python Code of Inferencing with Florence-2-large-ft onnx model

#6
by vishalkatheriya - opened

i am working in colab and i am new to onnx i want a python code to load and run Florence 2 onnx model. i tried but i can't..?

Did the issue of writing in Python code using ONNX got resolved or is it still persisting? @vishalkatheriya can you recommend me how to use those individual ONNX files of the Florence model ie encoder model, embed tokens model and decoder model?

Hi @Aurosampad , I used PyTorch to quantize the model for my use. I tested it with PyTorch dynamic quantization but haven't tested ONNX quantization yet.
https://pytorch.org/tutorials/recipes/recipes/dynamic_quantization.html

Sign up or log in to comment