Update app.py

app.py

@@ -4,10 +4,10 @@ import transformers
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
 import gradio as gr
-
+import sentencepiece
 
 title = "Welcome to Tonic's 🐋🐳Orca-2-13B!"
-description = "You can use [🐋🐳microsoft/Orca-2-13b](https://huggingface.co/microsoft/Orca-2-13b) Or clone this space to use it locally or on huggingface! [Join me on Discord to build together](https://discord.gg/VqTxc76K3u)."
+description = "You can use [🐋🐳microsoft/Orca-2-13b](https://huggingface.co/microsoft/Orca-2-13b) via API using Gradio by scrolling down and clicking Use 'Via API' or privately by [cloning this space on huggingface](https://huggingface.co/spaces/Tonic1/TonicsOrca2?duplicate=true) . [Join me on Discord to build together](https://discord.gg/VqTxc76K3u). Big thanks to the HuggingFace Organisation for the Community Grant."
 
 # os.environ['PYTORCH_CUDA_ALLOC_CONF'] = 'max_split_size_mb:24'
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
@@ -17,7 +17,7 @@ offload_folder = './model_weights'
 if not os.path.exists(offload_folder):
     os.makedirs(offload_folder)
 
-model = AutoModelForCausalLM.from_pretrained(model_name, device_map='auto', offload_folder=offload_folder)
+model = AutoModelForCausalLM.from_pretrained(model_name, device_map='auto')
 tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False,)
 model = model.to(torch.bfloat16)