Qwen 2.5 VL 7B vs.Llama 3 11B Vision
Passed:
Qwen 2.5 VL 7B 56.1% (46/82)
Llama 3 11B Vision 52.4% (43/82)
Average request time:
Qwen 2.5 VL 7B 10.18s
Llama 3 11B Vision 20.50s
Passed:
Qwen 2.5 VL 7B 56.1% (46/82)
Llama 3 11B Vision 52.4% (43/82)
Average request time:
Qwen 2.5 VL 7B 10.18s
Llama 3 11B Vision 20.50s
Summary
Annotation Understanding
Qwen 2.5 VL 7B
2 / 2
Llama 3 11B Vision
1 / 2
CAPTCHA
Qwen 2.5 VL 7B
2 / 2
Llama 3 11B Vision
0 / 2
Color Identification
Qwen 2.5 VL 7B
1 / 1
Llama 3 11B Vision
0 / 1
Counting nodes
Qwen 2.5 VL 7B
0 / 1
Llama 3 11B Vision
0 / 1
Defect Detection
Qwen 2.5 VL 7B
10 / 15
Llama 3 11B Vision
5 / 15
Document Understanding
Qwen 2.5 VL 7B
5 / 9
Llama 3 11B Vision
4 / 9
Find an author
Qwen 2.5 VL 7B
0 / 1
Llama 3 11B Vision
0 / 1
Localization
Qwen 2.5 VL 7B
0 / 1
Llama 3 11B Vision
0 / 1
OCR
Qwen 2.5 VL 7B
7 / 9
Llama 3 11B Vision
7 / 9
Object Counting
Qwen 2.5 VL 7B
1 / 10
Llama 3 11B Vision
4 / 10
Object Detection
Qwen 2.5 VL 7B
1 / 2
Llama 3 11B Vision
1 / 2
Object Measurement
Qwen 2.5 VL 7B
0 / 2
Llama 3 11B Vision
0 / 2
Object Understanding
Qwen 2.5 VL 7B
7 / 10
Llama 3 11B Vision
5 / 10
Receipt Reading
Qwen 2.5 VL 7B
1 / 1
Llama 3 11B Vision
1 / 1
Sign Understanding
Qwen 2.5 VL 7B
2 / 2
Llama 3 11B Vision
0 / 2
Spatial Relations
Qwen 2.5 VL 7B
10 / 17
Llama 3 11B Vision
14 / 17
Web Action Understanding
Qwen 2.5 VL 7B
3 / 4
Llama 3 11B Vision
4 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!