update

Dockerfile

@@ -4,7 +4,9 @@ WORKDIR /app
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 RUN mkdir -p /app/cache /app/nltk_data
-
+RUN mkdir -p /app/ChatbotCheckpoint 
+RUN chmod 777 /app/ChatbotCheckpoint  
+ENV CHECKPOINT_DIR=/app/ChatbotCheckpoint
 ENV HF_HOME=/app/cache
 ENV NLTK_DATA=/app/nltk_data
 RUN chmod -R 777 /app/cache /app/nltk_data

inference.py

@@ -3,22 +3,16 @@ from dataset import clean
 import re
 import gdown
 
-import os
-import gdown
-
 def load_tokenizer_model():
-    cache_dir = "/tmp/ChatbotCheckpoint"  
+    cache_dir="ChatbotCheckpoint"
     folder_url = "https://drive.google.com/drive/folders/1DDJ9t-HfMrf6OLYim5bVrP20QgyOZahc"
-    os.makedirs(cache_dir, exist_ok=True)  
-
-    gdown.download_folder(folder_url, output=cache_dir)
-    model_name = cache_dir  # Update to the cache directory
-    model = GPT2LMHeadModel.from_pretrained(model_name, cache_dir=cache_dir)
+    gdown.download_folder(folder_url, output="ChatbotCheckpoint")
+    model_name="ChatbotCheckpoint"
+    model = GPT2LMHeadModel.from_pretrained(model_name,cache_dir=cache_dir)
     model.eval()
-    tokenizer = GPT2Tokenizer.from_pretrained(model_name)
-    tokenizer.pad_token = tokenizer.eos_token
-    return tokenizer, model
-
+    tokenizer=GPT2Tokenizer.from_pretrained(model_name)
+    tokenizer.pad_token=tokenizer.eos_token
+    return tokenizer,model
 
 def generate_answer(query):
     tokenizer,model=load_tokenizer_model()