Crimson AI NewsA CrimsonLingua Network service
EN ع
← Back to news
Research paper Hugging Face

VCSD: Visual Contrastive Self-Distillation Boosts VLMs Without External Teachers

AI By Crimson AI Hugging Face Papers 24 July 2026 · 00:00 7 views
Share: X Telegram

Researchers propose Visual Contrastive Self-Distillation (VCSD), a novel on-policy self-distillation method that improves vision-language models by contrasting predictions conditioned on original vs. content-erased images, eliminating the need for external teachers, privileged answers, or visual evidence.

VCSD: Visual Contrastive Self-Distillation Boosts VLMs Without External Teachers

Key points

Researchers have introduced Visual Contrastive Self-Distillation (VCSD), a new on-policy self-distillation (OPSD) method for vision-language models (VLMs) that relies solely on input conditioning, removing the need for external teachers, privileged answers, or auxiliary visual evidence.

Traditional on-policy distillation (OPD) requires a separate teacher model, while existing OPSD methods create asymmetry through privileged answers or visual evidence. VCSD achieves this asymmetry by comparing the same EMA teacher's predictions under two conditions: the original image and a content-erased control, with identical prompt and student-generated response prefix.

The token-wise log-probability difference between these two conditions highlights candidates whose likelihood is specifically increased by the instance-level visual content. This contrast is used to sharpen the teacher's original-image distribution within its plausible support, and the resulting full-distribution target is distilled into the student model.

Evaluated on the ViRL39K dataset across six Qwen3-VL and Qwen3.5 models, VCSD consistently outperforms matched OPSD on every model, with gains ranging from +1.76 to +5.33 points. For example, on Qwen3-VL, the seven-benchmark aggregate improves from 62.27% to 67.04% at 2B, 71.30% to 73.16% at 4B, and 72.51% to 76.26% at 8B. The Qwen3.5-9B model reaches 79.24% average accuracy.

VCSD requires no external teacher, privileged answers, visual evidence signals, reasoning traces, or additional inference-time cost. The project and code are available at https://joliang17.github.io/VisualCSD/.

ModelBaseline (OPSD)VCSDImprovement
Qwen3-VL 2B62.27%67.04%+4.77%
Qwen3-VL 4B71.30%73.16%+1.86%
Qwen3-VL 8B72.51%76.26%+3.75%
Qwen3.5-9B79.24%
Source
Hugging Face · Hugging Face Papers
Related news
Research paper
Hugging Face 31 Aug 2026

Hugging Face Unveils StepGuard: Step-Level Guardrails for Safer AI Agents

StepGuard, a new step-level guard model from Hugging Face, audits agent actions before execution, reducing attack success rates by...

1
Research paper
Hugging Face 31 Aug 2026

Hugging Face Researchers Unveil ABot-Recon for Stable Long-Horizon 3D Reconstruction

ABot-Recon, a new streaming 3D reconstruction model from Hugging Face, achieves stable long-horizon performance using only local t...

1
Research paper
Hugging Face 31 Aug 2026

ContextPilot: Teaching Agents Proactive Context Management via Fine-Grained RL

Hugging Face researchers introduce ContextPilot, a framework that enhances long-horizon agent reasoning by expanding context-editi...

1