TempoFunk/makeavid-sd-jax
Text-to-Video
•
Updated
•
3
•
10
Dataset Viewer
The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider
removing the
loading script
and relying on
automated data support
(you can use
convert_to_parquet
from the datasets library). If this is not possible, please
open a discussion
for direct help.
TempoFunk Small
7.8k samples of metadata and encoded latents & prompts of random videos.
Data format
- Video frame latents
- Numpy arrays
- 120 frames, 512x512 source size
- Encoded shape (120, 4, 64, 64)
- CLIP (openai) encoded prompts
- Video description (as seen in metadata)
- Encoded shape (77,768)
- Video metadata as JSON (description, tags, categories, source URL, etc.)
- Downloads last month
- 4,832
Models trained or fine-tuned on TempoFunk/small