You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2021/01/05 05:18:00 UTC

[GitHub] [tvm] fantasyRqg opened a new pull request #7211: Build multi models into one system-lib

fantasyRqg opened a new pull request #7211:
URL: https://github.com/apache/tvm/pull/7211


   Deploy models on Android & iOS platform must use system-lib. Avoid usage of `dlopen`.  
   [bundle_deploy](https://github.com/apache/tvm/tree/main/apps/bundle_deploy) demonstrated how to deploy a model which build a model which target system-lib
   
   BUT we need more than one models
   
   #### Related Discussions:
   - [How to deploy two different tvm compiled model in c++ statically?](https://discuss.tvm.apache.org/t/how-to-deploy-two-different-tvm-compiled-model-in-c-statically/2492)  
   - [Combine multi SYSTEM-LIB module libs to one](https://discuss.tvm.apache.org/t/discuss-combine-multi-system-lib-module-libs-to-one/8724). The implementation steps are different from what I thought at the beginning
   
   cc @wweic @zhiics 
   
   Please forgive me for my poor English


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] fantasyRqg commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
fantasyRqg commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-771485029


   https://github.com/apache/tvm/blob/3635945e48d9d1a7d8e76df418f057a4a3b88dc4/src/relay/backend/compile_engine.cc#L803-L836
   
   `GetUniqueName` make sure function names will not collide
   
   `name_map_` will not be cleand even after `CompileEngine::Global()->Clear();`
   
   @ziyu-guo 


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] AlexanderSerov commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
AlexanderSerov commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-842951774


   Is this planned to be included in v0.8?


-- 
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] fantasyRqg commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
fantasyRqg commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-771490370


   Seems batch size is not handled by `relay.build` either.
   My implementation just follow `relay.build`


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] FrozenGene commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
FrozenGene commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-766760556


   @kazum Could you help to review it? As it is somehow related with iOS.


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] fantasyRqg commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
fantasyRqg commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-754468357


   ```log
   enabled targets: llvm -device=arm_cpu; llvm
   pytest marker:
   ================================================================== test session starts ==================================================================
   platform linux -- Python 3.8.0, pytest-6.1.1, py-1.9.0, pluggy-0.13.1 -- /usr/local/bin/python3
   cachedir: .pytest_cache
   rootdir: /root/tvm, configfile: pytest.ini
   collected 14 items
   
   tests/python/unittest/test_runtime_rpc.py::test_bigendian_rpc PASSED                                                                              [  7%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_simple PASSED                                                                                 [ 14%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_runtime_string PASSED                                                                         [ 21%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_array PASSED                                                                                  [ 28%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_large_array PASSED                                                                            [ 35%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_echo PASSED                                                                                   [ 42%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_file_exchange PASSED                                                                          [ 50%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_remote_module PASSED                                                                          [ 57%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_return_func PASSED                                                                            [ 64%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_session_constructor_args PASSED                                                               [ 71%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_return_ndarray PASSED                                                                         [ 78%]
   tests/python/unittest/test_runtime_rpc.py::test_local_func PASSED                                                                                 [ 85%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_tracker_register PASSED                                                                       [ 92%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_tracker_request PASSED                                                                        [100%]
   
   ================================================================== 14 passed in 7.20s ===================================================================
   ```
   
   `test_rpc_echo` passed on my ubuntu.  
   trigger ci again


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] ziyu-guo commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
ziyu-guo commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-768679644


   Quick question: how does this work with multiple TVM module libs compiled from same model, but with different batch size? Won't the names collide in the combined lib? 


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] fantasyRqg removed a comment on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
fantasyRqg removed a comment on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-754468357


   ```log
   enabled targets: llvm -device=arm_cpu; llvm
   pytest marker:
   ================================================================== test session starts ==================================================================
   platform linux -- Python 3.8.0, pytest-6.1.1, py-1.9.0, pluggy-0.13.1 -- /usr/local/bin/python3
   cachedir: .pytest_cache
   rootdir: /root/tvm, configfile: pytest.ini
   collected 14 items
   
   tests/python/unittest/test_runtime_rpc.py::test_bigendian_rpc PASSED                                                                              [  7%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_simple PASSED                                                                                 [ 14%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_runtime_string PASSED                                                                         [ 21%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_array PASSED                                                                                  [ 28%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_large_array PASSED                                                                            [ 35%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_echo PASSED                                                                                   [ 42%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_file_exchange PASSED                                                                          [ 50%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_remote_module PASSED                                                                          [ 57%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_return_func PASSED                                                                            [ 64%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_session_constructor_args PASSED                                                               [ 71%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_return_ndarray PASSED                                                                         [ 78%]
   tests/python/unittest/test_runtime_rpc.py::test_local_func PASSED                                                                                 [ 85%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_tracker_register PASSED                                                                       [ 92%]
   tests/python/unittest/test_runtime_rpc.py::test_rpc_tracker_request PASSED                                                                        [100%]
   
   ================================================================== 14 passed in 7.20s ===================================================================
   ```
   
   `test_rpc_echo` passed on my ubuntu.  
   trigger ci again


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] jroesch closed pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
jroesch closed pull request #7211:
URL: https://github.com/apache/tvm/pull/7211


   


-- 
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-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] FrozenGene commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
FrozenGene commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-766760556


   @kazum Could you help to review it? As it is somehow related with iOS.


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] fantasyRqg commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
fantasyRqg commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-771485029






----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



[GitHub] [tvm] jroesch commented on pull request #7211: Build multi models into one system-lib

Posted by GitBox <gi...@apache.org>.
jroesch commented on pull request #7211:
URL: https://github.com/apache/tvm/pull/7211#issuecomment-1016827415


   This PR appears to be out of date, please feel free to reopen it if this is not the case.
   
   As part of the new year we are attempting to triage the project's open pull requests to ensure that code which
   is ready for review and/or merging receives adequate attention.
   
   Thanks again for your contribution, and feel free to reach out to discuss these changes.


-- 
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-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org