Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -18,14 +18,14 @@ import re
|
|
| 18 |
GEMINI_API_KEY = "AIzaSyDGpNmvskXEAeOH6hG_BtT8GR043tMREYk" # Replace with actual API key
|
| 19 |
genai.configure(api_key=GEMINI_API_KEY)
|
| 20 |
|
| 21 |
-
# Set Hugging Face cache directory
|
| 22 |
CACHE_DIR = "/tmp/huggingface"
|
| 23 |
os.environ["HF_HOME"] = CACHE_DIR
|
| 24 |
os.environ["HF_DATASETS_CACHE"] = f"{CACHE_DIR}/datasets"
|
| 25 |
os.environ["TRANSFORMERS_CACHE"] = f"{CACHE_DIR}/transformers"
|
| 26 |
os.environ["HF_HUB_CACHE"] = f"{CACHE_DIR}/hub"
|
| 27 |
|
| 28 |
-
# Ensure directories exist
|
| 29 |
for path in [os.environ["HF_DATASETS_CACHE"], os.environ["TRANSFORMERS_CACHE"], os.environ["HF_HUB_CACHE"]]:
|
| 30 |
os.makedirs(path, exist_ok=True)
|
| 31 |
|
|
@@ -36,9 +36,9 @@ CORS(app)
|
|
| 36 |
# RAG Model Initialization
|
| 37 |
print("\U0001F680 Initializing RAG System...")
|
| 38 |
|
| 39 |
-
# Load medical guidelines dataset
|
| 40 |
print("\U0001F4C2 Loading dataset...")
|
| 41 |
-
dataset = load_dataset("epfl-llm/guidelines", split="train", cache_dir=
|
| 42 |
TITLE_COL = "title"
|
| 43 |
CONTENT_COL = "clean_text"
|
| 44 |
|
|
|
|
| 18 |
GEMINI_API_KEY = "AIzaSyDGpNmvskXEAeOH6hG_BtT8GR043tMREYk" # Replace with actual API key
|
| 19 |
genai.configure(api_key=GEMINI_API_KEY)
|
| 20 |
|
| 21 |
+
# Set Hugging Face cache directory explicitly
|
| 22 |
CACHE_DIR = "/tmp/huggingface"
|
| 23 |
os.environ["HF_HOME"] = CACHE_DIR
|
| 24 |
os.environ["HF_DATASETS_CACHE"] = f"{CACHE_DIR}/datasets"
|
| 25 |
os.environ["TRANSFORMERS_CACHE"] = f"{CACHE_DIR}/transformers"
|
| 26 |
os.environ["HF_HUB_CACHE"] = f"{CACHE_DIR}/hub"
|
| 27 |
|
| 28 |
+
# Ensure cache directories exist
|
| 29 |
for path in [os.environ["HF_DATASETS_CACHE"], os.environ["TRANSFORMERS_CACHE"], os.environ["HF_HUB_CACHE"]]:
|
| 30 |
os.makedirs(path, exist_ok=True)
|
| 31 |
|
|
|
|
| 36 |
# RAG Model Initialization
|
| 37 |
print("\U0001F680 Initializing RAG System...")
|
| 38 |
|
| 39 |
+
# Load medical guidelines dataset with explicit cache directory
|
| 40 |
print("\U0001F4C2 Loading dataset...")
|
| 41 |
+
dataset = load_dataset("epfl-llm/guidelines", split="train", cache_dir="/tmp/huggingface/datasets")
|
| 42 |
TITLE_COL = "title"
|
| 43 |
CONTENT_COL = "clean_text"
|
| 44 |
|