Upload vae checkpoint

README.md

@@ -0,0 +1,41 @@
+---
+license: mit
+tags:
+  - tiny-stable-diffusion
+  - vae
+  - image-generation
+  - diffusion
+library_name: pytorch
+---
+
+# tiny-sd-models
+
+This is a **VAE** model trained with [tiny-stable-diffusion](https://github.com/your-username/tiny-stable-diffusion).
+
+## Model Description
+
+This is a Variational Autoencoder (VAE) trained to compress images into a latent space.
+The VAE follows the SD3 architecture with 16 latent channels and f8 compression ratio.
+
+### Architecture
+- **Type**: AutoencoderKL
+- **Latent Channels**: 16
+- **Compression**: f8 (64x64 → 8x8)
+
+## Usage
+
+```python
+import torch
+from src.models.vae import create_vae  # or appropriate model import
+
+# Load checkpoint
+checkpoint = torch.load("model.pt", map_location="cpu")
+
+# Create model and load weights
+model = create_model(...)  # Use config from checkpoint
+model.load_state_dict(checkpoint["model_state_dict"])
+```
+
+## License
+
+MIT License

vae.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e98929e0bc3a5209600626aa83707660559d21c6a9c4a5140e34eadc7fbb6473
+size 252594535