Abob1ch commited on
Commit
103f1a6
·
verified ·
1 Parent(s): dd367fe

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +23 -18
app.py CHANGED
@@ -1,11 +1,15 @@
 
 
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("SicariusSicariiStuff/LLAMA-3_8B_Unaligned_BETA")
8
 
 
 
 
 
9
 
10
  def respond(
11
  message,
@@ -26,27 +30,26 @@ def respond(
26
  messages.append({"role": "user", "content": message})
27
 
28
  response = ""
29
-
30
- for message in client.chat_completion(
31
- messages,
32
  max_tokens=max_tokens,
33
  stream=True,
34
  temperature=temperature,
35
  top_p=top_p,
 
36
  ):
37
  token = message.choices[0].delta.content
38
-
39
  response += token
40
  yield response
 
 
41
 
42
-
43
- """
44
- For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
45
- """
46
  demo = gr.ChatInterface(
47
  respond,
48
  additional_inputs=[
49
- gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
50
  gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
51
  gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
52
  gr.Slider(
@@ -54,11 +57,13 @@ demo = gr.ChatInterface(
54
  maximum=1.0,
55
  value=0.95,
56
  step=0.05,
57
- label="Top-p (nucleus sampling)",
58
  ),
 
59
  ],
 
 
 
60
  )
61
-
62
-
63
  if __name__ == "__main__":
64
- demo.launch()
 
1
+ #refer llama recipes for more info https://github.com/huggingface/huggingface-llama-recipes/blob/main/inference-api.ipynb
2
+ #huggingface-llama-recipes : https://github.com/huggingface/huggingface-llama-recipes/tree/main
3
  import gradio as gr
4
+ from openai import OpenAI
5
+ import os
6
 
7
+ ACCESS_TOKEN = os.getenv("HF_TOKEN")
 
 
 
8
 
9
+ client = OpenAI(
10
+ base_url="https://api-inference.huggingface.co/v1/",
11
+ api_key=ACCESS_TOKEN,
12
+ )
13
 
14
  def respond(
15
  message,
 
30
  messages.append({"role": "user", "content": message})
31
 
32
  response = ""
33
+
34
+ for message in client.chat.completions.create(
35
+ model="SicariusSicariiStuff/LLAMA-3_8B_Unaligned_BETA",
36
  max_tokens=max_tokens,
37
  stream=True,
38
  temperature=temperature,
39
  top_p=top_p,
40
+ messages=messages,
41
  ):
42
  token = message.choices[0].delta.content
43
+
44
  response += token
45
  yield response
46
+
47
+ chatbot = gr.Chatbot(height=600)
48
 
 
 
 
 
49
  demo = gr.ChatInterface(
50
  respond,
51
  additional_inputs=[
52
+ gr.Textbox(value="", 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(
 
57
  maximum=1.0,
58
  value=0.95,
59
  step=0.05,
60
+ label="Top-P",
61
  ),
62
+
63
  ],
64
+ fill_height=True,
65
+ chatbot=chatbot,
66
+ theme="Nymbo/Alyx_Theme",
67
  )
 
 
68
  if __name__ == "__main__":
69
+ demo.launch()