AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

Teaching CLIP to Count to Ten

Roni Paiss, Ariel Ephrat, Omer Tov, Shiran Zada, Inbar Mosseri, Michal Irani, Tali Dekel

ICCV 2023 · Vision-language and multimodal › CLIP and zero-shot · approach: Fine-grained improvements to CLIP · 121 citations (OpenAlex)

This paper tried a common change: Add a contrastive loss

add counting contrastive loss with counterfactual captions on a filtered counting set · CountBench · accuracy: 31.67 -> 75.93 Improved

What 1,317 accepted papers report for this change

See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.

Read the paper

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
add counting contrastive loss with counterfactual captions on a filtered counting setCLIP-B/32 · CountBenchaccuracy: 31.67 -> 75.93Improved
add counting contrastive loss with counterfactual captions on a filtered counting setBASIC-S · CountBenchaccuracy: 17.97 -> 69.02Improved
add counting loss L_count on top of finetuning on the filtered counting set (ablation)CLIP-B/32 · CountBenchaccuracy: 44.26 -> 75.93Improved
use restrictive detector-based count filtering instead of naive number filtering (ablation)CLIP-B/32 · CountBenchaccuracy: 49.81 -> 75.93Improved
use restrictive filtering instead of naive filtering (ablation)BASIC-S · CountBenchaccuracy: 28.68 -> 69.02Improved
finetune on filtered counting set without counting loss (ablation)BASIC-S · CountBenchaccuracy: 22.75 -> 30.59Improved
apply counting finetuning to CLIP-B/32CLIP-B/32 · zero-shot classification benchmarksaccuracy: ImageNet 62.93 -> 64.06; EuroSAT 45.23 -> 37.69Mixed
increase counting fraction of batch p from 1/32 to 1/8 (ablation)CLIP-B/32 · CountBenchaccuracy: 75.93 -> 70.19Hurt
increase counting loss weight lambda from 1 to 5 or 10 (ablation)CLIP-B/32 · CountBenchaccuracy: 75.93 -> 73.15 (lambda 5), 72.59 (lambda 10)Hurt
lower counting loss weight lambda to 0.1 (ablation)CLIP-B/32 · CountBenchaccuracy: 75.93 -> 69.44Hurt
condition Imagen on counting-aware CLIP instead of baseline CLIPImagen · DrawBench countingaccuracy: 24.12 -> 40.35Improved
condition Imagen on counting-aware CLIP for numbers 2-10Imagen · CIFAR10 prompts 2-10accuracy: 20.00 -> 50.18Improved

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