Comments (10)
@zhouxinyu0723 Is there a way where we do not have to hardcode the following lines?
https://github.com/tum-ei-eda/mlonmcu/blob/less_pip_install_issue51/mlonmcu/setup/setup.py#L191-L239
Maybe something like this could be used:if config.name in requirements: ...I can hardly think of a method to do this without hardcode because the config structures are nested.
For example how to check whether autotuning is activated?
It is also possible that certain feature names can be shared by two different modules. So I think it will be safe to check both the module name and feature name for requirements group like "tvm-autotuning" "tflite-visualization"
Then let’s keep it hardcoded as-is until we have resolved #33 which removes the nested structure of the config.
from mlonmcu.
check the branch less_pip_install_issue51.
from mlonmcu.
The command is "mlonmcu setup -g".
from mlonmcu.
First some feedback, which is not related to your implementation:
It seems like, a lot of packages are required even if ther are not used just because they are included in some mlonmcu source files such i.e.:
tflite
flatbuffers
matplotlib
I will think about if we can do anything about this.
from mlonmcu.
In
https://github.com/tum-ei-eda/mlonmcu/blob/main/mlonmcu/cli/init.py#L64-L69
I initially wanted to create a MLonMCU-managed virtual environment for the user (optional), however I never bothered implementing it. With this new feature we could take care of this automatically by some degree.
from mlonmcu.
The following does not work as expected when installed into an python environment (e.g. using make install
):
mlonmcu/mlonmcu/setup/setup.py
Line 170 in eeb1616
Could you write the requirements_addition.txt
to the environment directory instead?
from mlonmcu.
@zhouxinyu0723 Is there a way where we do not have to hardcode the following lines?
https://github.com/tum-ei-eda/mlonmcu/blob/less_pip_install_issue51/mlonmcu/setup/setup.py#L191-L239
Maybe something like this could be used:
if config.name in requirements:
...
from mlonmcu.
After adressing these comments, feel free to open a PR
from mlonmcu.
The following does not work as expected when installed into an python environment (e.g. using
make install
):mlonmcu/mlonmcu/setup/setup.py
Line 170 in eeb1616
Could you write the
requirements_addition.txt
to the environment directory instead?
will check it.
from mlonmcu.
@zhouxinyu0723 Is there a way where we do not have to hardcode the following lines?
https://github.com/tum-ei-eda/mlonmcu/blob/less_pip_install_issue51/mlonmcu/setup/setup.py#L191-L239
Maybe something like this could be used:
if config.name in requirements: ...
I doubt whether it is necessary to do it without hardcode. The problem is that the config structures are nested.
For example how to check whether autotuning is activated?
It is also possible that certain feature names can be shared by two different modules. So I think it will be safe to check both the module name and feature name for requirement groups like "tvm-autotuning" "tflite-visualization".
Then this style of naming conventions like "tvm-autotuning" need to be distinguished from "tvm". Which means branching is needed.
from mlonmcu.
Related Issues (20)
- [SW] Allow custom (out-of-tree) target/toolchain definitions
- [ETISS] Only run get_metrics.py if tracing dynamic memory metrics.
- [Plugins] Provide Example on how to use plugin interface
- [SW] Eliminate all compiler warnings in MLIF
- Support New RVV Targets
- Support for multi-core / multi-hart / cluster targets
- Test 64-bit support in RISC-V targets
- [Infrastructure] Refactor toolchain support
- Support custom (user/feature-defined) stages
- FreeRTOS support
- Add support for non-ML benchmarks HOT 1
- Allow using MLonMCU as MicroTVM backend target
- [TVM][MicroTVM] Finish metascheduler integration
- [TVM][MicroTVM] Finish Autoscheduler integration
- [TVM][MicroTVM] support graph tuner
- Fix 3rd party platform (ESP-IDF and Zephyr)
- Postprocesses very slow due to Python’s GIL HOT 1
- Reduce RAM consumption during MLonMCU Flow
- Avoid Paramiko (SSH) warning HOT 1
- SSHTarget: Only connect once
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from mlonmcu.