AI Paper Atlas

Research map / Video / Video-language understanding

VMChill: A Dataset for Fine-Grained Visual-Musical Synergy

Xiaowei Chi, Zeyue Tian, Jialiang Chen, Wei Xue

AAAI 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
fine-tune VideoCrafter-2.0 on VMChill-2MVideoCrafter-2.0 · VBenchsubject consistency: 96.85 -> 98.62Improved
fine-tune VideoCrafter-2.0 on VMChill-2MVideoCrafter-2.0 · VBenchmotion smoothness: 97.73 -> 98.33Improved
fine-tune VideoCrafter-2.0 on VMChill-2MVideoCrafter-2.0 · VBenchoverall consistency: 28.23 -> 25.33Hurt
fine-tune VideoCrafter-2.0 on VMChill-2MVideoCrafter-2.0 · VBenchtemporal style: 25.84 -> 24.61Hurt
fine-tune VideoCrafter-2.0 on VMChill-2MVideoCrafter-2.0 · VBenchsum of 9 dimensions: 64.45 -> 64.57No clear effect
fine-tune VideoCrafter-2.0 on VMChill-2MVideoCrafter-2.0 · VBenchimage quality: 67.22 -> 69.78Improved
train Video-LLaMA on VMChill-2M instead of official fine-tuned weightsVideo-LLaMA · VMChill-TestCIDEr: 2.45 -> 24.79Improved
train Video-LLaMA on VMChill-2M instead of official fine-tuned weightsVideo-LLaMA · VMChill-TestBLEU-4: 3.94 -> 5.59Improved
train Video-LLaMA on VMChill-2M instead of official fine-tuned weightsVideo-LLaMA · VMChill-TestMETEOR: 14.05 -> 13.83Hurt
fine-tune Stable-Audio-Open on VMChill-2MStable-Audio-Open · MusicCapsFAD: 3.23 -> 2.40Improved
fine-tune Stable-Audio-Open on VMChill-2MStable-Audio-Open · MusicCapsKL: 1.51 -> 1.20Improved
use DiT trained from scratch on VMChill-2M (V2M-DiT) instead of VidMusevideo-to-music · V2M-BenchFAD: 2.46 -> 2.51No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy