paper
-with-
me
Papers
Browse State-of-the-Art
Datasets
Methods
AI Agents
Trends
Digest
🌙
image-sentence alignment
벤치마크
image-sentence alignment on VALSE actant swap
14개 결과 ·
⬇ CSV
·
JSON
pairwise accuracy
44.4
52.52
60.65
68.78
76.9
2021-12
2026-09
GPT2 — 76.9 (2021-12-14)
GPT1 — 72.2 (2021-12-14)
CLIP — 68.6 (2021-12-14)
ViLBERT — 68.3 (2021-12-14)
ViLBERT 12-in-1 — 58.9 (2021-12-14)
LXMERT — 45.8 (2021-12-14)
VisualBERT — 44.4 (2021-12-14)
GPT2 — 76.9 (2021-12-14)
GPT1 — 72.2 (2021-12-14)
CLIP — 68.6 (2021-12-14)
ViLBERT — 68.3 (2021-12-14)
ViLBERT 12-in-1 — 58.9 (2021-12-14)
LXMERT — 45.8 (2021-12-14)
VisualBERT — 44.4 (2021-12-14)
GPT2 — 76.9 (2021-12-14)
2021-12-14 — GPT2: pairwise accuracy 76.9
Rank
Model
pairwise accuracy
Accuracy (%)
Paper
Code
Year
1
GPT2
76.9
–
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
2
GPT1
72.2
–
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
3
CLIP
68.6
–
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
4
ViLBERT
68.3
50.4
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
5
ViLBERT 12-in-1
58.9
52.2
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
6
LXMERT
45.8
48.5
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
7
VisualBERT
44.4
49.7
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
8
GPT2
76.9
–
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
9
GPT1
72.2
–
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
10
CLIP
68.6
–
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
11
ViLBERT
68.3
50.4
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
12
ViLBERT 12-in-1
58.9
52.2
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
13
LXMERT
45.8
48.5
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
14
VisualBERT
44.4
49.7
VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
heidelberg-nlp/valse
2021
1–14 / 14
페이지당
10
20
50
100