Add support for vLLM and local models to GeminiVideoSummarization#436

Open

adsarver wants to merge 43 commits into

openai-api-video-summary

Commits on Jun 27, 2025

Start of GeminiSummarizationComponent
kburkewv
committed

Commits on Jul 15, 2025

Altered to work with VertexAI. Does no tracks or detections
kburkewv
committed
Added support for tracks
kburkewv
committed

Commits on Jul 16, 2025

Updated the default prompt to use structured output
kburkewv
committed

Commits on Jul 21, 2025

Added gcp certifcate file and other adjustments
kburkewv
committed

Commits on Jul 22, 2025

Added more properties and functionality
kburkewv
committed

Commits on Jul 24, 2025

Adjusted prompt to fix time formatting issues
kburkewv
committed
Changed two lines to work with time formatting
kburkewv
committed

Commits on Aug 7, 2025

Added segmentation and other functionality
kburkewv
committed
Added minimal unit testing
kburkewv
committed
Added more detail for timestamps (including milliseconds)
kburkewv
committed
Reverting the previous change. Adding more detail to README.
kburkewv
committed

Commits on Aug 29, 2025

Merge branch 'develop' into kburkewv/gemini-video
kburkewv
authored

Commits on Sep 17, 2025

Added a feature to disable/enable timelines
kburkewv
committed
Merging develop into kburkewv/gemini-video
kburkewv
committed

Commits on Oct 6, 2025

Merge remote-tracking branch 'origin/develop' into kburkewv/gemini-video
kburkewv
committed

Commits on Nov 20, 2025

Removing the subprocess
kburkewv
committed

Commits on Nov 25, 2025

Altered time formatting for gemini request
kburkewv
committed

Commits on Nov 27, 2025

Add comment explanation for retry logic
kburkewv
committed

Commits on Jan 20, 2026

Added to the Unit Tests
kburkewv
committed
Merge 'origin/develop' into kburkewv/gemini-video
kburkewv
committed

Commits on May 20, 2026

Commits on May 21, 2026

Commits on May 22, 2026

Videos now emphasize frames with higher motion scores to combat model rate limits
adsarver
committed

Commits on May 27, 2026

Commits on May 28, 2026

Updated prompts and motion sampler, reduced sampler overhead
adsarver
committed

Commits on Jun 4, 2026

Added support for audio and vLLM parameters, updated tests, updated prompt for audio
adsarver
committed

Commits on Jun 15, 2026

Updated testing docker file, added new diffusion gemma file, added ffmpeg to the config, changed how motion is profiled, added support for audio tracks
adsarver
committed

Commits on Jul 6, 2026

Changed audio encoding, removed tracks support
adsarver
committed

Commits on Jul 9, 2026

Torch import fixes and descriptor fixes
adsarver
committed

Commits on Jul 10, 2026

Added dockerignore so that the test data is not included in the container, fixed typo in vllm dockerfile, added optional testing at image build without GPU
adsarver
committed