Update README.md

README.md

@@ -70,7 +70,7 @@ generated_tokens = output.size(1) - prompt_tokens
 time_per_token = generated_tokens/total_time
 print(f"Generated {generated_tokens} tokens in {total_time:.3f} s ({time_per_token:.3f} tok/s)")
 
-response = processor.tokenizer.decode(output[0, prompt_tokens:])
+response = processor.tokenizer.decode(output[0, prompt_tokens:], skip_special_tokens=True)
 print(response)
 
 torch.cuda.empty_cache()