MiniGPT-4 VS stable-diffusion-webui-wd14-tagger

I just think the tech has been out for so long it's not as big of a deal. Mini-Gpt4 has been out for 6 months! Of course the descriptions aren't exactly gpt-4 grade, but with mistral 7b being used as the language model instead of llama 7b, the reasoning ability will improve noticeably.

[1] https://github.com/Vision-CAIR/MiniGPT-4

Isn't there only two open multimodal LLMs, LLaVA and mini-gpt4?

For LLM visual instruction, you can use LLaVA, LaVIN, or MiniGPT-4.

Please read the rules before posting. If you want a model for visual instruction, use LLaVA, LaVIN, or MiniGPT-4.

So we use MiniGPT-4 for image parsing, and yep it does return a pretty detailed (albeit not always accurate) description of the photo. You can actually play around with it on Huggingface here.

We use MiniGPT-4 first to interpret the image and then pass the results onto GPT-4. Hopefully, once GPT-4 makes its multi-modal functionality available, we can do it all in one request.

But I would like to bring up that there are some multi models(llava, miniGPT-4) that are built based on censored llama based models like vicuna. I tried several multi modal models like llava, minigpt4 and blip2. Llava has very good captioning and question answering abilities and it is also much faster than the others(basically real time), though it has some hallucination issue.

Create an extension like 'send pictures' that uses the WD14 tagger which is way more detailed and has options for nsfw etc. Its used in Automatic1111 and Koyha ss so there's extensions you can probably implement from. https://github.com/toriato/stable-diffusion-webui-wd14-tagger

If anyone is interested I made some changes to toriato's wd14-tagger, now it works also on vladmandic webui, repo here. You can do a new installation, or use your old automatic1111 one changing 3 files, instructions on my repo. The lora files also work (there were some problems in the vlad issue page). I'm not a programmer and it's not perfect though, in fact for now if you don't like the default tagger model you have to change it manually (instructions in the repo), and since it is basically a fork of toriato's version, if there were errors there, there will be here too.

Automatic captioning tools that can be used as an initial point for captions: this tool or this one.

wd14-tagger, to describe anime images and get a prompt idea

Use WD 1.4 tagger (https://github.com/toriato/stable-diffusion-webui-wd14-tagger) to extract prompt words from each frame (threshold 0.65), then use the dataset tag editor (https://github.com/toshiaki1729/stable-diffusion-webui-dataset-tag-editor) for batch editing, mainly:

I recommend using EveryDream2. You'll need an 11GB VRAM GPU. There's no need to crop or resize images, just caption them, which can be done automatically with CLIP Interrogator or WD14 taggers. Make sure to add the trigger word for your subject. It's not a Dreambooth script; it's actual training, so it shouldn't be as destructive to the model as Dreambooth. Typically, using an LR of 1e-6 with a cosine scheduler over two epochs and a batch size of 4 works fine. This script supports validation, so you can actually watch in real-time whether the training is going well or if you're overfitting. I got very good results using it.

In my current process, I use CLIP Interrogator to produce a high level caption and wd14 tagger for more granular booru tags. Typically in that order, because you can append the results from the latter to the former. Both tools perform with greater accuracy than the standard interrogators in img2img and give you more flexibility and features as well. You still have to do some manual adjustments, but I generally prefer this process over starting from scratch.