ired-reasoning / README.md
roy-W's picture
Upload README.md with huggingface_hub
461a5e4 verified
|
Raw
History Blame Contribute Delete
661 Bytes
---
library_name: ired
tags:
- autoencoder
- bart
- latent-diffusion
---
# IRED Autoencoder — `roy-W/ired-reasoning`
Frozen BART autoencoder with compression/reconstruction split.
- **Pool type:** `conv`
- **Latent slots (K):** 384
- **d_ae:** 768
- **Decoder fine-tuned:** yes
- **Base model:** `facebook/bart-base`
- **Training steps:** 5
## Usage
```python
from ired.model.autoencoder import FrozenBartAutoencoder
import torch
ae = FrozenBartAutoencoder(
model_name='facebook/bart-base',
k=384,
pool_type='conv',
d_ae=768,
train_decoder=True,
)
ckpt = torch.load('ae_checkpoint.pt', map_location='cpu')
ae.load_ae(ckpt['ae'])
```