-tvm/pull/5763#issuecomment-642417669 Yeah, apparently, the end-to-end checks aren't completely deterministic, I bumped them all to 1e-4 tolerance.

[GitHub] [incubator-tvm] tqchen closed issue #5771: How to test wasm32 target on Rust runtime binding

2020-06-10 Thread GitBox
tqchen closed issue #5771: URL: https://github.com/apache/incubator-tvm/issues/5771 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

[GitHub] [incubator-tvm] tqchen commented on issue #5771: How to test wasm32 target on Rust runtime binding

2020-06-10 Thread GitBox
tqchen commented on issue #5771: URL: https://github.com/apache/incubator-tvm/issues/5771#issuecomment-642392302 Thanks @leonwanghui for bringing up the problem, the community uses https://discuss.tvm.ai/ for general discussions, please bring a new thread there https://discuss.tvm.ai/

[GitHub] [incubator-tvm] tqchen opened a new pull request #5773: [TIR][REFACTOR][API-Change] Migrate the tvm/tir/expr.h to constructor

2020-06-10 Thread GitBox
tqchen opened a new pull request #5773: URL: https://github.com/apache/incubator-tvm/pull/5773 This PR migrate tvm/tir/expr.h to the new constructor style that is consistent with the rest of the codebase and changes the affected files accordingly.

[incubator-tvm] branch master updated (ed58309 -> c7274fd)

2020-06-10 Thread kevinthesun
This is an automated email from the ASF dual-hosted git repository. kevinthesun pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from ed58309 [topi] block sparse dense on cuda (#5746) add c7274fd [Relay] Fix for recursive let

[GitHub] [incubator-tvm] kevinthesun commented on pull request #5757: [Relay] Fix for recursive let

2020-06-10 Thread GitBox
kevinthesun commented on pull request #5757: URL: https://github.com/apache/incubator-tvm/pull/5757#issuecomment-642389407 Thanks @icemelon9 @jroesch @lixiaoquan This is an automated message from the Apache Git Service. To

[GitHub] [incubator-tvm] kevinthesun merged pull request #5757: [Relay] Fix for recursive let

2020-06-10 Thread GitBox
kevinthesun merged pull request #5757: URL: https://github.com/apache/incubator-tvm/pull/5757 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

[GitHub] [incubator-tvm] xqdan opened a new pull request #5772: [ARITH]add simplify rule for div

2020-06-10 Thread GitBox
xqdan opened a new pull request #5772: URL: https://github.com/apache/incubator-tvm/pull/5772 Thanks for contributing to TVM! Please refer to guideline https://tvm.apache.org/docs/contribute/ for useful information and tips. After the pull request is submitted, please request code

[GitHub] [incubator-tvm-vta] tmoreau89 commented on pull request #8: [Hardware][Xilinx] explicitly specify acc dep distance to avoid hidden pitfall

2020-06-10 Thread GitBox
tmoreau89 commented on pull request #8: URL: https://github.com/apache/incubator-tvm-vta/pull/8#issuecomment-642387343 @zhanghaohit I think that the warning approach is a reasonable compromise between user-defined flexibility and making sure that they set the value correctly. To

[GitHub] [incubator-tvm] leonwanghui closed pull request #5733: Make Tensor struct public in Rust runtime

2020-06-10 Thread GitBox
leonwanghui closed pull request #5733: URL: https://github.com/apache/incubator-tvm/pull/5733 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

[GitHub] [incubator-tvm] leonwanghui commented on pull request #5733: Make Tensor struct public in Rust runtime

2020-06-10 Thread GitBox
leonwanghui commented on pull request #5733: URL: https://github.com/apache/incubator-tvm/pull/5733#issuecomment-642384629 Already fixed by transferring `Tensor` to `DLTensor` ABI. This is an automated message from the

[GitHub] [incubator-tvm] leonwanghui opened a new issue #5771: How to test wasm32 target on Rust runtime binding

