Nicklas Börjesson

Is this repository maintained?

1

I am not seeing any responses on issues or PR:s in a long while. And Python 3.12 have been out a year soon?

Fix Python 3.12 compatibility - use importlib.util.find_spec("numpy") instead of imp

3

Hi! In the pip distribution (at least on windows), in flatbuffers\compat.py, it uses imp, which is removed in python 3.12, so I had to monkey patch it to use importlib...

model.export to TFLite with int8 doesn't yield a fully int8 quantized model

14

### Search before asking - [X] I have searched the Ultralytics YOLO [issues](https://github.com/ultralytics/ultralytics/issues) and found no similar bug report. ### Ultralytics YOLO Component Export ### Bug In my view, this,...

Adjustments for ESP32-cam and ESP-IDF 5.3.0

2

Hi, Not sure if this will work for you, the PR perhaps does way too much, but it makes the hello- and benchmark examples work on ESP-IDF 5.3.0 and a...

Inputs format misunderstood. (prepare_input) (AIV-743)

12

### Checklist - [X] Checked the issue tracker for similar issues to ensure this is not a duplicate. - [X] Provided a clear description of your suggestion. - [X] Included...

Cannot create an esp-ppq installation where quantize_custom_torch_model.py works properly. (AIV-740)

4

### Checklist - [X] Checked the issue tracker for similar issues to ensure this is not a duplicate - [X] Read the documentation to confirm the issue is not addressed...

Make LCPAN also pass the activator to DepthwiseSeparable

6

Without this, nn.Hardswish() will be selected in ConvBNLayer, which causes the opset version to be bumped to 14, in turn, and in my case, breaks ESP-DL ONNX quantization, which only...