i

app.py

@@ -26,10 +26,10 @@ def get_completion(prompt, model_name="dolly-v0-70m"):
 	#elif gpu_family == "a10" or gpu_family == "a100":
 	#pipeline_kwargs['torch_dtype'] = "bfloat16"
 
-	pipeline_kwargs['max_new_tokens'] = 10
+	pipeline_kwargs['max_new_tokens'] = 100
 	#pipeline_kwargs['temperature'] = float("inf")
 	#pipeline_kwargs['top_k'] = 1
-	pipeline_kwargs['top_p'] = 0.01
+	#pipeline_kwargs['top_p'] = 0.01
 
 	# Use the model to generate responses for each of the instructions above.
 	for instruction in instructions: