Spaces:
Running
Running
Commit
·
4f78dfa
1
Parent(s):
6c7ba9b
Update save_data.py
Browse files- save_data.py +5 -1
save_data.py
CHANGED
|
@@ -1,4 +1,5 @@
|
|
| 1 |
import os
|
|
|
|
| 2 |
import json
|
| 3 |
import shutil
|
| 4 |
import requests
|
|
@@ -6,11 +7,14 @@ import re as r
|
|
| 6 |
from urllib.request import urlopen
|
| 7 |
from datetime import datetime
|
| 8 |
from datasets import Image
|
|
|
|
| 9 |
from huggingface_hub import Repository, upload_file
|
| 10 |
|
| 11 |
HF_TOKEN = os.environ.get("HF_TOKEN")
|
| 12 |
DATASET_NAME = "OCR-img-to-text"
|
| 13 |
-
DATASET_REPO_URL =
|
|
|
|
|
|
|
| 14 |
DATASET_REPO_ID = "pragnakalp/OCR-img-to-text"
|
| 15 |
print("is none?", HF_TOKEN is None)
|
| 16 |
REPOSITORY_DIR = "data"
|
|
|
|
| 1 |
import os
|
| 2 |
+
import numpy as np
|
| 3 |
import json
|
| 4 |
import shutil
|
| 5 |
import requests
|
|
|
|
| 7 |
from urllib.request import urlopen
|
| 8 |
from datetime import datetime
|
| 9 |
from datasets import Image
|
| 10 |
+
from PIL import Image
|
| 11 |
from huggingface_hub import Repository, upload_file
|
| 12 |
|
| 13 |
HF_TOKEN = os.environ.get("HF_TOKEN")
|
| 14 |
DATASET_NAME = "OCR-img-to-text"
|
| 15 |
+
DATASET_REPO_URL = "https://huggingface.co/datasets/pragnakalp/OCR-img-to-text"
|
| 16 |
+
DATA_FILENAME = "ocr_data.csv"
|
| 17 |
+
DATA_FILE = os.path.join("ocr_data", DATA_FILENAME)
|
| 18 |
DATASET_REPO_ID = "pragnakalp/OCR-img-to-text"
|
| 19 |
print("is none?", HF_TOKEN is None)
|
| 20 |
REPOSITORY_DIR = "data"
|