Let's get it

app.py

@@ -38,7 +38,7 @@ SIDEBAR_INFO = f"""
 
 NB-GPT-J-6B NorPaca is a hybrid of a GPT-3 and Llama model, trained on the Norwegian Colossal Corpus and other Internet sources. It is a 6.7 billion parameter model, and is the largest model in the GPT-J family.
 
-This model has been trained with [Mesh Transformer JAX](https://github.com/kingoflolz/mesh-transformer-jax) using TPUs provided by Google through the Tensor Research Cloud program, starting off the [GPT-J-6B model weigths from EleutherAI](https://huggingface.co/EleutherAI/gpt-j-6B), and trained on the [Norwegian Colossal Corpus](https://huggingface.co/datasets/NbAiLab/NCC) and other Internet sources. *This demo runs on {DEVICE.split(':')[0].upper()}*.
+This model has been trained with [Mesh Transformer JAX](https://github.com/kingoflolz/mesh-transformer-jax) using TPUs provided by Google through the Tensor Research Cloud program, starting off the [GPT-J-6B model weigths from EleutherAI](https://huggingface.co/EleutherAI/gpt-j-6B), and trained on the [Norwegian Colossal Corpus](https://huggingface.co/datasets/NbAiLab/NCC) and other Internet sources. *This demo runs on {DEVICE}*.
 
 For more information, visit the [model repository](https://huggingface.co/CBSMasterThesis).