CLIP4STR / clip4str_large_3c9d881b88_log.txt
mzhaoshuai's picture
Upload CLIP4STR Pre-trained on DataComp-1B, LAION-2B, and DFN-5B
a70545b verified
Benchmark (Subset) set:
| Dataset | # samples | Accuracy | 1 - NED | Confidence | Label Length |
|:---------:|----------:|---------:|--------:|-----------:|-------------:|
| IIIT5k | 3000 | 99.60 | 99.88 | 97.77 | 5.09 |
| SVT | 647 | 98.61 | 99.63 | 97.18 | 5.87 |
| IC13_1015 | 1015 | 99.01 | 99.63 | 97.73 | 5.32 |
| IC15_1811 | 1811 | 91.88 | 97.52 | 94.77 | 5.37 |
| IC15_2077 | 2077 | 91.43 | 97.33 | 94.19 | 5.33 |
| SVTP | 645 | 98.14 | 99.44 | 96.42 | 5.86 |
| CUTE80 | 288 | 99.65 | 99.94 | 97.37 | 5.54 |
| HOST | 2416 | 81.08 | 94.91 | 88.33 | 5.38 |
| WOST | 2416 | 90.60 | 97.26 | 93.04 | 5.39 |
|-----------|-----------|----------|---------|------------|--------------|
| Combined | 14315 | 92.64 | 97.88 | 94.38 | 5.35 |
Benchmark set:
| Dataset | # samples | Accuracy | 1 - NED | Confidence | Label Length |
|:---------:|----------:|---------:|--------:|-----------:|-------------:|
| IIIT5k | 3000 | 99.60 | 99.88 | 97.77 | 5.09 |
| SVT | 647 | 98.61 | 99.63 | 97.18 | 5.87 |
| IC13_1015 | 1015 | 99.01 | 99.63 | 97.73 | 5.32 |
| IC15_1811 | 1811 | 91.88 | 97.52 | 94.77 | 5.37 |
| IC15_2077 | 2077 | 91.43 | 97.33 | 94.19 | 5.33 |
| SVTP | 645 | 98.14 | 99.44 | 96.42 | 5.86 |
| CUTE80 | 288 | 99.65 | 99.94 | 97.37 | 5.54 |
| HOST | 2416 | 81.08 | 94.91 | 88.33 | 5.38 |
| WOST | 2416 | 90.60 | 97.26 | 93.04 | 5.39 |
|-----------|-----------|----------|---------|------------|--------------|
| Combined | 14315 | 92.64 | 97.88 | 94.38 | 5.35 |
New set:
| Dataset | # samples | Accuracy | 1 - NED | Confidence | Label Length |
|:--------:|----------:|---------:|--------:|-----------:|-------------:|
| ArT | 35149 | 86.43 | 95.62 | 93.62 | 5.41 |
| COCOv1.4 | 9825 | 82.68 | 94.43 | 88.30 | 5.92 |
| Uber | 80551 | 92.20 | 96.79 | 92.85 | 5.37 |
|----------|-----------|----------|---------|------------|--------------|
| Combined | 125525 | 89.84 | 96.28 | 92.71 | 5.42 |
Time: Total 2185.2176456451416s, Average 15.626556390482992ms. Total samples 139840.