Upload app.py

app.py

@@ -10,19 +10,18 @@ from transformers import AutoTokenizer , AutoModelForSeq2SeqLM , TextIteratorStr
 from threading import Thread
 
 # Load model and tokenizer globally to avoid reloading for every request
-base_model = "Helsinki-NLP/europarl"
+base_model = "google-t5/t5-small"
 model_path = "Mat17892/t5small_enfr_opus"
 
 # Load tokenizer
 tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=True, legacy=False)
 
 # Load the base model (e.g., LLaMA)
-print(os.getenv('huggingface_token'))
-base_model = AutoModelForSeq2SeqLM.from_pretrained(base_model,token = os.getenv('huggingface_token'))
+base_model = AutoModelForSeq2SeqLM.from_pretrained(base_model, token = os.getenv('huggingface_token'))
 
 # Load LoRA adapter
 from peft import PeftModel
-model = PeftModel.from_pretrained(base_model, model_path)
+model = PeftModel.from_pretrained(base_model, model_path, token = os.getenv('huggingface_token'))
 
 def respond(
     message: str,