Update model card: Add zenlm tag, update to AutoPipelineForText2Image usage

README.md

@@ -8,6 +8,8 @@ tags:
 - text-to-image
 - image-generation
 - zen
+- zen3
+- zenlm
 - hanzo
 library_name: diffusers
 ---
@@ -30,10 +32,10 @@ Zen3 Image Dev prioritizes image quality and prompt fidelity over speed. It exce
 ## Usage
 
 ```python
-from diffusers import DiffusionPipeline
+from diffusers import AutoPipelineForText2Image
 import torch
 
-pipe = DiffusionPipeline.from_pretrained(
+pipe = AutoPipelineForText2Image.from_pretrained(
     "zenlm/zen3-image-dev",
     torch_dtype=torch.bfloat16
 )
@@ -57,35 +59,22 @@ image.save("output.png")
 | Architecture | Diffusion Transformer |
 | Precision | bfloat16 |
 | License | MIT |
-| Output Resolution | Up to 1024x1024 |
-| Inference Steps | 20-50 (recommended: 28) |
 
-## Intended Use
+## API Access
 
-- High-quality image generation
-- Creative and artistic production
-- Marketing and design assets
-- Research and experimentation
-
-## Limitations
+```bash
+curl https://api.hanzo.ai/v1/images/generations \
+  -H "Authorization: Bearer $HANZO_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{"model": "zen3-image-dev", "prompt": "A majestic mountain landscape"}'
+```
 
-- Slower inference compared to Zen3 Image Fast
-- Requires more VRAM for optimal quality
-- Text rendering in generated images may be imperfect
-- Complex multi-subject scenes may require prompt refinement
+Get your API key at [console.hanzo.ai](https://console.hanzo.ai) — $5 free credit on signup.
 
-## Citation
+## License
 
-```bibtex
-@misc{zen3-image-dev,
-  title={Zen3 Image Dev: High-Fidelity Text-to-Image Generation},
-  author={Hanzo AI},
-  year={2025},
-  url={https://huggingface.co/zenlm/zen3-image-dev}
-}
-```
+MIT
 
-## Links
+---
 
-- [Zen Model Family](https://huggingface.co/zenlm)
-- [Hanzo AI](https://hanzo.ai)
+*Zen LM is developed by [Hanzo AI](https://hanzo.ai) — Frontier AI infrastructure.*