ltx_video
LTX-Video-Activations-FP4LTX-Video-Activations-INT4azm-archive-20260909-ltx-i2v-videos-1370
ltx_i2v_videos_1370.tar
Backup of an existing dataset archive, preserving its original bytes.
File: ltx_i2v_videos_1370.tar
Size: 2,247,976,960 bytes
SHA256: b46c8369b4d8312ae1bae4c02aca0a8e31cd8d9e4b6a598a9689623f42ca8bbf
Verify the downloaded archive with sha256sum -c SHA256SUMS.
sketch-to-video-dataset
Sketch → Reality — IC-LoRA training set
15 clips from linoyts/repo-to-space-example-videos (CC0), each paired with an auto-derived coarse 'kid's-drawing' line sketch (*_reference.mp4) as the IC-LoRA control. Trainer-format dataset.json (media_path + caption + reference_video). The LoRA learns sketch → photoreal video.
Note: wide scenes (surfing, landscape) yield sparse sketches — drop them for a cleaner demo.
LTX-VIDEOLTX-training-videos