PyTorchCon Demo Theater afternoon: IBM Spyre, Modal, Arm and 5 more sponsor demos
PyTorch · x · 2026-10-21
PyTorchCon's Demo Theater features 14 sponsor presentations on training, inference, compilers, and deployment. The remaining afternoon sessions include:
- 2:00 PM IBM: Spyre integration with PyTorch and vLLM, plus insurance claim processing and document intelligence demos
- 2:15 PM Datadog: job-centric observability for Ray training failures
- 2:30 PM Modal: low-latency inference and LLM post-training via live Street Fighter III gameplay
- 3:55 PM Arm: connecting natural language requests to physical hardware with PyTorch, ExecuTorch, vLLM, and Arm Device Connect
- 4:10 PM Red Hat: quantization and speculative decoding under load, examining performance/cost/quality trade-offs
All sessions are held at the Community Expo on the Concourse Level.
More from Infra
- Anthropic plans $750B compute spend over 7-10 years; Meta reportedly drives a fifth of revenue — danshipper · 2026-10-21
- NVIDIA ships FP4-Local checkpoints of Qwen3.8-Flash-Next and GLM-5.3-Flash for local AI — NVIDIAAI · 2026-10-21
- Rumor: post-hoc GQA→MLA conversion cuts Qwen 27B KV-cache memory by up to 8.4x — ChrisGPT · 2026-10-21
- EAGLE-4 deep dive: up to 5.21× agentic coding speedup with lossless speculative decoding — zhyncs42 · 2026-10-21
- EAGLE-4 speculative decoding hits 961.6 tokens/s on Qwen3.8-27B, code open-sourced — zhyncs42 · 2026-10-21
- Bittensor miners speed up REAX decision model nearly 3× within one evening of competition — airesearch12 · 2026-10-21