--- license: cc-by-nc-sa-4.0 language: - en task_categories: - video-to-video configs: - config_name: default data_files: - split: train path: metadata.jsonl dataset_info: features: - name: id dtype: string - name: style_id dtype: string - name: style_label dtype: string - name: instruction dtype: string - name: source_video dtype: video - name: source_depth_video dtype: video - name: style_video dtype: video - name: provenance dtype: string --- ## 当前训练数据:不拉伸版 v2 当前新批次使用不拉伸DINO筛选,距离阈值保持0.05;训练准备也保持宽高比。旧版v1记录保留但不混入新版。 [新版数据与下载训练说明](training/ditto-depth-nostretch-v2-20261027/README.md) · [可下载样本索引](training/ditto-depth-nostretch-v2-20261027/index.json) ## 可下载训练数据(2026-10-27增量批次) 新增批次包含实际原视频、编辑视频和对应深度视频,不只是筛选记录。共享源视频及深度全局去重,采用约1GB分包并持续上传;清单仅收录素材已上传完成的样本。 - [下载与训练说明](training/ditto-depth-v1-20261027/README.md) - [当前可下载样本与分包索引](training/ditto-depth-v1-20261027/index.json) - [可训练清单](training/ditto-depth-v1-20261027/metadata.jsonl) - [实际视频与深度分包](https://huggingface.co/datasets/faraway00/StyleEdit/tree/main/training/ditto-depth-v1-20261027/shards) - [原视频预览](https://huggingface.co/datasets/faraway00/StyleEdit/resolve/main/training/ditto-depth-v1-20261027/preview/source.mp4) · [深度预览](https://huggingface.co/datasets/faraway00/StyleEdit/resolve/main/training/ditto-depth-v1-20261027/preview/depth.mp4) · [编辑视频预览](https://huggingface.co/datasets/faraway00/StyleEdit/resolve/main/training/ditto-depth-v1-20261027/preview/edited.mp4) 下面的旧版8400条默认集保留不变,新批次请使用上述目录及下载脚本。 # StyleEdit: Ditto + DA3 Tuples: original video, original-video depth, style video, and the unchanged editing instruction. Current verified tuples: **8400**. Target: 8,400 accepted tuples; processing is ongoing. Source: [QingyanBai/Ditto-1M](https://huggingface.co/datasets/QingyanBai/Ditto-1M), revision `aac1417418d8f7e77d4e257d5620c36b2c3a00e8`. Attribution: Qingyan Bai et al., *Ditto: Scaling Instruction-Based Video Editing with a High-Quality Synthetic Dataset*. Source license CC BY-NC-SA 4.0 is retained for this derivative dataset. The original media are unmodified. Depth was estimated using [depth-anything/DA3-LARGE](https://huggingface.co/depth-anything/DA3-LARGE). ## Selection 496 instruction variants are mapped to 84 explicit style groups: 35 paraphrases each for 12 named styles are folded together, then Jojo, Vector/vector animation, Poly/low-poly, and Paper Cutting/papercut aliases are merged. Distinct artist/game-specific styles remain distinct. Full mapping and candidate counts: `styles.json`. Target: **8,400 accepted tuples**, 100 distinct source clips per style. Availability after filtering is not guaranteed; exhausted quotas will be reported rather than filled with rejected data. Core-v1 checks all frames: full decoding, equal frame counts and FPS, relative timestamp tolerance 1 ms; EdenLive projected DINO mean distance <=0.05; full native-pixel darkness/overexposure checks. Exact settings and per-frame scores are in each sample's `filter.json`. No OCR, GMFlow, or VGGT filter is included in this version. ## Depth representation DA3-LARGE runs jointly on all source frames, process resolution 504, upper-bound resize. Only depth video is persisted; no NPZ arrays are generated or uploaded. Values are **not meters**. `depth.mp4` is a native-size, all-frame visual representation with a single inverse-depth 0.5/99.5 percentile mapping per entire clip (near bright). MP4 uses 8-bit quantization and lossy H.264 CRF 16 encoding; original float depth cannot be recovered. Every depth video is checked against all original frame timestamps (<=1 ms). Current worker explicitly reports an error for variable-frame-rate clips, >160-frame sequences, >2 GiB decoded RGB, or model failures; it never silently truncates, changes frame rate, or uploads a failed tuple. ## Storage and verification `data///{source.mp4,depth.mp4,edited.mp4,depth.json,filter.json,record.json}`. Each sample is committed atomically, then all remote file sizes and SHA256 values are verified at the immutable commit. Only verified samples count toward quotas and become eligible for local deletion. Immutable media URLs and original instructions are in `metadata.jsonl`. Source archives are downloaded, SHA256-verified, and extracted on the Pod root disk before processing. Six GPU workers filter full videos and generate DA3 depth; accepted tuples are uploaded asynchronously and verified at immutable commits before processing scratch is removed. Source archives and extracted input videos remain on the root disk. SQLite runs locally; standalone database snapshots and verification receipts persist on the network volume. Historical pilot and range-index benchmarks are retained for provenance. The current root-disk pipeline downloads the source, global_style1 and global_style2 split archives at the fixed source revision, then reads extracted files directly. On startup it reconciles missing local checkpoints from the remote verified manifest; publication rejects any manifest that would omit existing remote samples. ## Measured pilot See `pilot-report.json` and `benchmark.json`; wall time includes fresh pilot-range download, model load, processing and verified sample upload/cleanup, excluding this final report publication. Source clips can differ in aspect ratio and therefore DA3 compute cost. Concurrent stage durations must not be added as wall time. ## 2026-10-29 全风格续跑批次 根盘异步打包上传,保留既有数据。仅收录远端分包已校验的样本。 [下载训练说明](training/ditto-depth-nostretch-v2-resume-20261029/README.md) · [已上传样本索引](training/ditto-depth-nostretch-v2-resume-20261029/index.json)