2020-06-10 Thread GitBox
leonwanghui opened a new issue #5771: URL: https://github.com/apache/incubator-tvm/issues/5771 Hi @tqchen @jroesch , I was working for embedding Rust runtime for WASM target, but came across some problems when building test-wasm32 crate. ## Error elaboration Firstly I need to

[GitHub] [incubator-tvm] zhiics commented on a change in pull request #5699: [Frontend][TensorFlow] Improve Control Flow and TensorArray

2020-06-10 Thread GitBox
zhiics commented on a change in pull request #5699: URL: https://github.com/apache/incubator-tvm/pull/5699#discussion_r438517544 ## File path: python/tvm/relay/frontend/tensorflow.py ## @@ -2444,6 +2470,27 @@ def _in_while_loop(control_flow_node_map, op_name): return

[GitHub] [incubator-tvm] FrozenGene edited a comment on pull request #5754: [RFC] Improve quantized convolution performance for armv8 architectures

2020-06-10 Thread GitBox
FrozenGene edited a comment on pull request #5754: URL: https://github.com/apache/incubator-tvm/pull/5754#issuecomment-642374967 Thanks for the great work! I have some quick question: 1. Have you tested various models arm cpu? (like A53, A72, A55, A75 and so on). According to fb qnnpack

[GitHub] [incubator-tvm] FrozenGene commented on pull request #5754: [RFC] Improve quantized convolution performance for armv8 architectures

2020-06-10 Thread GitBox
FrozenGene commented on pull request #5754: URL: https://github.com/apache/incubator-tvm/pull/5754#issuecomment-642375802 cc @ajtulloch This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [incubator-tvm] FrozenGene commented on pull request #5754: [RFC] Improve quantized convolution performance for armv8 architectures

2020-06-10 Thread GitBox
FrozenGene commented on pull request #5754: URL: https://github.com/apache/incubator-tvm/pull/5754#issuecomment-642374967 Thanks for the great work! I have some quick question: 1. Have you tested various models arm cpu? (like A53, A72, A55, A75 and so on). According to fb qnnpack blog,

[GitHub] [incubator-tvm] masahi commented on a change in pull request #5737: [TOPI][RELAY][PYTORCH]Conv3d_transpose op support added

2020-06-10 Thread GitBox
masahi commented on a change in pull request #5737: URL: https://github.com/apache/incubator-tvm/pull/5737#discussion_r438492454 ## File path: python/tvm/relay/op/strategy/generic.py ## @@ -345,6 +345,46 @@ def conv2d_transpose_strategy(attrs, inputs, out_type, target):

[GitHub] [incubator-tvm] masahi commented on a change in pull request #5737: [TOPI][RELAY][PYTORCH]Conv3d_transpose op support added

2020-06-10 Thread GitBox
masahi commented on a change in pull request #5737: URL: https://github.com/apache/incubator-tvm/pull/5737#discussion_r438492376 ## File path: python/tvm/relay/op/strategy/generic.py ## @@ -345,6 +345,46 @@ def conv2d_transpose_strategy(attrs, inputs, out_type, target):

[GitHub] [incubator-tvm] leonwanghui commented on pull request #5527: [Rust] Second stage of Rust Refactor

2020-06-10 Thread GitBox
leonwanghui commented on pull request #5527: URL: https://github.com/apache/incubator-tvm/pull/5527#issuecomment-642349670 @jroesch Got it, thanks a lot for the elaboration : ) This is an automated message from the Apache

[GitHub] [incubator-tvm] masahi commented on pull request #5763: Fix gelu in PyTorch frontend, tighten numerical checks

2020-06-10 Thread GitBox
masahi commented on pull request #5763: URL: https://github.com/apache/incubator-tvm/pull/5763#issuecomment-642343462 CI is still not happy. Could be related to CUDA issues. This is an automated message from the Apache Git

[GitHub] [incubator-tvm-vta] tmoreau89 commented on pull request #5: Add c++ and python local deploy example

2020-06-10 Thread GitBox
tmoreau89 commented on pull request #5: URL: https://github.com/apache/incubator-tvm-vta/pull/5#issuecomment-642342815 Thanks so much @huajsj the PR has been approved and merged. This is an automated message from the Apache

