Thhis is the top ranked in VLM models benchmarks
Thhis is the top ranked in VLM models benchmarks
MYep even the 8b model tops many of the other ones
DDid you use any app or tool to load this model?
Mi manually used the huggingface space
Dnice, hope @Dr. Furkan Gözükara can create a set up like Cogvlm 2
DWhy does my terminated instance keep rising back up on Massed Compute? 
DGuess it finally died this time lol.
MHey any advices for training a Flux style Lora with 100 images? For now I have been using Ostris Colab.
Muse onetrainer now
Mits fixed and the loras work on forge and comfy
MAny rec. Settings?
TI can't find a solution how to generate a queue of 100 different promts in order. The context window allows you to enter a lot of text with promts where they are placed in one line, but only the last one is launched
DWhat are you using? SwarmUI generator?
DUse seperator || to separate the prompts.
D
Fyes . you can use hugging face for backup which i use
Fyou need to generate if folder missing
Fshared the prompts here , will put into our zip file as well : https://gist.github.com/FurkanGozukara/12c837939ca3fc59abc04d8bd44876ab
Gist
100 flux prompts.txt. GitHub Gist: instantly share code, notes, and snippets.
Fi think yes and hopefully gonna start tests today
Fdid you notice that in my grids?
Fi saw such thing when only i used 2x upscale of swarmui
Fdo you use system prompt? if so what is it
Mjust caption the image with details, colors, makeup, items, objects, emotions, art style, drawing style and objects but do not add any description or comment. do not miss any item in the given image, very simple sentance.
Mthey have a demo interface
M@Dr. Furkan Gözükara its quite nice actually, able to capture details better as well. i've tried longer prompts too works quite nice
Mi think it would be better than joy and cog for style training too
Xjust create a folder titled Umet in the model files inside of swarmui?
D@Dr. Furkan Gözükara Yes please see if InternVL2 is possible
Mit crazy that the 8b models outperforms many of the others
M26b 4bit it possilbe as well, its the best one from my testing
Fthey lack Gradio Demo
Fexactly as that
Mthey have streamlit demo
Ftheir demo down
FCalling st.rerun() within a callback is a no-op.
MMthis one? its working for me
Ftesting here
Mboth work
Meven for me
F
MSeems pretty good!
F
Fnow i will compare generated images on swarmui
Mi think in terms of training, when you train this many tokens for a person its not good, but the simplified but accurate ones from internvl gave me quite good results
Fwell i think for training a person no captions better :d
Mdepends, yours are usually overtrained
Xokay i made the unet file inside of models then restarted/updated but still getting same error saying "All available backends failed to load the model"
