edgeai-torchvision

edgeai-torchvision copied to clipboard

Accuracy drop in Centernet model QAT training

1

Hi, I got 4% accuracy drop in QAT trained centernet model wrapped up with Xnn. QuantTrainModule. I'd tried with other utilites functions as you mentioned like " xnn.utils.freeze_bn(model) and xnn.layers.freeze_quant_range(model)...

sathyapatel

### 🐛 Describe the bug I want to convert a pre-trained onnx file to a PyTorch model, but there is not function, named a onnx2pytorch_internal in this repository. In fact,...

123DreamBuilder

Dear authors @mathmanu @kumardesappan , when I try to train my model on edgeai-torchvision, but I got following error: `Traceback (most recent call last): File "train4.py", line 141, in args.func(config,...

GRE-EXAMINATION

### 📚 The doc issue I have test the QAT using the edgeai-torchvision/references/edgeailite/scripts/train_classification_main.py for float, quantize, Freezing BN, Freezing ranges training, the results show that it is upto 10x slower...

Jon-drugstore

### 🐛 Describe the bug I am trying to install this on BeagleBone AI-64 (TDA4VM SoC) running with Debian-11. On executing script `./setup.sh` gave failure with below output on console....

vbvchauthmal

Quantized Checkpoints have Floating-Point Weights

20

### 🐛 Describe the bug Hello, I'm using the QuantTrainModule to train a MobileNetV2 model (using the MobileNetV2 class in this repo), and the quantized checkpoints have 32-bit floating-point weights...

IsidoraR

### 🚀 The feature I trained yolov5 model. but the accuract drop a lot. I want to train it with QAT ### Motivation, pitch I trained yolov5 model. but the...

suyali

### 🐛 Describe the bug when compile torchvision by "bash setup.sh", get follow error ``` [1/14] c++ -MMD -MF /home/songzhu/river/project/apa_proj/edgeai-torchvision/build/temp.linux-x86_64-3.7/home/songzhu/river/project/apa_proj/edgeai-torchvision/torchvision/csrc/vision.o.d -pthread -B /home/songzhu/anaconda3/envs/quantization/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall...

rivercn

### 🐛 Describe the bug In the quantize_classification_exmple.py script, there is a warning that says 'quantized inference/test may fail as it is not yet supported in gpu'. However, the Quantization...

IsidoraR

Quantitative training official edgeai-yolov5 code does not converge

4

I refer to the official quantitative demo training edgeai-yolov5, but the training did not converge. Quantization training examples/quantization_example.py this demo can converge. Refer to quantization_example.py to modify the code of...

Serissa