Văn Thạnh comments

Results 6 comments of


                                            Văn Thạnh

NaN during training

the reason may come from your dataset and not from the worker as it only loads the data for yourself. Is it a small dataset?

NaN during training

> It is made of around 1200 images for training and around 200 for validation. it is really small

How to load my Yolo model locally

i got same issue. the model is correct and when i try with this [repositories](https://github.com/Hyuto/yolov5-tfjs) everything was fine. also, i have a model was exported before and it work fine...

model size does not change on export to fp16 and int8 export

in this project, onnx with fp16 or int8 does not support. to export onnx with fp16: ` import onnx from onnxconverter_common import float16 model = onnx.load("path/to/model_fp32.onnx") model_fp16 = float16.convert_float_to_float16(model) onnx.save(model_fp16,...

Unable to set Width/Height in percent

try set height as a string Height="@("50%")".

Blazor Web App WASM download performance issues

Hi. I have tried to create a new solution with net 9 and it work fine. but with net 10 (aslo new solution with the same config) browser did not...