test api

app.py

@@ -3,6 +3,7 @@ import spaces
 from transformers import pipeline
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
+import json
 
 model_name = "sapienzanlp/Minerva-7B-instruct-v1.0"
 model = AutoModelForCausalLM.from_pretrained(
@@ -38,8 +39,14 @@ def generate(prompts: list[str]) -> tuple[list[str], list[dict[str, float]]]:
 
     return responses, classifier([text + "[SEP]" + response for text, response in zip(texts, responses)])
 
+def scores(body: str) -> tuple[list[str], list[dict[str, float]]]:
+    data = json.loads(body)
+    return data
+
+
 with gr.Blocks() as demo:
     gr.Markdown("Welcome")
-    gr.api(generate, api_name="score", batch=True)
+    gr.api(scores, api_name="score", batch=True, max_batch_size=25)
 
+demo.queue()
 demo.launch()