You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I converted your onnx models to tf saved model and with the current tfjs converter to tfjs models and it works very well. here is my code (based a lot on yours): https://github.com/Jonathhhan/ofEmscriptenExamples/blob/main/emscriptenCLIP2/clip.js
This is how I converted the models:
tensorflowjs_converter --input_format tf_saved_model clip-text-vit-32-tf clip-text-vit-f16-tfjs --quantize_float16
tensorflowjs_converter --input_format tf_saved_model clip-image-vit-32-tf clip-image-vit-uint8-tfjs --quantize_uint8
I converted your onnx models to tf saved model and with the current tfjs converter to tfjs models and it works very well. here is my code (based a lot on yours):
https://github.com/Jonathhhan/ofEmscriptenExamples/blob/main/emscriptenCLIP2/clip.js
This is how I converted the models:
tensorflowjs_converter --input_format tf_saved_model clip-text-vit-32-tf clip-text-vit-f16-tfjs --quantize_float16
tensorflowjs_converter --input_format tf_saved_model clip-image-vit-32-tf clip-image-vit-uint8-tfjs --quantize_uint8
https://huggingface.co/Jona0123456789
The text was updated successfully, but these errors were encountered: