vincentg/
llama3.2-fluxassistant:latest

1,371 1 year ago

Finetuned llama 3.2 3b to enrich prompt for image generation

1 year ago

209b0e8a4303 · 3.4GB ·

model
llama
·
3.21B
·
Q8_0
template
{{- range .Messages }}<|start_header_id|>{{ .Role }}<|end_header_id|> {{ .Content }}<|eot_id|> {{- e
system
You are a prompt creation assistant for FLUX, an AI image generation model. Your mission is to help
params
{ "stop": [ "<|start_header_id|>", "<|end_header_id|>", "<|eot_id|>"

Readme

A llama 3.2 finetuned to enrich user prompt for image generation tools. based on this model : https://huggingface.co/VincentGOURBIN/Llama-3.2-3B-Fluxed

Input : a toucan coding on a mac

Picture generated with Flux.dev from user input 20241201_145537_251483619.png

Output : “Create an image of a toucan comfortably perched on top of a large desk, amidst coding-related tools and software windows. Focus on capturing vivid details of the bird’s bright plumage, especially its striking orange and yellow feathers. The toucan should be intently engaged in typing away on a Mac laptop screen, with a hint of concentration or curiosity. Incorporate coding symbols like brackets, semicolons, and other visual elements to add depth and engagement. Consider adding subtle ambient light effects around the computer screens to give an impression of being near computer screens. The overall style should be vibrant, fun, and engaging.”

Picture generated with Flux.dev from enriched prompt 20241201_144436_3304061800.png