[incubator-tvm-vta] branch master updated: Add c++ and python local deploy example (#5)

2020-06-10 Thread moreau
This is an automated email from the ASF dual-hosted git repository. moreau pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm-vta.git The following commit(s) were added to refs/heads/master by this push: new f1c3385 Add c++ and python local

[GitHub] [incubator-tvm-vta] tmoreau89 merged pull request #5: Add c++ and python local deploy example

2020-06-10 Thread GitBox
tmoreau89 merged pull request #5: URL: https://github.com/apache/incubator-tvm-vta/pull/5 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

[GitHub] [incubator-tvm] icemelon9 commented on pull request #5757: [Relay] Fix for recursive let

2020-06-10 Thread GitBox
icemelon9 commented on pull request #5757: URL: https://github.com/apache/incubator-tvm/pull/5757#issuecomment-642335433 @lixiaoquan Yes, that's possible. I could take a look as well. This is an automated message from the

[GitHub] [incubator-tvm] zhiics opened a new pull request #5770: [BYOC][runtime] Separate code and metadata for CSourceModule

2020-06-10 Thread GitBox
zhiics opened a new pull request #5770: URL: https://github.com/apache/incubator-tvm/pull/5770 This PR attempts to separate metadata initialization and source code compilation (when necessary). The discussion can be found here:

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tqchen commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438463038 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tqchen commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438462908 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] t-vi commented on pull request #5763: Fix gelu in PyTorch frontend, tighten numerical checks

2020-06-10 Thread GitBox
t-vi commented on pull request #5763: URL: https://github.com/apache/incubator-tvm/pull/5763#issuecomment-642311242 Two of the end-to-end networks seemed to have deviations slightly above 1e-5. I moved those to 1e-4 tolerance.

[GitHub] [incubator-tvm] masahi edited a comment on pull request #5755: Edit onnx parser to infer values in post order

2020-06-10 Thread GitBox
masahi edited a comment on pull request #5755: URL: https://github.com/apache/incubator-tvm/pull/5755#issuecomment-642300115 Correct me if I'm wrong, does this change make `infer_value` on down stream nodes faster because the arguments are constant (computed bottom up) rather than another

[GitHub] [incubator-tvm] masahi commented on pull request #5755: Edit onnx parser to infer values in post order

2020-06-10 Thread GitBox
masahi commented on pull request #5755: URL: https://github.com/apache/incubator-tvm/pull/5755#issuecomment-642300115 Correct me if I'm wrong, does this change mean `infer_value` on down stream nodes should be faster because the arguments are constant (computed bottom up) rather than

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438438414 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438437559 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438437168 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438437454 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438436184 ## File path: docs/conf.py ## @@ -203,7 +203,8 @@ '../tutorials/deployment', '../vta/tutorials/frontend',

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438436066 ## File path: tutorials/micro/README.txt ## @@ -0,0 +1,4 @@ +.. _tutorial-micro: Review comment: Removed

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438434494 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438434392 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #5601: [DataType] Add bfloat16

2020-06-10 Thread GitBox
tqchen commented on a change in pull request #5601: URL: https://github.com/apache/incubator-tvm/pull/5601#discussion_r438430324 ## File path: src/tir/transforms/bf16_legalize.cc ## @@ -0,0 +1,387 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or

[GitHub] [incubator-tvm] anijain2305 commented on issue #5458: [CI] Fix gluoncv tutorial under mxnet-mkl

2020-06-10 Thread GitBox
anijain2305 commented on issue #5458: URL: https://github.com/apache/incubator-tvm/issues/5458#issuecomment-642287159 Ping @icemelon9 If you get time. This is an automated message from the Apache Git Service. To respond to

[GitHub] [incubator-tvm] junrushao1994 edited a comment on pull request #5740: [Object][Runtime] Introduce runtime::Map

2020-06-10 Thread GitBox
junrushao1994 edited a comment on pull request #5740: URL: https://github.com/apache/incubator-tvm/pull/5740#issuecomment-642286757 Some actionable items per offline discussion with @tqchen: A1. We move things back to include/tvm/node, instead of runtime. This is mainly due to the

