Update funcs.py

funcs.py

@@ -101,10 +101,14 @@ import torch
 
 # Load model once globally for reuse
 tokenizer = AutoTokenizer.from_pretrained("tiiuae/Falcon-H1-1.5B-Deep-Instruct")
-model = AutoModelForCausalLM.from_pretrained("tiiuae/Falcon-H1-1.5B-Deep-Instruct", torch_dtype=torch.float16, device_map="auto")
-
+use_cuda = torch.cuda.is_available()
+model = AutoModelForCausalLM.from_pretrained(
+    "tiiuae/Falcon-H1-1.5B-Deep-Instruct",
+    torch_dtype=torch.float16 if use_cuda else torch.float32,
+    device_map="auto" if use_cuda else None
+)
 def generate_falcon_response(prompt, max_new_tokens=300):
-    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+    inputs = tokenizer(prompt, return_tensors="pt", truncation=True, max_length=1800).to(model.device)
     outputs = model.generate(
         **inputs,
         max_new_tokens=max_new_tokens,
@@ -115,6 +119,7 @@ def generate_falcon_response(prompt, max_new_tokens=300):
     decoded_output = tokenizer.decode(outputs[0], skip_special_tokens=True)
     return decoded_output[len(prompt):].strip()
 
+
 def summarize_and_recommend(therapy_session_conversation):
 
     session_time = str(datetime.now().strftime("%Y-%m-%d %H:%M:%S"))