README: 11 datasets
Browse files
README.md
CHANGED
|
@@ -60,6 +60,18 @@ configs:
|
|
| 60 |
path: data/handan_ffa2/validation-*.parquet
|
| 61 |
- split: test
|
| 62 |
path: data/handan_ffa2/test-*.parquet
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
- config_name: tongrenHandan
|
| 64 |
data_files:
|
| 65 |
- split: train
|
|
@@ -96,7 +108,7 @@ configs:
|
|
| 96 |
|
| 97 |
# OphGen — curated institutional ophthalmic datasets
|
| 98 |
|
| 99 |
-
Curated copies of
|
| 100 |
images as stored in the OphGen image store (de-identified, de-duplicated, native resolution), the train / validation / test
|
| 101 |
split used in the project, six captions per image (short / medium / long × English / Chinese) and all original task annotations.
|
| 102 |
|
|
@@ -106,8 +118,8 @@ split used in the project, six captions per image (short / medium / long × Engl
|
|
| 106 |
|
| 107 |
```python
|
| 108 |
from datasets import load_dataset
|
| 109 |
-
ds = load_dataset("Kaphathy/OphGen-Private", "
|
| 110 |
-
side = load_dataset("Kaphathy/OphGen-Private", "
|
| 111 |
pre = load_dataset("Kaphathy/OphGen-Private", "all") # every dataset (default config), e.g. for pretraining
|
| 112 |
```
|
| 113 |
|
|
@@ -123,11 +135,12 @@ In `all`, an image released by several datasets appears once per dataset (same `
|
|
| 123 |
| [FQ_Datasets](cards/FQ_Datasets.md) | FQ_Datasets | CFP | 2,681 | 402 | 802 | 1 |
|
| 124 |
| [handan_ffa](cards/handan_ffa.md) | handan_ffa | FFA, CFP | 26,560 | 1,478 | 1,471 | 0 |
|
| 125 |
| [handan_ffa2](cards/handan_ffa2.md) | handan_ffa2 | FFA | 17,465 | 980 | 991 | 0 |
|
|
|
|
| 126 |
| [tongrenHandan](cards/tongrenHandan.md) | tongrenHandan | CFP | 7,012 | 987 | 1,974 | 1 |
|
| 127 |
| [trhd3_FRD_Cls66_Train](cards/trhd3_FRD_Cls66_Train.md) | trhd3_FRD_Cls66_Train | CFP | 7,780 | 1,106 | 2,176 | 1 |
|
| 128 |
| [yinhai_ffa](cards/yinhai_ffa.md) | yinhai_ffa | FFA | 812 | 34 | 52 | 0 |
|
| 129 |
| [YinhaiUBM](cards/YinhaiUBM.md) | YinhaiUBM | UBM | 75,720 | 4,012 | 3,857 | 0 |
|
| 130 |
-
| **Total** | | |
|
| 131 |
|
| 132 |
Split rules: `official` = official split; `official_val_from_train` = 10% of the official training set (by patient) used as
|
| 133 |
validation; `labeled_7_1_2` = patient-level 7:1:2 split with a fixed seed (no official split); `pretrain_90_5_5` = 90:5:5 split
|
|
|
|
| 60 |
path: data/handan_ffa2/validation-*.parquet
|
| 61 |
- split: test
|
| 62 |
path: data/handan_ffa2/test-*.parquet
|
| 63 |
+
- config_name: tongren16G95DFundus
|
| 64 |
+
data_files:
|
| 65 |
+
- split: train
|
| 66 |
+
path: data/tongren16G95DFundus/train-*.parquet
|
| 67 |
+
- split: validation
|
| 68 |
+
path: data/tongren16G95DFundus/validation-*.parquet
|
| 69 |
+
- split: test
|
| 70 |
+
path: data/tongren16G95DFundus/test-*.parquet
|
| 71 |
+
- config_name: tongren16G95DFundus-sidecar
|
| 72 |
+
data_files:
|
| 73 |
+
- split: train
|
| 74 |
+
path: sidecar/tongren16G95DFundus/*.parquet
|
| 75 |
- config_name: tongrenHandan
|
| 76 |
data_files:
|
| 77 |
- split: train
|
|
|
|
| 108 |
|
| 109 |
# OphGen — curated institutional ophthalmic datasets
|
| 110 |
|
| 111 |
+
Curated copies of 11 institutional ophthalmic imaging datasets, prepared for the OphGen ophthalmic generative-model project:
|
| 112 |
images as stored in the OphGen image store (de-identified, de-duplicated, native resolution), the train / validation / test
|
| 113 |
split used in the project, six captions per image (short / medium / long × English / Chinese) and all original task annotations.
|
| 114 |
|
|
|
|
| 118 |
|
| 119 |
```python
|
| 120 |
from datasets import load_dataset
|
| 121 |
+
ds = load_dataset("Kaphathy/OphGen-Private", "tongren16G95DFundus") # one dataset, its own splits and labels
|
| 122 |
+
side = load_dataset("Kaphathy/OphGen-Private", "tongren16G95DFundus-sidecar") # its other annotation files (path, size, bytes), if any
|
| 123 |
pre = load_dataset("Kaphathy/OphGen-Private", "all") # every dataset (default config), e.g. for pretraining
|
| 124 |
```
|
| 125 |
|
|
|
|
| 135 |
| [FQ_Datasets](cards/FQ_Datasets.md) | FQ_Datasets | CFP | 2,681 | 402 | 802 | 1 |
|
| 136 |
| [handan_ffa](cards/handan_ffa.md) | handan_ffa | FFA, CFP | 26,560 | 1,478 | 1,471 | 0 |
|
| 137 |
| [handan_ffa2](cards/handan_ffa2.md) | handan_ffa2 | FFA | 17,465 | 980 | 991 | 0 |
|
| 138 |
+
| [tongren16G95DFundus](cards/tongren16G95DFundus.md) | tongren16G95DFundus | CFP | 172,345 | 24,419 | 49,171 | 1 |
|
| 139 |
| [tongrenHandan](cards/tongrenHandan.md) | tongrenHandan | CFP | 7,012 | 987 | 1,974 | 1 |
|
| 140 |
| [trhd3_FRD_Cls66_Train](cards/trhd3_FRD_Cls66_Train.md) | trhd3_FRD_Cls66_Train | CFP | 7,780 | 1,106 | 2,176 | 1 |
|
| 141 |
| [yinhai_ffa](cards/yinhai_ffa.md) | yinhai_ffa | FFA | 812 | 34 | 52 | 0 |
|
| 142 |
| [YinhaiUBM](cards/YinhaiUBM.md) | YinhaiUBM | UBM | 75,720 | 4,012 | 3,857 | 0 |
|
| 143 |
+
| **Total** | | | 415,785 | 39,633 | 67,141 | |
|
| 144 |
|
| 145 |
Split rules: `official` = official split; `official_val_from_train` = 10% of the official training set (by patient) used as
|
| 146 |
validation; `labeled_7_1_2` = patient-level 7:1:2 split with a fixed seed (no official split); `pretrain_90_5_5` = 90:5:5 split
|