Hugging Face's logo

Mark-CHAE
/
ViTCM-LLM

Visual Question Answering
PEFT
English
Korean
Chinese
vision
multimodal
qwen
lora
tcm
traditional-chinese-medicine
tongue-diagnosis
Model card Files
xet
Community
ViTCM-LLM
File size: 5,321 Bytes 
8374b0f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
 
import streamlit as st
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM, AutoProcessor
from peft import PeftModel
from PIL import Image
import io

# Page configuration
st.set_page_config(
    page_title="ViTCM_LLM Tongue Diagnosis",
    page_icon="๐Ÿ–ผ๏ธ",
    layout="wide"
)

# Title
st.title("๐Ÿ–ผ๏ธ ViTCM_LLM Tongue Diagnosis")
st.markdown("**ViTCM_LLM - Traditional Chinese Medicine Tongue Diagnosis Model**")

# Model loading
@st.cache_resource
def load_model():
    """Load the ViTCM_LLM model for TCM tongue diagnosis."""
    try:
        # Tokenizer and processor
        tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-VL-32B-Instruct")
        processor = AutoProcessor.from_pretrained("Qwen/Qwen2.5-VL-32B-Instruct")
        
        # Base model
        base_model = AutoModelForCausalLM.from_pretrained(
            "Qwen/Qwen2.5-VL-32B-Instruct",
            torch_dtype=torch.float16,
            device_map="auto"
        )
        
        # LoRA adapter
        model = PeftModel.from_pretrained(base_model, "Mark-CHAE/shezhen")
        
        return model, tokenizer, processor
    except Exception as e:
        st.error(f"Model loading failed: {e}")
        return None, None, None

# Sidebar
with st.sidebar:
    st.header("โš™๏ธ Settings")
    
    # Inference parameters
    max_length = st.slider("Max tokens", 100, 1024, 512)
    temperature = st.slider("Temperature", 0.1, 2.0, 0.7, 0.1)
    top_p = st.slider("Top-p", 0.1, 1.0, 0.9, 0.05)
    
    # Model load button
    if st.button("๐Ÿš€ Load Model", type="primary"):
        with st.spinner("Loading ViTCM_LLM model..."):
            model, tokenizer, processor = load_model()
            if model is not None:
                st.session_state.model = model
                st.session_state.tokenizer = tokenizer
                st.session_state.processor = processor
                st.session_state.model_loaded = True
                st.success("โœ… ViTCM_LLM model loaded successfully!")

# Main content
if not st.session_state.get('model_loaded', False):
    st.info("๐Ÿ‘ˆ Click 'Load Model' button in the sidebar to start tongue diagnosis.")
    st.stop()

# Image upload
st.header("๐Ÿ“ธ Tongue Image Upload")
uploaded_file = st.file_uploader(
    "Upload a tongue image for TCM diagnosis",
    type=['png', 'jpg', 'jpeg']
)

if uploaded_file is not None:
    # Display image
    image = Image.open(uploaded_file)
    st.image(image, caption="Uploaded tongue image", use_column_width=True)
    
    # Question input
    st.header("โ“ Tongue Diagnosis Question")
    question = st.text_area(
        "Ask a question about the tongue image for TCM diagnosis",
        placeholder="e.g., ๆ นๆฎๅ›พ็‰‡ๅˆคๆ–ญ่ˆŒ่ฏŠๅ†…ๅฎน",
        height=100
    )
    
    # Analyze button
    if st.button("๐Ÿ” Analyze Tongue", type="primary") and question.strip():
        with st.spinner("Analyzing tongue for TCM diagnosis..."):
            try:
                # Construct prompt
                prompt = f"<|im_start|>user\n<image>\n{question}<|im_end|>\n<|im_start|>assistant\n"
                
                # Process inputs
                inputs = st.session_state.processor(
                    text=prompt,
                    images=image,
                    return_tensors="pt"
                )
                
                # Inference
                with torch.no_grad():
                    outputs = st.session_state.model.generate(
                        **inputs,
                        max_length=max_length,
                        temperature=temperature,
                        top_p=top_p,
                        do_sample=True,
                        pad_token_id=st.session_state.tokenizer.eos_token_id
                    )
                
                # Process results
                response = st.session_state.tokenizer.decode(outputs[0], skip_special_tokens=True)
                answer = response.split("<|im_start|>assistant")[-1].strip()
                
                # Display results
                st.header("๐Ÿ’ก TCM Tongue Diagnosis")
                st.markdown(f"**Question:** {question}")
                st.markdown(f"**Diagnosis:** {answer}")
                
            except Exception as e:
                st.error(f"Error occurred during tongue analysis: {e}")

# Usage examples
with st.expander("๐Ÿ“š Tongue Diagnosis Examples"):
    st.markdown("""

    ### Tongue Diagnosis Questions:

    - ๆ นๆฎๅ›พ็‰‡ๅˆคๆ–ญ่ˆŒ่ฏŠๅ†…ๅฎน

    - ๅˆ†ๆž่ˆŒๅคด็š„้ขœ่‰ฒๅ’Œๅฝข็Šถ

    - ๅˆคๆ–ญ่ˆŒ่‹”็š„ๅŽš่–„ๅ’Œ้ขœ่‰ฒ

    - ๅˆ†ๆž่ˆŒๅคด็š„่ฃ‚็บนๅ’Œๆ–‘็‚น

    - ่ฏ„ไผฐ่ˆŒๅคด็š„ๆ•ดไฝ“ๅฅๅบท็Šถๅ†ต

    """)

# Model information
with st.expander("โ„น๏ธ Model Information"):
    st.markdown("""

    ### ViTCM_LLM - Traditional Chinese Medicine Tongue Diagnosis Model

    

    - **Base Model**: Qwen/Qwen2.5-VL-32B-Instruct

    - **Adapter**: Mark-CHAE/shezhen (ViTCM_LLM)

    - **Language**: Chinese

    - **License**: Apache-2.0

    - **Specialization**: Traditional Chinese Medicine Tongue Diagnosis

    """)

st.markdown("---")
st.markdown("**ViTCM_LLM Tongue Diagnosis** | Powered by Qwen2.5-VL-32B-Instruct")