File size: 2,215 Bytes
7103ccc
 
 
748826b
 
1c026a2
7103ccc
714a27c
748826b
416fea8
748826b
2371338
 
 
 
748826b
 
d590a55
 
 
2371338
d590a55
 
748826b
 
d590a55
748826b
416fea8
 
748826b
7103ccc
1c026a2
 
 
 
5c51be8
1c026a2
 
 
 
 
 
 
 
 
 
 
 
 
 
f6551d5
5c51be8
f6551d5
5e78e4f
9006e63
 
 
1c026a2
9006e63
 
 
1c026a2
9006e63
 
 
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
import gradio as gr
import spaces
import torch
from transformers import AutoTokenizer, AutoModel
import plotly.graph_objects as go
import numpy as np

model_name = "mistralai/Mistral-7B-v0.1"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = None

# Set pad token to eos token if not defined
if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token

@spaces.GPU
def get_embedding(text):
    global model
    if model is None:
        model = AutoModel.from_pretrained(model_name).cuda()
        model.resize_token_embeddings(len(tokenizer))
    
    inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True, max_length=512).to('cuda')
    with torch.no_grad():
        outputs = model(**inputs)
    return outputs.last_hidden_state.mean(dim=1).squeeze().cpu().numpy()

def reduce_to_3d(embedding):
    return embedding[:3]

@spaces.GPU
def compare_embeddings(text_input):
    texts = text_input.split('\n')
    embeddings = [get_embedding(text) for text in texts]
    embeddings_3d = [reduce_to_3d(emb) for emb in embeddings]
    
    fig = go.Figure()

    # Add origin point (black)
    fig.add_trace(go.Scatter3d(x=[0], y=[0], z=[0], mode='markers', name='Origin',
                               marker=dict(size=5, color='black')))

    # Add lines and points for each text embedding
    colors = ['red', 'blue', 'green', 'purple', 'orange', 'cyan', 'magenta', 'yellow']
    for i, emb in enumerate(embeddings_3d):
        color = colors[i % len(colors)]
        fig.add_trace(go.Scatter3d(x=[0, emb[0]], y=[0, emb[1]], z=[0, emb[2]], 
                                   mode='lines+markers', name=f'Text {i+1}',
                                   line=dict(color=color), marker=dict(color=color)))

    fig.update_layout(scene=dict(xaxis_title='X', yaxis_title='Y', zaxis_title='Z'))
    
    return fig

iface = gr.Interface(
    fn=compare_embeddings,
    inputs=[
        gr.Textbox(label="Input Texts", lines=5, placeholder="Enter multiple texts, each on a new line")
    ],
    outputs=gr.Plot(),
    title="3D Embedding Comparison",
    description="Compare the embeddings of multiple strings visualized in 3D space using Mistral 7B."
)

iface.launch()