Update app.py

app.py

@@ -46,9 +46,8 @@ h1 {
 """
 
 # Load the tokenizer and model
-with torch.autocast(device_type='cpu'):
-    tokenizer = AutoTokenizer.from_pretrained("CyberNative-AI/Colibri_8b_v0.1_gptq_128_4bit")
-    model = AutoGPTQForCausalLM.from_quantized("CyberNative-AI/Colibri_8b_v0.1_gptq_128_4bit", dtype=torch.bfloat16, device="cpu")
+tokenizer = AutoTokenizer.from_pretrained("CyberNative-AI/Colibri_8b_v0.1_gptq_128_4bit")
+model = AutoGPTQForCausalLM.from_quantized("CyberNative-AI/Colibri_8b_v0.1_gptq_128_4bit", dtype=torch.float32, device="cpu")
 
 terminators = [
     tokenizer.eos_token_id,