additional changes

app.py

@@ -158,5 +158,6 @@ iface = gr.Interface(
     examples=examples,
     cache_examples=True,
     title="BLOOM vs BLOOMZ",
-    description='''<p>Compare outputs of the BLOOM and BLOOMZ 176 billion parameter models using the [Petals](https://petals.ml/) network. Please consider joining the Petals network to help speed up inference.</p><p>Big thanks to [RFTCapital](https://www.rftcapital.com) for providing initial compute resources.</p>'''
+    description='''<p>Compare outputs of the BLOOM and BLOOMZ 176 billion parameter models using the [Petals](https://petals.ml/) network. WARNING: Inference may take a long time. Keep the max_tokens low to speed things up.<p>
+    <p>Please consider [joining](https://github.com/bigscience-workshop/petals) the Petals network to help speed up inference.</p><p>Big thanks to [RFTCapital](https://www.rftcapital.com) for providing initial compute resources.</p>'''
 ).launch()

requirements.txt

@@ -1,4 +1,9 @@
 threadpoolctl==3.1.0
 transformers==4.25.1
+bitsandbytes>=0.35.4
+huggingface-hub>=0.7.0
+protobuf>=3.12.2,<4.0.0
 torch==1.13.1
+hivemind>=1.1.2
+gradio
 petals