雷娃 commited on
Commit
b3dfe3c
·
1 Parent(s): 291372b

modify output length

Browse files
Files changed (2) hide show
  1. app.py +3 -3
  2. app_api.py +7 -5
app.py CHANGED
@@ -12,7 +12,7 @@ client = OpenAI(
12
  )
13
 
14
  # define chat function
15
- def chat(user_input, max_new_tokens=20480):
16
  # chat history
17
  messages_template = [
18
  # {"role": "system", "content": "You are Ling, an assistant created by inclusionAI"},
@@ -22,7 +22,7 @@ def chat(user_input, max_new_tokens=20480):
22
  response = client.chat.completions.create(
23
  model="Ling-lite-1.5-250604",
24
  messages=messages_template,
25
- max_tokens=11264,
26
  temperature=0.01,
27
  top_p=1,
28
  )
@@ -47,7 +47,7 @@ with gr.Blocks(css="""
47
  )
48
 
49
  with gr.Row():
50
- max_tokens_slider = gr.Slider(minimum=128, maximum=20480, step=160, label="Generated length")
51
 
52
  # output_box = gr.Textbox(lines=10, label="Response")
53
  output_box = gr.Markdown(label="Response", elem_id="markdown-output")
 
12
  )
13
 
14
  # define chat function
15
+ def chat(user_input, max_tokens=11264):
16
  # chat history
17
  messages_template = [
18
  # {"role": "system", "content": "You are Ling, an assistant created by inclusionAI"},
 
22
  response = client.chat.completions.create(
23
  model="Ling-lite-1.5-250604",
24
  messages=messages_template,
25
+ max_tokens=max_tokens,
26
  temperature=0.01,
27
  top_p=1,
28
  )
 
47
  )
48
 
49
  with gr.Row():
50
+ max_tokens_slider = gr.Slider(minimum=100, maximum=10000, step=100, label="Generated length")
51
 
52
  # output_box = gr.Textbox(lines=10, label="Response")
53
  output_box = gr.Markdown(label="Response", elem_id="markdown-output")
app_api.py CHANGED
@@ -12,21 +12,23 @@ client = OpenAI(
12
  )
13
 
14
  # define chat function
15
- def chat(user_input, max_new_tokens=2048):
16
  # chat history
17
  messages_template = [
18
- # {"role": "system", "content": "You are Ling, an assistant created by inclusionAI"},
19
  {"role": "user", "content": user_input}
20
  ]
21
 
22
  response = client.chat.completions.create(
23
  model="Ling-lite-1.5-250604",
24
  messages=messages_template,
25
- max_tokens=11264,
26
  temperature=0.01,
27
  top_p=1,
28
  )
29
- yield response.choices[0].message.content
 
 
30
 
31
 
32
 
@@ -45,7 +47,7 @@ with gr.Blocks(css="""
45
  )
46
 
47
  with gr.Row():
48
- max_tokens_slider = gr.Slider(minimum=128, maximum=2048, step=16, label="Generated length")
49
 
50
  # output_box = gr.Textbox(lines=10, label="Response")
51
  output_box = gr.Markdown(label="Response", elem_id="markdown-output")
 
12
  )
13
 
14
  # define chat function
15
+ def chat(user_input, max_tokens=11264):
16
  # chat history
17
  messages_template = [
18
+ # {"role": "system", "content": "You are Ling, an assistant created by inclusionAI"},
19
  {"role": "user", "content": user_input}
20
  ]
21
 
22
  response = client.chat.completions.create(
23
  model="Ling-lite-1.5-250604",
24
  messages=messages_template,
25
+ max_tokens=max_tokens,
26
  temperature=0.01,
27
  top_p=1,
28
  )
29
+ resp_text = response.choices[0].message.content
30
+ print(resp_text)
31
+ yield resp_text
32
 
33
 
34
 
 
47
  )
48
 
49
  with gr.Row():
50
+ max_tokens_slider = gr.Slider(minimum=100, maximum=10000, step=100, label="Generated length")
51
 
52
  # output_box = gr.Textbox(lines=10, label="Response")
53
  output_box = gr.Markdown(label="Response", elem_id="markdown-output")