Nadav Eden commited on
Commit
e425a6f
·
1 Parent(s): 7f774ed

first version, only llms are functional

Browse files
Files changed (2) hide show
  1. app.py +81 -52
  2. requirements.txt +4 -1
app.py CHANGED
@@ -1,64 +1,93 @@
 
 
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(
11
- message,
12
- history: list[tuple[str, str]],
13
- system_message,
14
- max_tokens,
15
- temperature,
16
- top_p,
17
- ):
18
- messages = [{"role": "system", "content": system_message}]
19
 
20
- for val in history:
21
- if val[0]:
22
- messages.append({"role": "user", "content": val[0]})
23
- if val[1]:
24
- messages.append({"role": "assistant", "content": val[1]})
 
 
 
 
 
25
 
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(
53
- minimum=0.1,
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
+ #!/usr/bin/env python3
2
+
3
  import gradio as gr
4
+ from transformers import AutoTokenizer, AutoModelForCausalLM
5
 
6
+ llms = {
7
+ "Qwen2-1.5B": {"model": "Qwen/Qwen2-1.5B-Instruct", "prefix": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
8
+ "Qwen2-3B": {"model": "Qwen/Qwen2-3B-Instruct", "prefix": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
9
+ "Qwen2-7B": {"model": "Qwen/Qwen2-7B-Instruct", "prefix": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
10
+ "Qwen2.5-1.5B": {"model": "Qwen/Qwen2.5-1.5B-Instruct", "prefix": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
11
+ "Qwen2.5-3B": {"model": "Qwen/Qwen2.5-3B-Instruct", "prefix": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
12
+ "DeepSeek-Coder": {"model": "DeepSeek/DeepSeek-Coder", "prefix": "You are a helpful assistant."},
13
+ }
14
 
15
+ vlms = dict()
16
 
17
+ def run_example(text_input, model_id="Qwen2-1.5B"):
18
+ global messages
19
+ tokenizer = AutoTokenizer.from_pretrained(llms[model_id]["model"], trust_remote_code=True)
20
+ model = AutoModelForCausalLM.from_pretrained(llms[model_id]["model"], trust_remote_code=True)
 
 
 
 
 
21
 
22
+ system_prompt = llms[model_id]["prefix"]
23
+
24
+ if messages is None:
25
+ messages = [
26
+ {"role": "system", "content": system_prompt},
27
+ {"role": "user", "content": text_input},
28
+ ]
29
+ else:
30
+ messages.append({"role": "user", "content": text_input})
31
+
32
 
33
+ text = tokenizer.apply_chat_template(
34
+ messages,
35
+ tokenize=False,
36
+ add_generation_prompt=True,
37
+ )
38
 
39
+ model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
40
+
41
+ generated_ids = model.generate(
42
+ **model_inputs,
43
+ max_new_tokens=512,
44
+ )
45
+
46
+ generated_ids = [
47
+ output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
48
+ ]
49
+
50
+ response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
51
+
52
+ return response
53
+
54
+
55
+ messages = list()
56
+ def reset_conversation():
57
+ global messages
58
+ messages = list()
59
+
60
+ with gr.Blocks() as demo:
61
+ gr.Markdown(
62
+ """
63
+ # LLM & VLM Demo
64
+ Use the different LLMs or VLMs to experience the different models.
65
+ """)
66
+ with gr.Tab(label="LLM"):
67
+ with gr.Row():
68
+ with gr.Column():
69
+ model_selector = gr.Dropdown(choices=list(llms.keys()), label="Model", value="Qwen2-1.5B")
70
+ text_input = gr.Textbox(label="User Prompt")
71
+ submit_btn = gr.Button(value="Submit")
72
+ reset_btn = gr.Button(value="Reset conversation")
73
+ with gr.Column():
74
+ model_output_text = gr.Textbox(label="Model Output Text")
75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
76
 
77
+ submit_btn.click(run_example,
78
+ [text_input, model_selector],
79
+ [model_output_text])
80
+
81
+ reset_btn.click(reset_conversation)
82
+
83
+ with gr.Tab(label="VLM (WIP)"):
84
+ with gr.Row():
85
+ with gr.Column():
86
+ input_img = gr.Image(label="Input Image", type="pil")
87
+ model_selector = gr.Dropdown(choices=list(vlms.keys()), label="Model", value="Qwen2-1.5B")
88
+ text_input = gr.Textbox(label="User Prompt")
89
+ submit_btn = gr.Button(value="Submit")
90
+ reset_btn = gr.Button(value="Reset conversation")
91
 
92
  if __name__ == "__main__":
93
  demo.launch()
requirements.txt CHANGED
@@ -1 +1,4 @@
1
- huggingface_hub==0.25.2
 
 
 
 
1
+ huggingface_hub==0.25.2
2
+ torch
3
+ transformers
4
+ gradio