[GitHub] [incubator-tvm] junrushao1994 commented on pull request #5740: [Object][Runtime] Introduce runtime::Map

2020-06-10 Thread GitBox
junrushao1994 commented on pull request #5740: URL: https://github.com/apache/incubator-tvm/pull/5740#issuecomment-642286757 Some actionable items per offline discussion with @tqchen: A1. We move things back to include/tvm/node, instead of runtime. This is mainly due to the concern

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #5740: [Object][Runtime] Introduce runtime::Map

2020-06-10 Thread GitBox
junrushao1994 commented on a change in pull request #5740: URL: https://github.com/apache/incubator-tvm/pull/5740#discussion_r438423203 ## File path: src/ir/expr.cc ## @@ -171,17 +171,17 @@ TVM_STATIC_IR_FUNCTOR(ReprPrinter, vtable) .set_dispatch([](const ObjectRef& node,

[GitHub] [incubator-tvm] jwfromm commented on pull request #5755: Edit onnx parser to infer values in post order

2020-06-10 Thread GitBox
jwfromm commented on pull request #5755: URL: https://github.com/apache/incubator-tvm/pull/5755#issuecomment-642273871 I'm definitely in favor of this change, some of the models that take a long time to import are very unpleasant to deal with and taking a small hit on much faster

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438405649 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] tqchen commented on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
tqchen commented on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642241286 Thanks @t-vi I do not disagree, that is why we felt perhaps we should go with float in the first place, how about also open a dicussion thread about this topic?

[GitHub] [incubator-tvm] tom-gall commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
tom-gall commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438382565 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] jroesch closed pull request #5764: [WIP] `tvm` crate stage 3 of Rust refactor

2020-06-10 Thread GitBox
jroesch closed pull request #5764: URL: https://github.com/apache/incubator-tvm/pull/5764 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

[GitHub] [incubator-tvm] jroesch opened a new pull request #5769: [WIP] `tvm` crate stage 3 of Rust refactor

2020-06-10 Thread GitBox
jroesch opened a new pull request #5769: URL: https://github.com/apache/incubator-tvm/pull/5769 See #5764 I made a mistake with the remotes. This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-tvm] jroesch commented on pull request #5764: [WIP] `tvm` crate stage 3 of Rust refactor

2020-06-10 Thread GitBox
jroesch commented on pull request #5764: URL: https://github.com/apache/incubator-tvm/pull/5764#issuecomment-642226801 I made a mistake with remotes, going to reopn. This is an automated message from the Apache Git Service.

[GitHub] [incubator-tvm] tqchen commented on pull request #5748: [REFACTOR][Relay] Migrate Id ObjectRef to not-null

2020-06-10 Thread GitBox
tqchen commented on pull request #5748: URL: https://github.com/apache/incubator-tvm/pull/5748#issuecomment-642226324 @junrushao1994 I will try to send a proposal around UnsafeNull This is an automated message from the

[GitHub] [incubator-tvm] t-vi edited a comment on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
t-vi edited a comment on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-64584 I'll probably send fixes for the PyTorch frontend, I don't think I'm qualified to fix it throughout the codebase, though. I might say that I have the feeling

[GitHub] [incubator-tvm] t-vi commented on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
t-vi commented on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-64584 I'll probably send fixes for the PyTorch frontend, I don't think I'm qualified to fix it throughout the codebase, though.

[GitHub] [incubator-tvm] t-vi closed pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
t-vi closed pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756 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

[GitHub] [incubator-tvm] tqchen commented on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
tqchen commented on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642218720 How about we first fix the usage of "float" throughout the codebase, then we can send a PR to warn about "float" as dtype in the parser.

[GitHub] [incubator-tvm] notoraptor commented on pull request #5716: [topi][relay] Add operation gather to relay.

