Spaces:
Tonic
/
Running on Zero

tonic commited on
Commit
f4b9dcc
1 Parent(s): 1b4f65b

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +11 -25
app.py CHANGED
@@ -2,7 +2,7 @@ import spaces
2
  import gradio as gr
3
  import torch
4
  import transformers
5
- from transformers import AutoModelForCausalLM, AutoTokenizer
6
  import os
7
 
8
  title = """# Welcome to 🌟Tonic's✨StarCoder
@@ -14,28 +14,13 @@ default_system_prompt = """SYSTEM: You are an AI that code. Answer with code."""
14
 
15
  model_path = "bigcode/starcoder2-15b"
16
 
17
-
18
  hf_token = os.getenv("HF_TOKEN")
19
  if not hf_token:
20
  raise ValueError("Hugging Face token not found. Please set the HF_TOKEN environment variable.")
21
 
22
- model = AutoModelForCausalLM.from_pretrained(
23
- model_path,
24
- torch_dtype=torch.bfloat16,
25
- device_map="auto",
26
- # trust_remote_code=True,
27
- token=hf_token,
28
- )
29
-
30
- tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
31
-
32
- # import gradio as gr
33
- # from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
34
-
35
- # checkpoint = "bigcode/starcoder2-15b"
36
- # quantization_config = BitsAndBytesConfig(load_in_8bit=True)
37
- # tokenizer = AutoTokenizer.from_pretrained(checkpoint)
38
- # model = AutoModelForCausalLM.from_pretrained(checkpoint, quantization_config=quantization_config).to("cuda")
39
  @spaces.GPU
40
  def generate_text(prompt, temperature, max_length):
41
  inputs = tokenizer.encode(prompt, return_tensors="pt").to("cuda")
@@ -45,14 +30,12 @@ def generate_text(prompt, temperature, max_length):
45
  def gradio_app():
46
  with gr.Blocks() as demo:
47
  gr.Markdown(title)
48
- output = gr.Code(label="Generated Code", lines=40)
49
- with gr.Row():
50
- generate_btn = gr.Button("Generate")
51
  with gr.Row():
52
  temperature = gr.Slider(minimum=0.1, maximum=1.0, step=0.1, value=0.5, label="Temperature")
53
  max_length = gr.Slider(minimum=100, maximum=1024, step=10, value=100, label="Generate Length")
54
- with gr.Row():
55
- prompt = gr.Textbox(label="Enter your code prompt", placeholder="def print_hello_world():")
56
 
57
  generate_btn.click(
58
  fn=generate_text,
@@ -60,4 +43,7 @@ def gradio_app():
60
  outputs=output
61
  )
62
 
63
- demo.launch()
 
 
 
 
2
  import gradio as gr
3
  import torch
4
  import transformers
5
+ from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
6
  import os
7
 
8
  title = """# Welcome to 🌟Tonic's✨StarCoder
 
14
 
15
  model_path = "bigcode/starcoder2-15b"
16
 
 
17
  hf_token = os.getenv("HF_TOKEN")
18
  if not hf_token:
19
  raise ValueError("Hugging Face token not found. Please set the HF_TOKEN environment variable.")
20
 
21
+ tokenizer = AutoTokenizer.from_pretrained(model_path)
22
+ quantization_config = BitsAndBytesConfig(load_in_8bit=True)
23
+ model = AutoModelForCausalLM.from_pretrained( model_path, quantization_config=quantization_config).to("cuda")
 
 
 
 
 
 
 
 
 
 
 
 
 
 
24
  @spaces.GPU
25
  def generate_text(prompt, temperature, max_length):
26
  inputs = tokenizer.encode(prompt, return_tensors="pt").to("cuda")
 
30
  def gradio_app():
31
  with gr.Blocks() as demo:
32
  gr.Markdown(title)
33
+ prompt = gr.Textbox(label="Enter your code prompt", placeholder="def print_hello_world():")
 
 
34
  with gr.Row():
35
  temperature = gr.Slider(minimum=0.1, maximum=1.0, step=0.1, value=0.5, label="Temperature")
36
  max_length = gr.Slider(minimum=100, maximum=1024, step=10, value=100, label="Generate Length")
37
+ generate_btn = gr.Button("Try✨StarCoder")
38
+ output = gr.Code(label="Generated Code", lines=40)
39
 
40
  generate_btn.click(
41
  fn=generate_text,
 
43
  outputs=output
44
  )
45
 
46
+ demo.launch()
47
+
48
+ if __name__ == "__main__":
49
+ gradio_app()