Keynote: Panel(...)
summary
Panel maps concrete work: Nx brings multi‑dimensional tensors with CPU/GPU backends, Axon is a deep‑learning library on top of Nx, and Livebook is a collaborative Elixir notebook; data preparation consumes 80–90% of the work, and production needs model serialization (ONNX is the likely target) plus techniques like quantization for billion‑parameter models. Practical gaps called out: fast data tooling (pandas/DataFrame equivalents), tree‑based algorithms (LightGBM/XGBoost) support, reliable GPU memory management (ref‑counting), and model versioning/deployment (MLflow/serving).
AI-written summary · about
chapters (8)
- 0:00Introductions and overview
- 2:08Machine learning lifecycle
- 7:22Elixir's machine learning tools
- 12:47Industry challenges and needs
- 15:48Model deployment standards
- 22:16Deployment strategies
- 32:19Ecosystem and future vision
- 38:32Adoption and closing advice
related talks
23mDeep dive in Nx Backends
25mApplied Machine Learning in Elixir
24mKeynote: Next-generation frameworks for Large-scale AI
32mKeynote: AI & the stuff built for AI - are they actually useful for data science?
12mKeynote: Build Accelerated LLM Applications with Open Source Tools from NVIDIA
36mKeynote: Scientific Discovery: From the Lab Bench to the GPU
14mLightning Talk: Every Millisecond Counts: The Fine-tuning Journey of an Ultra-Eff...
01h 05mKeynote: Builders Day