CaT-Movie
Collection
CaT-Movie: 5B weights, 75k blocks, umT5-XXL & Wan2.2/2.1 VAE caches, 18,173 source clips, cut-gated CaT-MBench, 736 videos & labels. • 8 items • Updated
image imagewidth (px) 480 8.88k | label class label 3
classes |
|---|---|
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames | |
0frames |
Archive Notice: Preserves the November 6 preliminary single-author calibration set used while developing the
CaT-MBenchVLM judge prompts.
- Do NOT treat
"Ours"rows here as the releasedCaT-Moviemodel: they come from an earlier internal development checkpoint rather thanstep-3000.safetensors.- For the final dual-annotator blind evaluation (
557reviewed items) reported in the paper, usehuman/human_labels.jsonandhuman/agreement.jsoninJinPLU/CaT-MBench.
| Path | Description |
|---|---|
consistency/labels.jsonl & consistency/frames/ |
Frame strips (.webp) and author consistency annotations for VLM calibration |
cut_execution/labels.jsonl & cut_execution/frames/ |
Consecutive-shot boundary frame pairs (a1/a2, b1/b2) and author cut-execution labels |
event_state/labels.jsonl & event_state/clips/ |
Short evaluation clips (.mp4) and author labels for event & state checks |
hard_negatives/labels.jsonl |
Hard-negative diagnostic cases verifying judge refusal of copied framing |
hf download JinPLU/CaT-MBench-Author-Labels \
--repo-type dataset \
--local-dir data/cat_mbench_author_labels_archive
| Repository | Tier | Contents |
|---|---|---|
JinPLu/CaT-Movie (GitHub) |
Core | Planner, training, inference, curation & CaT-MBench code (Apache-2.0) |
JinPLU/CaT-Movie |
Core | Renderer weights (step-3000.safetensors) |
JinPLU/CaT-Movie-Data |
Core | 75,461 blocks, Qwen3.7-Flash schemas, grounding, 96,000 train windows & SAM 3.1 weights |
JinPLU/CaT-Movie-T5-Cache |
Core | Precomputed umT5-XXL BF16 caches (332,516 unique prompts) |
JinPLU/CaT-Movie-VAE-Caches |
Core | Precomputed VAE target latents (Wan2.2 704p + Wan2.1 480p) |
JinPLU/CaT-Movie-Source-Videos |
Supporting | Complete 18,173 block-yielding source MP4 clips |
JinPLU/CaT-MBench |
Core | 122 stories, cut-gated GPT-6.1-sol VQA bank, Gemini 3.8 Flash verdicts & 557 human labels |
JinPLU/CaT-MBench-Videos |
Supporting | 736 generated test videos (92 test stories × 8 methods) |
JinPLU/CaT-MBench-Author-Labels (this repo) |
Archive | Early Nov 6 single-author calibration archive (superseded by CaT-MBench/human/) |
Released under CC BY-NC-SA 4.0 strictly for non-commercial academic research.
@misc{catmovie2026,
title = {CaT-Movie: Balancing Global Planning and Local Rendering for Long Multi-Shot Video Generation},
author = {Jinpeng Lu and Haoyuan Shi and Yuanshen Guan and Jiyuan Xia and Bin Wu and Yinda Chen and Zhiwei Xiong},
year = {2026},
url = {https://jinplu.github.io/CaT-Movie/}
}