Nekochu's picture
Fix warn2 Metadata
491a3d8 verified
metadata
language:
  - en
  - multilingual
base_model: Tomlim/myt5-large
model_type: myt5
model_creator: Nekochu
pretty_name: myt5-large-Stable-Diffusion-Prompt
library_name: transformers
pipeline_tag: text-generation
tags:
  - stable-diffusion
  - prompt-generation
  - text-generation
  - encoder-decoder
  - structured-output
  - human-curated
  - finetune
datasets:
  - Nekochu/discord-unstable-diffusion-SD-prompts
license: mit
prompt_template: >-
  ### Instruction: Create stable diffusion metadata based on the given english
  description. {prompt}

  ### Response: {output}
widget:
  - text: >
      ### Instruction:

      Create stable diffusion metadata based on the given english description. a
      futuristic city


      ### Response:
    example_title: Cyberpunk City
  - text: >
      ### Instruction:

      Create stable diffusion metadata based on the given english description. a
      dragon in a mystical forest


      ### Response:
    example_title: Fantasy Dragon
🎨 Stable Diffusion Text2Text
Architecture
myT5 1.2B
400k samples trained
Run the Model
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
model = AutoModelForSeq2SeqLM.from_pretrained("Nekochu/myt5-large-SD-prompts", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("Nekochu/myt5-large-SD-prompts")

prompt = "### Instruction:\nCreate stable diffusion metadata based on the given english description. a futuristic city\n\n### Response:\n" inputs = tokenizer(prompt, return_tensors="pt", max_length=256, truncation=True).to(model.device) outputs = model.generate(**inputs, max_length=256, num_beams=5, early_stopping=True) result = tokenizer.decode(outputs[0], skip_special_tokens=True) print(result)

🏙️
Cyberpunk City
SFW
Nikon Z9 200mm f_8 ISO 160, (giant rifle structure), flawless ornate architecture, cyberpunk, neon lights, busy street, realistic, ray tracing, hasselblad
🐉
Fantasy Dragon
SFW
masterpiece, best quality, cinematic lighting, 1girl, solo,
😈
Anime Succubus
NSFW
masterpiece, best quality, highly detailed background, intricate, 1girl, (full-face blush, aroused:1.3), long hair, medium breasts, nipples
Experimental Model
  • Structured output tasks perform best.
🔬 Training Evolution & Alternative Attempts
torchtune Qwen3-0.6B
decoder-only
pytorch-lightning Zamba2-1.2B
hybrid arch
tokenizer failed google/byt5
byte-level