Update app.py

app.py

@@ -4,7 +4,8 @@ from threading import Thread
 from transformers import StoppingCriteria, StoppingCriteriaList
 import torch
 import spaces
-import os 
+import os
+
 model_name = "microsoft/Phi-3-medium-128k-instruct"
 from transformers import AutoModelForCausalLM, AutoTokenizer
 model = AutoModelForCausalLM.from_pretrained(model_name, device_map='cuda', torch_dtype=torch.float16, trust_remote_code=True)
@@ -17,22 +18,22 @@ class StopOnTokens(StoppingCriteria):
             if input_ids[0][-1] == stop_id:
                 return True
         return False
-@spaces.GPU()
-def predict(message, history):
+
+@spaces.GPU(duration=120)
+def predict(message, history, temperature, max_tokens, top_p, top_k):
     history_transformer_format = history + [[message, ""]]
     stop = StopOnTokens()
     messages = "".join(["".join(["\n<|end|>\n<|user|>\n"+item[0], "\n<|end|>\n<|assistant|>\n"+item[1]]) for item in history_transformer_format])
-    #messages = "".join(["".join(["<user>"+item[0], "<output>"+item[1]]) for item in history_transformer_format])
     model_inputs = tokenizer([messages], return_tensors="pt").to("cuda")
     streamer = TextIteratorStreamer(tokenizer, timeout=10., skip_prompt=True, skip_special_tokens=True)
     generate_kwargs = dict(
         model_inputs,
         streamer=streamer,
-        max_new_tokens=8192,
+        max_new_tokens=max_tokens,
         do_sample=True,
-        top_p=0.8,
-        top_k=40,
-        temperature=0.9,
+        top_p=top_p,
+        top_k=top_k,
+        temperature=temperature,
         stopping_criteria=StoppingCriteriaList([stop])
     )
     t = Thread(target=model.generate, kwargs=generate_kwargs)
@@ -43,5 +44,14 @@ def predict(message, history):
             partial_message += new_token
         yield partial_message
 
-demo = gr.ChatInterface(fn=predict, examples=["Write me a python snake game code", "Write me a ping pong game code"], title="Phi-3-medium-128k-instruct")
-demo.launch(share=True)
+demo = gr.ChatInterface(
+    fn=predict,
+    title="Phi-3-medium-128k-instruct",
+    additional_inputs=[
+        gr.Slider(0.1, 0.9, value=0.7, label="Temperature"),
+        gr.Slider(512, 8192, value=4096, label="Max Tokens"),
+        gr.Slider(0.1, 0.9, value=0.7, label="top_p"),
+        gr.Slider(10, 90, value=40, label="top_k"),
+    ]
+)
+demo.launch(share=True)