t

app.py

@@ -110,7 +110,8 @@ def f(context, temperature, top_p, max_length, model_idx, SPACE_VERIFICATION_KEY
                 while len(generated_text) < max_length:#(max_length > 0): NOTE NOTE commented out this line and added new check
                     # context becomes the previous generated context
                     # NOTE I've set return_full_text to false, see how this plays out
-                    payload = {"inputs": context, "parameters": {"return_full_text":False, "max_new_tokens": max_length>250 and 250 or max_length, "temperature": temperature, "top_p": top_p}}
+                    # change max_length from max_length>250 and 250 or max_length to 250
+                    payload = {"inputs": context, "parameters": {"return_full_text":False, "max_new_tokens": 250, "temperature": temperature, "top_p": top_p}}
                     response = requests.request("POST", API_URL, data=json.dumps(payload), headers=headers)
                     context = json.loads(response.content.decode("utf-8"))#[0]['generated_text']
                     # context = get_generated_text(generated_context)
@@ -121,7 +122,7 @@ def f(context, temperature, top_p, max_length, model_idx, SPACE_VERIFICATION_KEY
                     # else:
                     #     context = context[0][0]['generated_text']
                         
-                    context = get_generated_text(context)
+                    context = get_generated_text(context).strip()
 
                     generated_text += context
                     # max_length -= 250