why's this taking so long

README.md

@@ -13,24 +13,4 @@ pinned: false
 # image2textapp
 demo of 🤗 spaces deployment of a streamlit python app
 
-
-Installation instructions
-
-```docker compose run dev-environment```
-
-Procedure used:
-
-Reasoned that it would make the most amount of sense to be able to modify the
-source code while the container is still running to allow for iterative
-debugging in the environment in which it is being deployed. To avoid writing
-back to the system, a readonly option was provided to the filesystem.
-
-Docker compose was used to provide a separation of concerns, and to move testing
-logic outside of the container that is to be deployed. This decouples the logic
-of the tests from the application logic. I have familiarity with docker compose
-and am happy to work with it again.
-
-A bare-metal python Dockerfile base image was used to provide a stable python
-deployment version. This version will be targeted in the poetry files, and
-packages necessary will be installed into the system python with the appropriate
-poetry arguments.
+deployed at https://huggingface.co/spaces/NativeVex/large-language-models

language_models_project/app.py

@@ -10,16 +10,26 @@ st.title("Easy OCR - Extract Text from Images")
 #subtitle
 st.markdown("## Optical Character Recognition - Using `easyocr`, `streamlit` -  hosted on 🤗 Spaces")
 
+model_name = st.selectbox(
+    'Select a pre-trained model',
+    [
+        'finiteautomata/bertweet-base-sentiment-analysis',
+        'ahmedrachid/FinancialBERT-Sentiment-Analysis',
+        'finiteautomata/beto-sentiment-analysis'
+    ],
+)
+
 input_sentences = st.text_area("Sentences", value="", height=200)
 
 data = input_sentences.split('\n')
 
 #if st.button("Classify"):
 for i in data:
-    j = classify(i)[0]
-    sentiment = j['label'] == 'POS'
+    st.write(i)
+    j = classify(model_name.strip(), i)[0]
+    sentiment = j['label']
     confidence = j['score']
-    st.write(f"{i} :: Classification - {'positive' if sentiment else 'negative'} with confidence {confidence}")
+    st.write(f"{i} :: Classification - {sentiment} with confidence {confidence}")
  
 
 st.markdown("Link to the app - [image-to-text-app on 🤗 Spaces](https://huggingface.co/spaces/Amrrs/image-to-text-app)")

language_models_project/main.py

@@ -1,8 +1,12 @@
 from transformers import pipeline
 from transformers import AutoTokenizer, AutoModelForSequenceClassification
 
-def classify(*args, **kwargs):
-    tokenizer = AutoTokenizer.from_pretrained("finiteautomata/bertweet-base-sentiment-analysis")
-    model = AutoModelForSequenceClassification.from_pretrained("finiteautomata/bertweet-base-sentiment-analysis")
-    sentiment_pipeline = pipeline("sentiment-analysis", model=model, tokenizer=tokenizer)
+def classify(model_string: str, *args, **kwargs):
+    tokenizer = AutoTokenizer.from_pretrained(model_string)
+    model = AutoModelForSequenceClassification.from_pretrained(model_string)
+    sentiment_pipeline = pipeline(
+        "sentiment-analysis",
+        model=model,
+        tokenizer=tokenizer
+    )
     return sentiment_pipeline(*args, **kwargs)