Stability AI JapanProprietary

sd-vae-ft-mse

Compare this model

A fine-tuned VAE model for Stable Diffusion developed by Stability AI. Trained using MSE loss for improved reconstruction quality.

Parameters

Undisclosed

Context Window

License

Proprietary

Release Date

2022-10-13

API Pricing

API pricing for this model is not yet available

Strengths

  • Fine-tuned with MSE loss for better accuracy
  • Compatible with Stable Diffusion models
  • Improved image reconstruction quality

Weaknesses

  • May produce slightly blurrier outputs
  • Specific to certain Stable Diffusion versions
  • Trade-off between sharpness and accuracy

Use Cases

  • Image generation with Stable Diffusion
  • High-fidelity image reconstruction
  • Latent space encoding and decoding