Thanks to visit codestin.com
Credit goes to github.com

Skip to content

VCD2 - #4

Open
Babak (babaknaderi) wants to merge 4 commits into
mainfrom
vcd2
Open

Babak (babaknaderi) wants to merge 4 commits into
mainfrom
vcd2

Conversation

@babaknaderi

Copy link
Copy Markdown
Contributor

Summary

Add the VCD2 camera-native talking-head dataset and update the existing VCD download tooling and documentation.

Changes

  • Document VCD2, its benchmarking subset, metadata, scenarios, and citation.
  • Add VCD2 raw and benchmarking download scripts.
  • Add a parallel Python downloader with retries and partial-download cleanup.
  • Add VCD2 file manifests and visualization assets.
  • Improve the existing VCD1 downloader and scripts.
  • Standardize Microsoft MIT license headers across source files.
  • Update supporting documentation and links.

Dataset details

  • 838 camera-native, losslessly encoded FFV1 clips.
  • 120-clip benchmarking subset covering talking head, background blur, and background replacement scenarios.
  • Metadata includes MOS, SI/TI, perceptual-quality tokens, camera information, and demographics

Babak (babaknaderi) and others added 4 commits June 8, 2026 21:46
- Update README with VCD2 section (839 clips, FFV1 lossless) and VCD2-Benchmarking subset (120 clips)
- Add download scripts (bash + Python with parallel download support)
- Add file lists for raw MKV (839) and benchmarking MKV (120)
- Add VCD2 thumbnail atlas and SI/TI images for TH, TH-BB, TH-BR groups
- Add VFR note explaining CFR correction for benchmarking subset

Co-authored-by: Copilot <[email protected]>

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR adds documentation and download tooling for the new VCD2 camera-native talking-head dataset while modernizing the existing VCD (VCD1) download scripts (bash + Python) and support guidance.

Changes:

  • Document VCD2 (dataset summary, citation, download instructions, metadata, scenarios) and refresh the top-level README.
  • Add VCD2 download tooling (bash scripts, Python parallel downloader, file manifests) under download/vcd2.
  • Improve VCD1 download tooling by updating the Python downloader and hardening bash scripts (retries/resume/partial cleanup), plus update SUPPORT.md.

Reviewed changes

Copilot reviewed 12 out of 18 changed files in this pull request and generated 4 comments.

Show a summary per file
File Description
SUPPORT.md Replaces template placeholder content with concrete issue/support guidance.
README.md Documents VCD2 + benchmarking subset, adds download + metadata + scenario notes, keeps VCD1 section.
download/vcd2/requirements.txt Adds Python dependency list for VCD2 downloader.
download/vcd2/raw_mkv_list.txt Adds VCD2 raw dataset manifest (838 MKV paths).
download/vcd2/benchmarking_mkv_list.txt Adds VCD2 benchmarking subset manifest (120 MKV paths).
download/vcd2/downloader.py Adds a parallel Python downloader with retries and partial-download cleanup.
download/vcd2/download-vcd2-raw-mkv.sh Adds bash downloader for VCD2 raw dataset (retry + resume + .part).
download/vcd2/download-vcd2-benchmarking-mkv.sh Adds bash downloader for VCD2 benchmarking subset (retry + resume + .part).
download/requirements.txt Simplifies requirements list (drops stdlib argparse).
download/downloader.py Replaces VCD1 downloader with a more robust version (retries, .part, size validation).
download/download-vcd1-yuv.sh Updates VCD1 YUV downloader to use safer bash defaults and retry/resume behavior.
download/download-vcd1-mp4.sh Updates VCD1 MP4 downloader to use safer bash defaults and retry/resume behavior.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

Comment on lines +88 to +95
lines = [
line.strip()
for line in file_list.read_text(encoding="utf-8-sig").splitlines()
if line.strip()
]
if len(lines) != len(set(lines)):
raise ValueError(f"{file_list} contains duplicate paths")

Comment thread download/downloader.py
Comment on lines +31 to +38
paths = [
line.strip()
for line in file_list.read_text(encoding="utf-8-sig").splitlines()
if line.strip()
]
if len(paths) != len(set(paths)):
raise ValueError(f"{file_list} contains duplicate paths")

Comment on lines +16 to +20
BLOB="${BLOB%$'\r'}"
[ -z "$BLOB" ] && continue
URL="$AZURE_URL/$BLOB"
DES="$OUT_DIR/$BLOB"
TMP="$DES.part"
Comment on lines +16 to +20
BLOB="${BLOB%$'\r'}"
[ -z "$BLOB" ] && continue
URL="$AZURE_URL/$BLOB"
DES="$OUT_DIR/$BLOB"
TMP="$DES.part"
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants