File size: 2,084 Bytes
9439556
 
 
 
 
51833bf
9439556
75128ac
391a4ae
 
 
84c4fac
391a4ae
 
 
 
 
 
9439556
 
 
 
 
 
 
 
51833bf
84c4fac
9439556
391a4ae
9439556
391a4ae
 
 
 
84c4fac
9439556
391a4ae
 
 
 
 
 
 
 
 
 
a79d902
9439556
 
 
 
 
 
 
 
ec05561
9439556
 
 
 
9de310e
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
import torch
import torch.nn.functional as F
from transformers import AutoTokenizer, AutoModel, AutoImageProcessor
import gradio as gr
import spaces
import torch

# neuralmind/bert-base-portuguese-cased
#ModelName = "neuralmind/bert-base-portuguese-cased"
#model = AutoModel.from_pretrained(ModelName)
#tokenizer = AutoTokenizer.from_pretrained(ModelName, do_lower_case=False)


#processor = AutoImageProcessor.from_pretrained("nomic-ai/nomic-embed-vision-v1.5")
#vision_model = AutoModel.from_pretrained("nomic-ai/nomic-embed-vision-v1.5", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained('nomic-ai/nomic-embed-text-v1.5')
text_model = AutoModel.from_pretrained('nomic-ai/nomic-embed-text-v1.5', trust_remote_code=True)
text_model.eval()

def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0]
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
        
@spaces.GPU
def TxtEmbed(text):
   

    
    #input_ids = tokenizer.encode(text, return_tensors='pt')
    
    #with torch.no_grad():
    #    outs = model(input_ids)
    #    encoded = outs[0][0, 1:-1]  # Ignore [CLS] and [SEP] special tokens
    #return (encoded.tolist())[0];

    
   sentences = [text]
   encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
   
   with torch.no_grad():
       model_output = text_model(**encoded_input)
   
   text_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
   text_embeddings = F.layer_norm(text_embeddings, normalized_shape=(text_embeddings.shape[1],))
   text_embeddings = F.normalize(text_embeddings, p=2, dim=1)
    
   return (text_embeddings.tolist())[0]




with gr.Blocks() as demo:      
       txt = gr.Text();
       out = gr.Text();
       
       btn = gr.Button("Generate embeddings")
       btn.click(TxtEmbed, [txt], [out])
       
       
if __name__ == "__main__":
    demo.launch(show_api=True)