Machine intelligence is a unique computational workload with distinctly different characteristics to HPC algorithms or graphics programs. With the slowing down of Moore’s Law and model sizes on the rise, there is a need for specialised machine learning hardware designed to run AI workloads efficiently.
Phil Brown, Graphcore's Director of Applications, recently spoke to Founder of Weights & Biases, Lukas Biewald, about the role of AI processors such as the IPU in driving forward progress in machine intelligence, from enabling sparsity to accelerating BERT.
Listen to the Weights & Biases Podcast
Scaling Experiments with Weights & Biases
Pursuing new approaches to machine learning can be a challenge, particularly once AI workloads move from pilot to production. At scale, even a slight drop in performance can be costly. Recognising this, Weights & Biases have created a suite of tools to help developers scale up their projects more easily.
Graphcore engineers have been using Weights & Biases tools for AI and machine learning to support their work scaling IPU experiments. In a recent Weights & Biases case study, Graphcore’s Phil Brown shared how the complexities of tracking experiments across IPU-POD systems and multiple deployment locations led his team to turn to the Weights & Biases platform to track their large-scale experiments, including their BERT-Large Training implementation on the IPU.
Read the Weights & Biases Case Study
