Update pages/gpt_v1.py

pages/gpt_v1.py

@@ -4,7 +4,8 @@ import torch
 import textwrap
 import plotly.express as px
 
-st.header(':green[Генерация текста GPT-моделью]')
+
+st.header(':green[Text generation by GPT2 model]')
 
 tokenizer = GPT2Tokenizer.from_pretrained('sberbank-ai/rugpt3small_based_on_gpt2')
 model = GPT2LMHeadModel.from_pretrained(
@@ -12,26 +13,33 @@ model = GPT2LMHeadModel.from_pretrained(
     output_attentions = False,
     output_hidden_states = False,
 )
-# Вешаем сохраненные веса на нашу модель
-model.load_state_dict(torch.load('model.pt', map_location=torch.device('cpu')))
+
+model.load_state_dict(torch.load('models/model.pt', map_location=torch.device('cpu')))
 
 
-length = st.sidebar.slider('**Длина генерируемой последовательности:**', 8, 256, 15)
-num_samples = st.sidebar.slider('**Число генераций:**', 1, 10, 1)
-temperature = st.sidebar.slider('**Температура:**', 1.0, 10.0, 2.0)
-top_k = st.sidebar.slider('**Количество наиболее вероятных слов генерации:**', 10, 200, 50)
-top_p = st.sidebar.slider('**Минимальная суммарная вероятность топовых слов:**', 0.4, 1.0, 0.9)
+length = st.sidebar.slider('**Generated sequence length:**', 8, 256, 15)
+if length > 100:
+    st.warning("This is very hard for me, please have pity on me. Could you lower the value?", icon="🤖")
+num_samples = st.sidebar.slider('**Number of generations:**', 1, 10, 1)
+if num_samples > 4:
+    st.warning("OH MY ..., I have to work late again!!! Could you lower the value?", icon="🤖")
+temperature = st.sidebar.slider('**Temperature:**', 0.1, 10.0, 3.0)
+if temperature > 6.0:
+    st.info('What? You want to get some kind of bullshit as a result? Turn down the temperature', icon="🤖")
+top_k = st.sidebar.slider('**Number of most likely generation words:**', 10, 200, 50)
+top_p = st.sidebar.slider('**Minimum total probability of top words:**', 0.4, 1.0, 0.9)
 
 
-prompt = st.text_input('**Введите текст 👇:**')
-if st.button('**Сгенерировать текст**'):
-    
+prompt = st.text_input('**Enter text 👇:**')
+if st.button('**Generate text**'):
+    image_container = st.empty()
+    image_container.image("pict/wait.jpeg", caption="that's so long!!!", use_column_width=True)
     with torch.inference_mode():
         prompt = tokenizer.encode(prompt, return_tensors='pt')
         out = model.generate(
             input_ids=prompt,
             max_length=length,
-            num_beams=5,
+            num_beams=8,
             do_sample=True,
             temperature=temperature,
             top_k=top_k,
@@ -39,9 +47,17 @@ if st.button('**Сгенерировать текст**'):
             no_repeat_ngram_size=3,
             num_return_sequences=num_samples,
             ).cpu().numpy()
+        image_container.empty()
+        st.write('**_Результат_** 👇')
         for i, out_ in enumerate(out):
-            st.write(f'Текст {i+1}:')
-            st.write(textwrap.fill(tokenizer.decode(out_), 100))
+            # audio_file = open('pict/pole-chudes-priz.mp3', 'rb')
+            # audio_bytes = audio_file.read()
+            # st.audio(audio_bytes, format='audio/mp3')
+            
+            with st.expander(f'Текст {i+1}:'):
+                st.write(textwrap.fill(tokenizer.decode(out_), 100))
+                st.image("pict/wow.png")
+