Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -48,6 +48,39 @@ wsi_dataset = load_dataset("Lab-Rasool/TCGA", "wsi", split="train")
|
|
48 |
molecular_dataset = load_dataset("Lab-Rasool/TCGA", "molecular", split="train")
|
49 |
```
|
50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
51 |
## Dataset Creation
|
52 |
|
53 |
#### Data Collection and Processing
|
|
|
48 |
molecular_dataset = load_dataset("Lab-Rasool/TCGA", "molecular", split="train")
|
49 |
```
|
50 |
|
51 |
+
Example code for loading HF dataset into a PyTorch Dataloader.
|
52 |
+
**Note**: Some embeddings are stored as buffers due to their multi-dimensional shape.
|
53 |
+
|
54 |
+
```python
|
55 |
+
from datasets import load_dataset
|
56 |
+
import os
|
57 |
+
from torch.utils.data import Dataset
|
58 |
+
import numpy as np
|
59 |
+
|
60 |
+
class CustomDataset(Dataset):
|
61 |
+
def __init__(self, hf_dataset):
|
62 |
+
self.hf_dataset = hf_dataset
|
63 |
+
|
64 |
+
def __len__(self):
|
65 |
+
return len(self.hf_dataset)
|
66 |
+
|
67 |
+
def __getitem__(self, idx):
|
68 |
+
hf_item = self.hf_dataset[idx]
|
69 |
+
embedding = np.frombuffer(hf_item["embedding"], dtype=np.float32)
|
70 |
+
embedding_shape = hf_item["embedding_shape"]
|
71 |
+
embedding = embedding.reshape(embedding_shape)
|
72 |
+
return embedding
|
73 |
+
|
74 |
+
if __name__ == "__main__":
|
75 |
+
|
76 |
+
clinical_dataset = load_dataset("Lab-Rasool/TCGA", "clinical", split="train")
|
77 |
+
wsi_dataset = load_dataset("Lab-Rasool/TCGA", "wsi", split="train")
|
78 |
+
|
79 |
+
for index, item in enumerate(clinical_dataset):
|
80 |
+
print(np.frombuffer(item.get("embedding"), dtype=np.float32).reshape(item.get("embedding_shape")).shape)
|
81 |
+
break
|
82 |
+
```
|
83 |
+
|
84 |
## Dataset Creation
|
85 |
|
86 |
#### Data Collection and Processing
|