2020-06-10 Thread GitBox
notoraptor commented on pull request #5716: URL: https://github.com/apache/incubator-tvm/pull/5716#issuecomment-642209115 Rebased to fix conflicts. @tqchen PR is approved by @abergeron ! This is an automated message from the

[GitHub] [incubator-tvm] mbrookhart commented on a change in pull request #5766: [BYOC][FIX] Infer types in MergeComposite

2020-06-10 Thread GitBox
mbrookhart commented on a change in pull request #5766: URL: https://github.com/apache/incubator-tvm/pull/5766#discussion_r438355902 ## File path: src/relay/transforms/merge_composite.cc ## @@ -36,17 +36,24 @@ namespace tvm { namespace relay { namespace merge_composite {

[GitHub] [incubator-tvm] junrushao1994 commented on pull request #5748: [REFACTOR][Relay] Migrate Id ObjectRef to not-null

2020-06-10 Thread GitBox
junrushao1994 commented on pull request #5748: URL: https://github.com/apache/incubator-tvm/pull/5748#issuecomment-642203863 Do we have conclusions yet on what the best way is to support not-null with composite objects?

[GitHub] [incubator-tvm] jroesch commented on a change in pull request #5764: [WIP] `tvm` crate stage 3 of Rust refactor

2020-06-10 Thread GitBox
jroesch commented on a change in pull request #5764: URL: https://github.com/apache/incubator-tvm/pull/5764#discussion_r438348392 ## File path: rust/tvm/src/ir/array.rs ## @@ -0,0 +1,74 @@ +use std::convert::TryFrom; +use std::marker::PhantomData; + +use

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #5766: [BYOC][FIX] Infer types in MergeComposite

2020-06-10 Thread GitBox
mbaret commented on a change in pull request #5766: URL: https://github.com/apache/incubator-tvm/pull/5766#discussion_r438334823 ## File path: src/relay/transforms/merge_composite.cc ## @@ -36,17 +36,24 @@ namespace tvm { namespace relay { namespace merge_composite {

[GitHub] [incubator-tvm] zhiics commented on a change in pull request #5766: [BYOC][FIX] Infer types in MergeComposite

2020-06-10 Thread GitBox
zhiics commented on a change in pull request #5766: URL: https://github.com/apache/incubator-tvm/pull/5766#discussion_r438322794 ## File path: src/relay/transforms/merge_composite.cc ## @@ -36,17 +36,24 @@ namespace tvm { namespace relay { namespace merge_composite {

[GitHub] [incubator-tvm] mbrookhart commented on pull request #5766: [BYOC][FIX] Infer types in MergeComposite

2020-06-10 Thread GitBox
mbrookhart commented on pull request #5766: URL: https://github.com/apache/incubator-tvm/pull/5766#issuecomment-642173517 @comaniac Match won't traverse the input expression, it will only compare the function to the exact expression you give. Partition and Rewrite, however, traverse the

[GitHub] [incubator-tvm] tqchen commented on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
tqchen commented on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642170343 Aha, that is definitely interesting. It is still worth to have a discussion about the numpy compatibility vs favors common deep learning land. That could leads to the

[GitHub] [incubator-tvm] mwillsey commented on a change in pull request #5764: [WIP] `tvm` crate stage 3 of Rust refactor

2020-06-10 Thread GitBox
mwillsey commented on a change in pull request #5764: URL: https://github.com/apache/incubator-tvm/pull/5764#discussion_r438302423 ## File path: rust/runtime/tests/test_wasm32/Cargo.toml ## @@ -22,5 +22,6 @@ license = "Apache-2.0" authors = ["TVM Contributors"]

[GitHub] [incubator-tvm] kevinthesun commented on a change in pull request #5762: [TF] Support symbolic inputs of Fill

2020-06-10 Thread GitBox
kevinthesun commented on a change in pull request #5762: URL: https://github.com/apache/incubator-tvm/pull/5762#discussion_r438299882 ## File path: python/tvm/relay/frontend/tensorflow.py ## @@ -1352,12 +1352,17 @@ def _impl(inputs, attr, params, mod): # Output shape

[GitHub] [incubator-tvm] icemelon9 commented on pull request #5600: [TOPI] Improve CUDA softmax scheduling

