Skip to content

Add local prompt-based video-to-video editing pipeline (video_edit/)#6

Open
AndriiShramko wants to merge 2 commits into
mainfrom
claude/rtx6000-local-video-model-8cziG
Open

Add local prompt-based video-to-video editing pipeline (video_edit/)#6
AndriiShramko wants to merge 2 commits into
mainfrom
claude/rtx6000-local-video-model-8cziG

Conversation

@AndriiShramko

Copy link
Copy Markdown
Owner

Node-free, fully scripted video editing for RTX PRO 6000 Blackwell (96GB).
Research-backed model selection (June 2026):

  • Kiwi-Edit (MIT) as default — OpenVE-Bench NVIDIA-lyra ? for-commerial-use. #1 open editor
  • LIVEditor-14B — runs single-GPU on 96GB, triple-benchmark claim
  • Ditto — IVEBench instruction-following leader (research only)
  • Lucy-Edit-Dev — fast lightweight option
  • SeedVR2-7B — temporally-consistent upscale via --upscale

Unified CLI (edit.py) dispatches to per-model backends; install.sh pins a
Blackwell sm_120 PyTorch nightly + SageAttention and clones upstream repos.

https://claude.ai/code/session_01ADDj1qHAaD7XW1AbDoKNP2

claude added 2 commits June 7, 2026 13:55
Node-free, fully scripted video editing for RTX PRO 6000 Blackwell (96GB).
Research-backed model selection (June 2026):
- Kiwi-Edit (MIT) as default — OpenVE-Bench #1 open editor
- LIVEditor-14B — runs single-GPU on 96GB, triple-benchmark claim
- Ditto — IVEBench instruction-following leader (research only)
- Lucy-Edit-Dev — fast lightweight option
- SeedVR2-7B — temporally-consistent upscale via --upscale

Unified CLI (edit.py) dispatches to per-model backends; install.sh pins a
Blackwell sm_120 PyTorch nightly + SageAttention and clones upstream repos.

https://claude.ai/code/session_01ADDj1qHAaD7XW1AbDoKNP2
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