mediapipe VS Deep Java Library (DJL)

mediapipe/docs/solutions/pose.md at master · google/mediapipe · GitHub

Today, we announce MediaPipe diffusion plugins, which enable controllable text-to-image generation to be run on-device. Expanding upon our prior work on GPU inference for on-device large generative models, we introduce new low-cost solutions for controllable text-to-image generation that can be plugged into existing diffusion models and their Low-Rank Adaptation (LoRA) variants.

You can just grab the TFLite version! https://github.com/google/mediapipe/blob/master/docs/solutions/models.md

I believe Google already released transformers under an apache 2 license with a patent grant:

https://github.com/google/mediapipe/blob/master/mediapipe/mo...

I was going to say that I like the MediaPipe Selfie Segmentation model for doing this sort of thing in a web page, but I've just noticed (when getting the GitHub link[1]) that Google have marked the code as legacy[2] ... no idea if the new solution is better/easier to use[3].

For what it's worth, my CodePen using the old model is here: https://codepen.io/kaliedarik/pen/PopBxBM

[1] - https://github.com/google/mediapipe/blob/master/docs/solutio...

[2] - "Attention: Thank you for your interest in MediaPipe Solutions. As of April 4, 2023, this solution was upgraded to a new MediaPipe Solution."

[3] - https://developers.google.com/mediapipe/solutions/vision/ima...

I have used mediapipe very successfully in multiple projects and it's very easy to get running. You can choose from many different vision tasks including hand landmarks ( https://github.com/google/mediapipe/blob/master/docs/solutions/hands.md )

I found MediaPipe's Face Mesh and was impressed with how simple it was to get going, but it just gives you the landmark points and I've not gone any further yet.

We've trained ControlNet on a subset of the LAION-Face dataset using modified output from MediaPipe's face mesh annotator to provide a new level of control when generating images of faces.

Medipaipe library link: https://mediapipe.dev/

It seems to have been picked up by Eclipse and there is also Oracle Labs' Tribuo and Deep Java Library. All seem active, but I don't know much about any of them. I agree it's probably best to follow the community and use a more popular tool like PyTorch.

Although it may be a bit more work, you can do both machine learning and AI in Java. If you are doing deep learning, you can use DeepJavaLibrary (I do work on this one at Amazon). If you are looking for other ML algorithms, I have seen Smile, Tribuo, or some around Spark.

Image preprocessing I know less about, but tokenization is something I've dealt with a bunch. There are a few options, either push the tokenizer into the ONNX model and use MS's ONNX Runtime extensions (we've used this when working with sentencepiece tokenizers), port the tokenizer entirely to Java (we did this for BERT), or use a sentencepiece or HF tokenizers wrapper directly (e.g. Amazon's DJL did this - HF, sentencepiece).

https://djl.ai/ seems very promising. I've played around with it quite a bit, not in real production though. It's a very well documented (https://d2l.djl.ai/) and active project, with Amazon working on it.

Here is link number 1 - Previous text "DJL"

To give a contrasting perspective, I think the Java ecosystem is much better suited for many data science tasks, and has a growing and well-maintained set of libraries for general purpose machine learning. I won't list them all, but TF-Java, DJL et al. have implementations of many modern architectures and there are a number of excellent libraries (CoreNLP, Lucene et al.) for working with text.

I think DJL also use use it for their tutorials - https://docs.djl.ai/jupyter/tutorial/01_create_your_first_network.html.

AWS Deep Learning more deep learning.