google-ai-edge/mediapipe

What is the .task LoRA file format that the Image Generator wants?

Closed this issue · 1 comments

Have I written custom code (as opposed to using a stock example script provided in MediaPipe)

No

OS Platform and Distribution

Android 14

MediaPipe Tasks SDK version

0.10.14

Task name (e.g. Image classification, Gesture recognition etc.)

Image Generation

Programming Language and version (e.g. C++, Python, Java)

Java

Describe the actual behavior

Error when loading .safetensor or pytorch .bin files

Describe the expected behaviour

Model and LoRA both initialize

Standalone code/steps you may have used to try to get what you need

Download the example LoRA files from the LoRA colab file in the Image Generation guide. Causes exception when loaded as LoRA. https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_mediapipe_image_generation.ipynb

The demo app points the LoRA file path to a teapot_lora.task. Are there any example LoRA files in the .task format, or even conversion scripts from pytorch to .task?

Other info / Complete Logs

No response

LoRA task file is in the example app readme.