ashupanda123 commited on
Commit
902ca20
·
verified ·
1 Parent(s): 6f6c791

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +17 -18
app.py CHANGED
@@ -1,10 +1,8 @@
1
  import gradio as gr
2
- from huggingface_hub import InferenceClient
3
 
4
- """
5
- For more information on huggingface_hub Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
6
- """
7
- client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
8
 
9
 
10
  def respond(
@@ -29,18 +27,20 @@ def respond(
29
 
30
  messages.append({"role": "user", "content": template_instruction})
31
 
32
- response = ""
 
 
33
 
34
- for message in client.chat_completion(
35
- messages,
36
- max_tokens=max_tokens,
37
- stream=True,
38
- temperature=temperature,
39
- top_p=top_p,
40
- ):
41
- token = message.choices[0].delta.content
42
 
43
- response += token
 
 
 
 
 
 
 
44
  yield response
45
 
46
  """
@@ -51,7 +51,7 @@ demo = gr.ChatInterface(
51
  additional_inputs=[
52
  gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
53
  gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
54
- gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
55
  gr.Slider(
56
  minimum=0.1,
57
  maximum=1.0,
@@ -63,6 +63,5 @@ demo = gr.ChatInterface(
63
  ],
64
  )
65
 
66
-
67
  if __name__ == "__main__":
68
- demo.launch(share=True)
 
1
  import gradio as gr
2
+ from transformers import pipeline
3
 
4
+ # Initialize the text generation pipeline
5
+ pipe = pipeline("text-generation", model="HuggingFaceH4/zephyr-7b-beta")
 
 
6
 
7
 
8
  def respond(
 
27
 
28
  messages.append({"role": "user", "content": template_instruction})
29
 
30
+ # Convert messages to a single prompt
31
+ user_messages = "\n".join([msg["content"] for msg in messages if msg["role"] == "user"])
32
+ system_message_content = next(msg["content"] for msg in messages if msg["role"] == "system")
33
 
34
+ prompt = f"{system_message_content}\n{user_messages}"
 
 
 
 
 
 
 
35
 
36
+ response = ""
37
+ # Generate text using the pipeline
38
+ outputs = pipe(prompt, max_length=max_tokens, do_sample=True, temperature=temperature, top_p=top_p)
39
+ generated_text = outputs[0]['generated_text']
40
+
41
+ # Yield the response token by token
42
+ for token in generated_text.split():
43
+ response += token + " "
44
  yield response
45
 
46
  """
 
51
  additional_inputs=[
52
  gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
53
  gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
54
+ gr.Slider(minimum=0.1, maximum 4.0, value=0.7, step=0.1, label="Temperature"),
55
  gr.Slider(
56
  minimum=0.1,
57
  maximum=1.0,
 
63
  ],
64
  )
65
 
 
66
  if __name__ == "__main__":
67
+ demo.launch(share=True)