Increasing the Impact of AI Through Better Software (Pytorch Updates)
summary
A practical update on PyTorch: core pieces are an automatic-differentiation engine, a GPU-backed tensor library, optimizers and data loaders, and new tooling for production. Distributed data-parallel training scales near-linearly to 64-node scaling (roughly 60x speedup reported). Models can be traced/scripted and exported to a C++ runtime export for deployment, and tensors interoperate with NumPy with no copies enabling mixed pipelines (e.g., send NN outputs into scikit‑learn). Quantization and domain libraries like medical imaging round out the stack.
AI-written summary · about
chapters (4)
related talks
22mIncreasing the Impact of AI Through Better Software
47mGet to Know PyTorch - A Cutting Edge Open Source AI Framework for Deep Learning
19mGet to know PyTorch - a cutting edge open source AI framework for Machine Learning
54mKeynote: PyTorch 2.1 Technical Deep Dive
43mPyTorch Technical Deep Dive
06mKeynote: PyTorch Lightning: Powering the GenAI Revolution from Research to the...
12mKeynote: How PyTorch Became the Foundation of the AI Revolution
13mKeynote