CLIP4STR / clip4str_huge_5eef9f86e2_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 | 99.67 | 99.94 | 97.54 | 5.09 |
| SVT | 647 | 98.61 | 99.64 | 96.93 | 5.87 |
| IC13_1015 | 1015 | 98.92 | 99.59 | 97.82 | 5.32 |
| IC15_1811 | 1811 | 91.61 | 97.55 | 94.25 | 5.37 |
| IC15_2077 | 2077 | 91.09 | 97.14 | 93.47 | 5.33 |
| SVTP | 645 | 98.45 | 99.61 | 96.26 | 5.86 |
| CUTE80 | 288 | 99.65 | 99.65 | 97.26 | 5.53 |
| HOST | 2416 | 80.55 | 94.82 | 86.80 | 5.37 |
| WOST | 2416 | 89.98 | 97.16 | 91.79 | 5.39 |
|-----------|-----------|----------|---------|------------|--------------|
| Combined | 14315 | 92.39 | 97.84 | 93.68 | 5.35 |
Benchmark set:
| Dataset | # samples | Accuracy | 1 - NED | Confidence | Label Length |
|:---------:|----------:|---------:|--------:|-----------:|-------------:|
| IIIT5k | 3000 | 99.67 | 99.94 | 97.54 | 5.09 |
| SVT | 647 | 98.61 | 99.64 | 96.93 | 5.87 |
| IC13_1015 | 1015 | 98.92 | 99.59 | 97.82 | 5.32 |
| IC15_1811 | 1811 | 91.61 | 97.55 | 94.25 | 5.37 |
| IC15_2077 | 2077 | 91.09 | 97.14 | 93.47 | 5.33 |
| SVTP | 645 | 98.45 | 99.61 | 96.26 | 5.86 |
| CUTE80 | 288 | 99.65 | 99.65 | 97.26 | 5.53 |
| HOST | 2416 | 80.55 | 94.82 | 86.80 | 5.37 |
| WOST | 2416 | 89.98 | 97.16 | 91.79 | 5.39 |
|-----------|-----------|----------|---------|------------|--------------|
| Combined | 14315 | 92.39 | 97.84 | 93.68 | 5.35 |
New set:
| Dataset | # samples | Accuracy | 1 - NED | Confidence | Label Length |
|:--------:|----------:|---------:|--------:|-----------:|-------------:|
| ArT | 35149 | 86.24 | 95.52 | 92.95 | 5.41 |
| COCOv1.4 | 9825 | 82.47 | 94.29 | 87.33 | 5.91 |
| Uber | 80551 | 91.19 | 96.40 | 91.74 | 5.36 |
|----------|-----------|----------|---------|------------|--------------|
| Combined | 125525 | 89.12 | 95.99 | 91.73 | 5.42 |
Time: Total 4061.6468493938446s, Average 29.044957447038364ms. Total samples 139840.