limited the choice of models to just 2

app.py

@@ -7,6 +7,7 @@ In the demo, the user can write a prompt
 import streamlit as st
 from torch.cuda import CudaError
 
+from available_models import AVAILABLE_MODELS
 from hanlde_form_submit import on_form_submit
 from on_server_start import main as on_server_start_main
 
@@ -15,12 +16,10 @@ on_server_start_main()
 st.title("Grouped Sampling Demo")
 
 with st.form("request_form"):
-    selected_model_name: str = st.text_input(
-        label="Model name",
-        value="gpt2",
-        help="The name of the model to use."
-             "Supported models are all the models in:"
-             " https://huggingface.co/models?pipeline_tag=text-generation&library=pytorch",
+    selected_model_name: str = st.selectbox(
+        label="Select a model",
+        options=AVAILABLE_MODELS,
+        help="opt-iml-max-30b generates better texts but is slower",
     )
 
     output_length: int = st.number_input(
@@ -70,5 +69,3 @@ with st.form("request_form"):
 with open("user_instructions_hebrew.md", "r") as fh:
     long_description = fh.read()
 st.markdown(long_description)
-
-await on_server_start_main()

available_models.py

@@ -0,0 +1,4 @@
+AVAILABLE_MODELS = (
+    "facebook/opt-iml-max-1.3b",
+    "facebook/opt-iml-max-30b",
+)

on_server_start.py

@@ -6,6 +6,7 @@ from concurrent.futures import ThreadPoolExecutor
 from transformers import logging as transformers_logging
 from huggingface_hub import logging as huggingface_hub_logging
 
+from available_models import AVAILABLE_MODELS
 from download_repo import download_pytorch_model
 
 
@@ -22,17 +23,13 @@ def download_useful_models():
     Downloads the models that are useful for this project.
     So that the user doesn't have to wait for the models to download when they first use the app.
     """
-    print("Downloading useful models...")
-    useful_models = (
-        "facebook/opt-125m",
-        "facebook/opt-iml-max-30b",
-    )
+    print("Downloading useful models. It might take a while...")
     with ThreadPoolExecutor() as executor:
-        executor.map(download_pytorch_model, useful_models)
+        executor.map(download_pytorch_model, AVAILABLE_MODELS)
 
 
-async def main():
-    disable_progress_bar()
+def main():
+    # disable_progress_bar()
     download_useful_models()