Downloads · 30 days
0
knightfall_001_models
Karajan42/knightfall_001_models
knightfall_001_models is a machine learning model from Karajan42. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Diamond world model trained on Knightfall combat gameplay dataset.
Access
Public
Updated Jan 31, 2026
Repo size
250 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt250 MB · 100%
At a glance
- Access
- Public
- Created
- Jan 31, 2026
- Updated
- Jan 31, 2026
- SHA
- bb1d9d35
Model card
From the Hugging Face model README
Knightfall 001 World Model
Diamond world model trained on Knightfall combat gameplay dataset.
Models
| File | Description | Size |
|---|---|---|
autoencoder_small_best_epoch_050.pt | Full autoencoder (encoder + decoder) | ~900KB |
autoencoder_small_decoder_epoch_050.pt | Decoder only (for inference) | ~150KB |
agent_epoch_00100.pt | World model checkpoint (epoch 100) | ~83MB |
agent_epoch_00125.pt | World model checkpoint (epoch 125) | ~83MB |
agent_epoch_00150.pt | World model checkpoint (epoch 150) | ~83MB |
Action Space
| Index | Action |
|---|---|
| 0 | idle |
| 1 | attack |
| 2 | block |
Training Details
- Dataset: knightfall_001 (20 episodes, 37,579 frames)
- Resolution: 64x64 (low-res), 256x256 (full-res)
- Architecture: Latent diffusion (~5M params denoiser)
- Autoencoder: Small (~70K params)
Usage
# Load world model
import torch
agent = torch.load("agent_epoch_00150.pt")
# Load decoder for upsampling
decoder = torch.load("autoencoder_small_decoder_epoch_050.pt")
License
Research use only.
- Created
- Jan 31, 2026
- Updated
- Jan 31, 2026