Notebook ML Domain Framework Models Used Type Enhanced Pod16 Capability Pod4 Capability Link URL
Notebook
Model(s) used
ML Domain
Type
POD4 Capability
Enhanced POD16 capability
Links
Stable Diffusion Image-to-Image Generation on IPU
Stable Diffusion
CV
Inference
-
Stable Diffusion Text-to-Image Generation on IPU
Stable Diffusion
CV
Inference
-
Stable Diffusion Text Guided In-Painting on IPU
Stable Diffusion
CV
Inference
-
Training a ViT Hugging Face model in PyTorch using the IPU using your own dataset
ViT
CV
Fine-Tuning
Yes (run faster on POD16)
Fine-tuning for Image Classification with Hugging Face Optimum on IPU
ViT
CV
Fine-Tuning
Yes (run faster on POD16)
BERT-Large Fine Tuning on IPU
BERT-Large
NLP
Fine-Tuning
Yes (run faster on POD16)
Fast sentiment analysis using pre-trained models on Graphcore IPU
BERT, RoBERTa
NLP
Inference
-
Text Generation with GPT2 using IPUs
GPT2-S/M/XL
NLP
Inference
Support for larger GPT2 models
Inference for Named Entity Recognition with BERT on IPU
BERT
NLP
Inference
-
Fine-Tuning for Named Entity Recognition with BERT on IPU
BERT
NLP
Fine-Tuning
Yes (run faster on POD16)
Fine-tuning a model on a multiple choice task on IPU
RoBERTa
NLP
Fine-Tuning
Yes (run faster on POD16)
Fine-tuning a model on a question-answering task on IPU
RoBERTa
NLP
Fine-Tuning
Yes (run faster on POD16)
Fine-tuning a model on a summarization task on IPU
T5 Small
NLP
Fine-Tuning
Yes (run faster on POD16)
Fine-tuning a model on a text classification task on IPU
RoBERTa
NLP
Fine-Tuning
Yes (run faster on POD16)
Fine-tuning a model on a token classification task on IPU
BERT
NLP
Fine-Tuning
Yes (run faster on POD16)
Fine-tuning a model on a translation task on IPU
BART-Base
NLP
Fine-Tuning
Yes (run faster on POD16)
Training large graphs efficiently with Cluster-GCN on IPU
Cluster-GCN
GNN
Training
-
Training Dynamic Graphs with Temporal Graph Networks (TGN) on IPU
TGN
GNN
Training
-
Prediction of molecular properties using SchNet on IPU
SchNet
GNN
Training
Yes (run faster on POD16)
Predicting of molecular properties using GPS++ on IPU (OGB-LSC)
GPS++
GNN
Inference
-
Training for molecular property prediction using GPS++ on IPU (OGB-LSC)
GPS++
GNN
Training
-
Yes (POD16 enables Training GPS++)
Link prediction training for knowledge graphs using Distributed KGE on IPU (OGB-LSC)
Dist KGE - TransE (256)
GNN
Training
Yes (run faster on POD16)
Fine-tune a wav2vec 2 checkpoint for Automatic Speech Recognition (ASR) on IPU
wav2vec
Speech Processing
Fine-Tuning
Yes (run faster on POD16)
Running Automated Speech Recognition (ASR) using a fine-tuned wav2vec 2.0 checkpoint on IPU
wav2vec
Speech Processing
Inference
Yes (run faster on POD16)