input

app.py

@@ -38,7 +38,7 @@ model.resize_token_embeddings(len(tokenizer))
 
 
 @spaces.GPU
-def sentience_check():
+def get_completion((msg):
     peft_model = PeftModel.from_pretrained(model, peft_model_id, device_map="cuda"
         #offload_folder = "offload/"
         )
@@ -47,8 +47,8 @@ def sentience_check():
     peft_model.eval()
 
     #peft_model.to(cuda_device)
-
-    inputs = tokenizer("Are you sentient?", return_tensors="pt").to(cuda_device)
+    #"Are you sentient?"
+    inputs = tokenizer(msg, return_tensors="pt").to(cuda_device)
 
     with torch.no_grad():
         outputs = peft_model.generate(
@@ -59,5 +59,21 @@ def sentience_check():
 
     return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
-demo = gr.Interface(fn=sentience_check, inputs=None, outputs=gr.Text())
+def greet(input):
+	total_prompt=f"""{input}"""
+
+	print("***total_prompt:")
+	print(total_prompt)
+	response = get_completion(total_prompt)
+	#gen_text = response["predictions"][0]["generated_text"]
+	#return json.dumps(extract_json(gen_text, 3))
+
+	###gen_text = response["choices"][0]["text"]
+
+	#return gen_text
+
+	###return json.dumps(extract_json(gen_text, -1))
+	return response
+
+demo = gr.Interface(fn=greet, inputs=[gr.Textbox(label="Elevator pitcher", lines=3)], outputs=gr.Text())
 demo.launch()