christoph/wamr
1
Fork 0
Code Releases Activity

Update sample workload tensorflow (#2101)

Browse Source 
Auto download emsdk under core/deps and hack it, and fix multi-thread issues.
This commit is contained in:
Wenyong Huang
2023-04-03 14:56:46 +08:00
committed by GitHub
parent 5c201995f1
commit 24a7e5c1e6
2 changed files with 49 additions and 53 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
samples/workload/tensorflow/README.md
View File

@ -1,29 +1,19 @@
"tensorflow" sample introduction
==============
This sample demonstrates how to build [tensorflow](https://github.com/tensorflow/tensorflow) into WebAssembly with emsdk toolchain and run it with iwasm. Please first install [emsdk](https://github.com/emscripten-core/emsdk):
```bash
git clone https://github.com/emscripten-core/emsdk.git
cd emsdk
./emsdk install 2.0.26
./emsdk activate 2.0.26
```
And set up ensdk environment:
```bash
source emsdk_env.sh
```
Then run
This sample demonstrates how to build [tensorflow](https://github.com/tensorflow/tensorflow) into WebAssembly with emsdk toolchain and run it with iwasm.:
```bash
./build.sh
# for linux platform, or
./build.sh --sgx
# for linux-sgx platform or
./build.sh --threads
# for multi-thread execution (on linux platform)
# for multi-threading on linux platform
./build.sh --sgx
# for linux-sgx platform
```
to build tensorflow and run it with iwasm, which basically contains the following steps:
- clone emsdk under `<wamr_dir>/core/deps`, install and activate 2.0.26
- hack emcc to delete some objects in libc.a
- build tf-lite with emcc compiler
- build iwasm with pthread enable and include libiary under libc-emcc
- build iwasm with lib-pthread and libc-emcc enabled
- run benchmark model with iwasm:
--max-secs 300: means the max training time cost is 5 minutes, you can adjust it by yourself