Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/marty1885/scirknn
Convert and run scikit-learn MLPs on Rockchip NPU.
https://github.com/marty1885/scirknn
inference-acceleration npu rk3566 rk3588 rknn rknpu2 rockchip scikit-learn
Last synced: 17 days ago
JSON representation
Convert and run scikit-learn MLPs on Rockchip NPU.
- Host: GitHub
- URL: https://github.com/marty1885/scirknn
- Owner: marty1885
- License: mpl-2.0
- Created: 2023-07-12T09:57:29.000Z (over 1 year ago)
- Default Branch: master
- Last Pushed: 2023-07-13T11:00:42.000Z (over 1 year ago)
- Last Synced: 2024-10-16T13:18:14.667Z (2 months ago)
- Topics: inference-acceleration, npu, rk3566, rk3588, rknn, rknpu2, rockchip, scikit-learn
- Language: Python
- Homepage:
- Size: 34.2 KB
- Stars: 7
- Watchers: 4
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE.txt
Awesome Lists containing this project
README
# scirknn
Unholy but necessary code that converts scikit-learn's MLP classifer/regresser into RKNN2 files in order to run on Rockchip NPUs. Because `rknn-toolkit2` does not support the operations `sklearn-onnx` spits out. So you can put your busness logic on the edge (or other things you want to do). Also works around a ton of stuff that ONNX and rknn-toolkit2 expects the other party to solve.
For real, who is training models large enough to need a NPU on sklearn?
## How it works
* Grabs weights from the sklearn model
* Build ONNX graph with nodes that RKNN accepts
* Move operations RKNN does not support into Python (these are lite operations)
* Convert ONNX into RKNN
* Provides a lite Python wrapper with the same API as scikit-learn if you need it
### Example
```python
from sklearn.neural_network import MLPClassifier
import sklearn2rknn # Convers scikit models into RKNN
# obtain a MLP model
x = [[0., 0.], [1., 1.], [2., 2.], [3., 3.]]
y = [0, 1, 2, 3]
clf = MLPClassifier(solver='lbfgs', alpha=1e-5, hidden_layer_sizes=(32, 32), random_state=1)
clf.fit(x, y)
# Convert and save the model targeting the RK3588 chip. If successful, this step produces
# two files. `example.rknn` and `example.rknn.json`. The latter is metadata that is used
# by wrappers later
sklearn2rknn.convert(clf, "example.rknn", "rk3588", quantization=True, example_input=x)
```
Or, use the commandline.
```bash
python -m sklearn2rknn model.pkl model.rknn --quantization --example_input /path/to/data.npy
```
Now, copy `example.rknn` and `example.rknn.json` to your development board.
```python
import scirknn # Wrapper to provide easy to use API
model = scirknn.MLPClassifier("example.krnn")
pred = model.predict([0, 0])
print(pred) # [0]
```
## Install
* Install [rknn-toolkit2][rknn-toolkit-whl] from the official `.whl` file
* Install the remaining dependencies `pip install -r requirments`
* only needed for model conversion
[rknn-toolkit-whl]: https://github.com/rockchip-linux/rknn-toolkit2/tree/master/packages
On your dev board. Install [rknn-toolkit-lite][rknn-tookit-lite-whl] and put the correspending [librknnrt.so][rknnrt] in `/usr/lib`
[rknn-tookit-lite-whl]: https://github.com/rockchip-linux/rknn-toolkit2/tree/master/rknn_toolkit_lite2/packages
[rknnrt]: https://github.com/rockchip-linux/rknpu2/tree/master/runtime
## Running the examples
The examples are stored in the `examples` folder. Every example has 2 files. `train.py` which trains a MLP model and coverts it into RKNN format. By default the training scripts converts for RK3588. And `infer.py` which shold run on the terget device. To run them, invoke them as a module.
```bash
> python -m example.xor.train
```
This prduces `xor.rknn` and `xor.rknn.json` in the current working directory. Copy them to your target device. Then run the following command to infer
```bash
> ls xor.rknn*
xor.rknn xor.rknn.json
> python -m example.xor.infer
I RKNN: [14:45:42.681] RKNN Runtime Information: librknnrt version: 1.5.0 (e6fe0c678@2023-05-25T08:09:20)
...
[1]
```
The final line is the output!