update code

app.py

@@ -12,31 +12,6 @@ st.set_page_config(
      initial_sidebar_state="expanded",
 )
 
-# st.markdown(
-#     """
-# <style>
-
-#     .sidebar .sidebar-content {
-#         background-image: linear-gradient(#3377ff,  #80aaff);
-#     }
-
-#     footer {
-#         color:white;
-#         visibility: hidden;
-#     }
-#     input {
-#         direction: rtl;
-#     }
-#     .stTextInput .instructions {
-#         color: grey;
-#         font-size: 9px;}
-
-# </style>
-# <div style="color:white; font-size:13px; font-family:monospace;position: fixed; z-index: 1; bottom: 0; right:0; background-color: #f63766;margin:3px;padding:8px;border-radius: 5px;"><a href="https://huggingface.co/onlplab/alephbert-base"  target="_blank" style="text-decoration: none;color: white;">Use aleph-bert in your project </a></div>
-# """,
-#     unsafe_allow_html=True,
-# )
-
 models = {
     "AlephBERT-base": {
         "name_or_path":"onlplab/alephbert-base",
@@ -116,16 +91,8 @@ if mode == 'Models':
 
     st.markdown(''.join([f'<span style="color:white; font-size:13px; font-family:monospace; background-color: #f63766;margin:3px;padding:8px;border-radius: 5px;">{tag}</span>' for tag in model_tags]),unsafe_allow_html=True)
     st.markdown('___')
-    ####
-    #prepare the model
-    ####
-    
+
     unmasker, tokenize = load_model(model)
-    
-    
-    ####
-    # get inputs
-    ####
             
     input_text = st.text_input('Insert text you want to mask', '')
     if input_text:
@@ -136,9 +103,9 @@ if mode == 'Models':
         
         if masking_level == 'Tokens':
             tokens = str(input_text).split()
-            masked_token = st.selectbox('Select token to mask:', [''] + tokens)
-            if masked_token != '':
-                input_masked = ' '.join(token if token != masked_token else '[MASK]' for token in tokens)
+            mask_idx = st.selectbox('Select token to mask:', [None] + list(range(len(tokens))), format_func=lambda i: tokens[i] if i else '')
+            if mask_idx is not None:
+                input_masked = ' '.join(token if i != mask_idx else '[MASK]' for i, token in enumerate(tokens))
                 display_input = input_masked
         if masking_level == 'SubWords':
             tokens = subwords
@@ -157,25 +124,13 @@ if mode == 'Models':
                         unsafe_allow_html=True,
             )
             st.markdown('#### Outputs:')
-            res = unmasker(input_masked, tokenized=masking_level == 'SubWords', top_k=n_res)
+            with st.spinner('Running model...'):
+                res = unmasker(input_masked, tokenized=masking_level == 'SubWords', top_k=n_res)
             if res:
                 res = [{'Prediction':r['token_str'], 'Completed Sentence':r['sequence'].replace('[SEP]', '').replace('[CLS]', ''), 'Score':r['score']} for r in res]
                 res_table = pd.DataFrame(res)
                 st.table(res_table)
             
-            
-        
-#         cols = st.beta_columns(len(tokens))
-#         genre = st.radio(
-#      'Select token to mask:', tokens)
-#         for col, token in zip(cols, reversed(tokens)):
-#             col.text(token)
-        
-#         st.text(tokens)
-#         res = unmasker(input_text)
-#         res_table = pd.DataFrame(res)
-#         st.table(res_table)
-#         st.text(res)
         
 
 

requirements.txt

@@ -1,5 +1,5 @@
 torch
 sentencepiece
-transformers==4.4.2
+transformers==4.6.1 
 tokenizers
 pandas