app definition

.streamlit/config.toml

@@ -0,0 +1,2 @@
+[theme]
+base = "dark"

app.py

@@ -0,0 +1,69 @@
+from transformers import AutoTokenizer, AutoModelForCausalLM,pipeline
+import torch
+import streamlit as st
+
+
+
+MODELS={
+        'uribe':{
+                'tokenizer':AutoTokenizer.from_pretrained("jhonparra18/uribe-twitter-assistant-30ep"),
+                'model':AutoModelForCausalLM.from_pretrained("jhonparra18/uribe-twitter-assistant-30ep")},
+        'petro':{
+                'tokenizer':AutoTokenizer.from_pretrained("jhonparra18/petro-twitter-assistant-30ep"),
+                'model':AutoModelForCausalLM.from_pretrained("jhonparra18/petro-twitter-assistant-30ep")}}
+
+# MODELS={'petro':pipeline('text-generation', model='jhonparra18/petro-twitter-assistant-30ep-large'),
+#         'uribe':pipeline('text-generation', model='jhonparra18/uribe-twitter-assistant-30ep')}
+
+
+def text_completion(tokenizer,model,input_text:str,max_len:int=100):
+    torch.manual_seed(444)    ##for reproducibility
+    tokenizer.padding_side="left" ##start padding from left to right
+    tokenizer.pad_token = tokenizer.eos_token 
+    input_ids = tokenizer([input_text], return_tensors="pt",truncation=True,max_length=128)
+    outputs = model.generate(**input_ids, do_sample=True, max_length=max_len,top_k=100,top_p=0.95)
+    out_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0]
+    return out_text
+
+
+
+if "input_user_txt" not in st.session_state:
+    st.session_state["input_user_txt"] = None
+
+if "user_max_length" not in st.session_state:
+    st.session_state["user_max_length"] = 100
+
+
+st.markdown("<h3 style='text-align: center; color: white;'>Tweet de Político Colombiano: Autocompletado/generación de texto a partir de GPT2</h3>", unsafe_allow_html=True)
+st.text("")
+st.markdown("<h3 style='text-align: center; color: white;'>Causal Language Modeling</h3>", unsafe_allow_html=True)
+st.text("")
+
+
+col1,col2 = st.columns(2)
+
+
+with col1:
+    with st.form("input_values"):
+        politician = st.selectbox(
+            "Selecciona el político",
+            ("Uribe", "Petro")
+        )
+        st.text("")
+        max_length_text=st.slider('Num Max Tokens', 100, 200, 100,10,key="user_max_length")
+        st.text("")
+        input_user_text = st.text_area('Input Text', 'Mi gobierno es',key="input_user_txt")
+        st.text("")
+        go_button=st.form_submit_button('Generate',use_container_width=True)
+
+
+with col2:
+    
+    if go_button: ##avoid re running script
+        with st.spinner('Generating Text...'):
+            output_text=text_completion(MODELS[politician.lower()]['tokenizer'],MODELS[politician.lower()]['model'],input_user_text,max_length_text)
+            st.text_area("output text",output_text,height=380,key="output_text")
+    else:
+        st.text_area("output text","",height=380,key="output_text")
+    
+

requirements.txt

@@ -0,0 +1,175 @@
+#
+# This file is autogenerated by pip-compile with Python 3.9
+# by the following command:
+#
+#    pip-compile --output-file=requirements.txt requirements.in
+#
+aiohttp==3.8.4
+    # via
+    #   datasets
+    #   fsspec
+aiosignal==1.3.1
+    # via aiohttp
+appdirs==1.4.4
+    # via wandb
+async-timeout==4.0.2
+    # via aiohttp
+attrs==22.2.0
+    # via aiohttp
+beautifulsoup4==4.11.2
+    # via snscrape
+certifi==2022.12.7
+    # via
+    #   requests
+    #   sentry-sdk
+charset-normalizer==3.0.1
+    # via
+    #   aiohttp
+    #   requests
+click==8.1.3
+    # via wandb
+datasets==2.9.0
+    # via -r requirements.in
+dill==0.3.6
+    # via
+    #   datasets
+    #   multiprocess
+docker-pycreds==0.4.0
+    # via wandb
+filelock==3.9.0
+    # via
+    #   huggingface-hub
+    #   snscrape
+    #   transformers
+frozenlist==1.3.3
+    # via
+    #   aiohttp
+    #   aiosignal
+fsspec[http]==2023.1.0
+    # via datasets
+gitdb==4.0.10
+    # via gitpython
+gitpython==3.1.31
+    # via wandb
+huggingface-hub==0.12.1
+    # via
+    #   datasets
+    #   transformers
+idna==3.4
+    # via
+    #   requests
+    #   yarl
+lxml==4.9.2
+    # via snscrape
+multidict==6.0.4
+    # via
+    #   aiohttp
+    #   yarl
+multiprocess==0.70.14
+    # via datasets
+numpy==1.24.2
+    # via
+    #   datasets
+    #   pandas
+    #   pyarrow
+    #   transformers
+nvidia-cublas-cu11==11.10.3.66
+    # via
+    #   nvidia-cudnn-cu11
+    #   torch
+nvidia-cuda-nvrtc-cu11==11.7.99
+    # via torch
+nvidia-cuda-runtime-cu11==11.7.99
+    # via torch
+nvidia-cudnn-cu11==8.5.0.96
+    # via torch
+packaging==23.0
+    # via
+    #   datasets
+    #   huggingface-hub
+    #   transformers
+pandas==1.5.3
+    # via
+    #   -r requirements.in
+    #   datasets
+pathtools==0.1.2
+    # via wandb
+protobuf==4.22.0
+    # via wandb
+psutil==5.9.4
+    # via wandb
+pyarrow==11.0.0
+    # via datasets
+pysocks==1.7.1
+    # via requests
+python-dateutil==2.8.2
+    # via pandas
+pytz==2022.7.1
+    # via pandas
+pyyaml==6.0
+    # via
+    #   datasets
+    #   huggingface-hub
+    #   transformers
+    #   wandb
+regex==2022.10.31
+    # via transformers
+requests[socks]==2.28.2
+    # via
+    #   datasets
+    #   fsspec
+    #   huggingface-hub
+    #   responses
+    #   snscrape
+    #   transformers
+    #   wandb
+responses==0.18.0
+    # via datasets
+sentry-sdk==1.15.0
+    # via wandb
+setproctitle==1.3.2
+    # via wandb
+six==1.16.0
+    # via
+    #   docker-pycreds
+    #   python-dateutil
+smmap==5.0.0
+    # via gitdb
+snscrape==0.5.0.20230113
+    # via -r requirements.in
+soupsieve==2.4
+    # via beautifulsoup4
+tokenizers==0.13.2
+    # via transformers
+torch==1.13.1
+    # via -r requirements.in
+tqdm==4.64.1
+    # via
+    #   datasets
+    #   huggingface-hub
+    #   transformers
+transformers==4.26.1
+    # via -r requirements.in
+typing-extensions==4.5.0
+    # via
+    #   huggingface-hub
+    #   torch
+    #   wandb
+urllib3==1.26.14
+    # via
+    #   requests
+    #   responses
+    #   sentry-sdk
+wandb==0.13.10
+    # via -r requirements.in
+wheel==0.38.4
+    # via
+    #   nvidia-cublas-cu11
+    #   nvidia-cuda-runtime-cu11
+xxhash==3.2.0
+    # via datasets
+yarl==1.8.2
+    # via aiohttp
+
+# The following packages are considered to be unsafe in a requirements file:
+# setuptools