2020-06-10 Thread GitBox
icemelon9 commented on pull request #5600: URL: https://github.com/apache/incubator-tvm/pull/5600#issuecomment-642150126 That's great. I'll take a look at it. This is an automated message from the Apache Git Service. To

[GitHub] [incubator-tvm] vinx13 commented on pull request #5746: [topi] block sparse dense on cuda

2020-06-10 Thread GitBox
vinx13 commented on pull request #5746: URL: https://github.com/apache/incubator-tvm/pull/5746#issuecomment-642142044 thanks @ceruleangu @antinucleon @junrushao1994 This is an automated message from the Apache Git Service.

[incubator-tvm] branch master updated: [topi] block sparse dense on cuda (#5746)

2020-06-10 Thread wuwei
This is an automated email from the ASF dual-hosted git repository. wuwei pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git The following commit(s) were added to refs/heads/master by this push: new ed58309 [topi] block sparse dense on

[GitHub] [incubator-tvm] vinx13 merged pull request #5746: [topi] block sparse dense on cuda

2020-06-10 Thread GitBox
vinx13 merged pull request #5746: URL: https://github.com/apache/incubator-tvm/pull/5746 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

[GitHub] [incubator-tvm] t-vi edited a comment on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
t-vi edited a comment on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642134657 @tqchen The thing I'm finding awkward to use is ```python >>> tvm.relay.expr.const(1, dtype="float").data.dtype

[GitHub] [incubator-tvm] areusch commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
areusch commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r438270610 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] t-vi edited a comment on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
t-vi edited a comment on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642134657 @tqchen The thing I'm finding awkward to use is ```python >>> tvm.relay.expr.const(1, dtype="float").data.dtype

[GitHub] [incubator-tvm] t-vi edited a comment on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
t-vi edited a comment on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642134657 @tqchen The thing I'm finding awkward to use is ```python >>> tvm.relay.expr.const(1, dtype="float").data.dtype

[GitHub] [incubator-tvm] t-vi commented on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
t-vi commented on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642134657 @tqchen The thing I'm finding awkward to use is ``` tvm.relay.expr.const(1, dtype="float").data.dtype

[GitHub] [incubator-tvm] seanlatias opened a new pull request #5768: [Relay] Fix Calibration Pass to Support Modules with Multiple Functions

2020-06-10 Thread GitBox
seanlatias opened a new pull request #5768: URL: https://github.com/apache/incubator-tvm/pull/5768 With the existing calibration pass, it only returns the module that is built from the main function. However, the input can have more than one functions (e.g., after using BYOC). This PR

[GitHub] [incubator-tvm] comaniac commented on a change in pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
comaniac commented on a change in pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#discussion_r438263324 ## File path: python/tvm/relay/expr.py ## @@ -487,6 +487,8 @@ def const(value, dtype=None): - bool maps to "bool" - other using the same

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #5766: [BYOC][FIX] Infer types in MergeComposite

2020-06-10 Thread GitBox
mbaret commented on a change in pull request #5766: URL: https://github.com/apache/incubator-tvm/pull/5766#discussion_r438261289 ## File path: tests/python/relay/test_pass_merge_composite.py ## @@ -953,26 +985,28 @@ def _check_type_false(extract): return

[GitHub] [incubator-tvm] mbrookhart commented on a change in pull request #5766: [BYOC][FIX] Infer types in MergeComposite

2020-06-10 Thread GitBox
mbrookhart commented on a change in pull request #5766: URL: https://github.com/apache/incubator-tvm/pull/5766#discussion_r438258138 ## File path: tests/python/relay/test_pass_merge_composite.py ## @@ -953,26 +985,28 @@ def _check_type_false(extract): return

[GitHub] [incubator-tvm] tqchen edited a comment on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
tqchen edited a comment on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642115243 Previously we use this approach. Personally I was in favor of default to fp32. However, one reason for us to convert to the current way is to just make sure

[GitHub] [incubator-tvm] tqchen commented on pull request #5756: make dtype="float" mean float32 for relay.expr.const

