CoolFace
Datasetpublic

TempoFunk/tempofunk-sdance

TempoFunk S(mall)Dance 10k samples of metadata and encoded latents & prompts of videos themed around dance. 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 URLs, etc.)

sourceHugging Faceagpl-3.0updated 3y agoView on Hugging Face
5likes22kdownloads

TempoFunk/tempofunk-sdance · main · files are served by the source, never re-hosted here

TempoFunk/tempofunk-sdance · CoolFace