Fixed a bug in the optimization process for arithmetic operations - Githubissues

PINTO0309 / onnx2tf

Self-Created Tools to convert ONNX files (NCHW) to TensorFlow/TFLite/Keras format (NHWC). The purpose of this tool is to solve the massive Transpose extrapolation problem in onnx-tensorflow (onnx-tf). I don't need a Star, but give me a pull request.

MIT License

708 stars 73 forks source link

Fixed a bug in the optimization process for arithmetic operations #686

Closed PINTO0309 closed 2 months ago

PINTO0309 commented 2 months ago

1. Content and background

Add, Sub
- https://huggingface.co/onnx-community/metric3d-vit-small/blob/main/onnx/model.onnx
- metric3d-vit-small.onnx
- Fixed a bug in the optimization process for arithmetic operations. The error of the final output was less than 1e-4.
- A bug in the optimization process for the y = (200 - x) - 200 operation caused an incorrect Sub merge operation to be performed.

2. Summary of corrections

3. Before/After (If there is an operating log that can be used as a reference)

4. Issue number (only if there is a related issue)

Model gives inaccurate results post conversion to tflite #685