In many cases, the HF space is operated in a way where the main program runs as a server process and the GUI accesses it. If the server process crashes, I think that error will occur. The cause is various, but it is when you use up all the RAM, including the virtual RAM.
However, there is also the possibility of a new HF bug.
Small update. I created a new space in order to test this. It worked the first time. I deleted a line of code and then it got stuck in the building process. I tried a restart and the same thing occured. No logs were provided either.
Here’s the code in the app.py file:
from pyngrok import ngrok
import gradio as gr
def generate_response(instruction):
“”“Generates a response using your fine-tuned model.”“”
##FastLanguageModel.for_inference(model) # Enable native 2x faster inference within the function
prompt = f"“”### Instruction:
Answer the provided question with the knowledge provided to you
####### Question:
{instruction}
####### Answer:
“”"
return prompt
So the deletion of the ngrok library from the requirements and app files solved the issue. This is suprising to me because it worked before with that. Thanks nonetheless