2020-06-10 Thread GitBox
tqchen commented on pull request #5756: URL: https://github.com/apache/incubator-tvm/pull/5756#issuecomment-642115243 Previously we use this approach. Personally I was in favor of default to fp32. However, one reason for us to convert to the current way is to just make sure that

[GitHub] [incubator-tvm] siju-samuel commented on issue #5765: Feature-request: State-of-art Yolo v4 Detector

2020-06-10 Thread GitBox
siju-samuel commented on issue #5765: URL: https://github.com/apache/incubator-tvm/issues/5765#issuecomment-642114545 @AlexeyAB `load_network_custom` with `batch=1` memory is ok. I will do the rest. I think the `load_network` api is allocating memory for training as well. Thanks.

[GitHub] [incubator-tvm] AlexeyAB commented on issue #5765: Feature-request: State-of-art Yolo v4 Detector

2020-06-10 Thread GitBox
AlexeyAB commented on issue #5765: URL: https://github.com/apache/incubator-tvm/issues/5765#issuecomment-642102804 @siju-samuel Try to use `network *load_network_custom(char *cfg, char *weights, int clear, int batch)` And set `batch=1`.

[GitHub] [incubator-tvm] abergeron commented on pull request #5758: Fix the meaning of conv{1,2}d_transpose output_padding parameter.

2020-06-10 Thread GitBox
abergeron commented on pull request #5758: URL: https://github.com/apache/incubator-tvm/pull/5758#issuecomment-642101273 I've confirmed manually that with the tophub changes the vta tests pass. This is an automated message

[GitHub] [incubator-tvm] mbrookhart commented on pull request #5760: [PatternLang] Add ShapePattern and DataTypePattern

2020-06-10 Thread GitBox
mbrookhart commented on pull request #5760: URL: https://github.com/apache/incubator-tvm/pull/5760#issuecomment-642092107 @jroesch The problem comes in handling the matcher in the backend. Since Types belong to a lower level than the pattern language, we can't really embed something like

[GitHub] [incubator-tvm] siju-samuel opened a new pull request #5767: [TENSORFLOW]Sparse2Dense support

2020-06-10 Thread GitBox
siju-samuel opened a new pull request #5767: URL: https://github.com/apache/incubator-tvm/pull/5767 Sparse2Dense support added to tf frontend. @masahi @srkreddy1238 Please help to review and merge this PR. TIA This is

[GitHub] [incubator-tvm] siju-samuel edited a comment on issue #5765: Feature-request: State-of-art Yolo v4 Detector

2020-06-10 Thread GitBox
siju-samuel edited a comment on issue #5765: URL: https://github.com/apache/incubator-tvm/issues/5765#issuecomment-642084669 @AlexeyAB Thanks for the YoloV4 paper. the output is impressive. Few weeks back I already tried integrating this to tvm. In tvm we use `network

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #5740: [Object][Runtime] Introduce runtime::Map

2020-06-10 Thread GitBox
tqchen commented on a change in pull request #5740: URL: https://github.com/apache/incubator-tvm/pull/5740#discussion_r438213646 ## File path: include/tvm/runtime/container.h ## @@ -1554,6 +1604,925 @@ struct PackedFuncValueConverter> { } }; +/*! \brief map node content

[GitHub] [incubator-tvm] siju-samuel commented on issue #5765: Feature-request: State-of-art Yolo v4 Detector

2020-06-10 Thread GitBox
siju-samuel commented on issue #5765: URL: https://github.com/apache/incubator-tvm/issues/5765#issuecomment-642084669 @AlexeyAB Thanks for the YoloV4 paper. the output is impressive. Few weeks back I already tried integrating this to tvm. In tvm we use `network *load_network(char

[GitHub] [incubator-tvm] mbaret opened a new pull request #5766: [BYOC][FIX] Infer types in MergeComposite

2020-06-10 Thread GitBox
mbaret opened a new pull request #5766: URL: https://github.com/apache/incubator-tvm/pull/5766 If InferType isn't run between partitioning passes, function calls are inserted which don't have a type. This can result in failures for patterns which want to check types. This works

[GitHub] [incubator-tvm] mbaret commented on pull request #5766: [BYOC][FIX] Infer types in MergeComposite

2020-06-10 Thread GitBox
mbaret commented on pull request #5766: URL: https://github.com/apache/incubator-tvm/pull/5766#issuecomment-642072780 cc @mbrookhart @comaniac This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-tvm] AlexeyAB opened a new issue #5765: Feature-request: State-of-art Yolo v4 Detector

2020-06-10 Thread GitBox
AlexeyAB opened a new issue #5765: URL: https://github.com/apache/incubator-tvm/issues/5765 Feature-request: State-of-art **Yolo v4** Detector * paper: https://arxiv.org/abs/2004.10934 * source code: https://github.com/AlexeyAB/darknet * cfg-file Yolov4:

[GitHub] [incubator-tvm] lixiaoquan commented on pull request #5757: [Relay] Fix for recursive let

2020-06-10 Thread GitBox
lixiaoquan commented on pull request #5757: URL: https://github.com/apache/incubator-tvm/pull/5757#issuecomment-641943041 Is it possbile to happen in other passes? I run into a stack overflow in WellFormChecker, and it keeps visiting LetNode. Not sure whether it is the same issue.

[GitHub] [incubator-tvm] liangfu commented on a change in pull request #5655: Add MicroTVM tutorial using the STM32F746 discovery board

2020-06-10 Thread GitBox
liangfu commented on a change in pull request #5655: URL: https://github.com/apache/incubator-tvm/pull/5655#discussion_r437896159 ## File path: tutorials/micro/micro_tflite.py ## @@ -0,0 +1,219 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [incubator-tvm] t-vi commented on a change in pull request #5752: Make batch matrix multiplication on GPU tunable

2020-06-10 Thread GitBox
t-vi commented on a change in pull request #5752: URL: https://github.com/apache/incubator-tvm/pull/5752#discussion_r437993129 ## File path: topi/python/topi/cuda/batch_matmul.py ## @@ -51,55 +62,73 @@ def _schedule(op): C = s.outputs[0].output(0) b, y,

[GitHub] [incubator-tvm] t-vi commented on a change in pull request #5752: Make batch matrix multiplication on GPU tunable

2020-06-10 Thread GitBox
t-vi commented on a change in pull request #5752: URL: https://github.com/apache/incubator-tvm/pull/5752#discussion_r437993129 ## File path: topi/python/topi/cuda/batch_matmul.py ## @@ -51,55 +62,73 @@ def _schedule(op): C = s.outputs[0].output(0) b, y,

[GitHub] [incubator-tvm] jroesch opened a new pull request #5764: [WIP] `tvm` crate stage 3 of Rust refactor

2020-06-10 Thread GitBox
jroesch opened a new pull request #5764: URL: https://github.com/apache/incubator-tvm/pull/5764 Currently WIP this adds compiler bindings to the TVM Rust bindings. This is an automated message from the Apache Git Service.

[incubator-tvm] branch rust-tvm updated (caa6904 -> 43325ab)

2020-06-10 Thread jroesch
This is an automated email from the ASF dual-hosted git repository. jroesch pushed a change to branch rust-tvm in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. discard caa6904 All tests pass discard 0c55c39 Finish removing anyhow and work with new external! macro

[incubator-tvm] 02/02: Add tvm crate

2020-06-10 Thread jroesch
This is an automated email from the ASF dual-hosted git repository. jroesch pushed a commit to branch rust-tvm in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git commit 43325abcdf9b1aa8d405a9769aa43dc12f66112d Author: Jared Roesch AuthorDate: Wed Jun 10 02:30:10 2020 -0700

[incubator-tvm] 01/02: Adapt to new macro

2020-06-10 Thread jroesch
This is an automated email from the ASF dual-hosted git repository. jroesch pushed a commit to branch rust-tvm in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git commit fcdfffa2adfce22244ed19d3234060d7f4968147 Author: Jared Roesch AuthorDate: Wed Jun 10 02:29:39 2020 -0700

  1   2   >