{"data":{"slug":"pytorch","etype":"software","name":"PyTorch","cat":"Training","subtitle":"The framework nearly every modern model is written in.","summary":"PyTorch won research with Pythonic eager execution, then won production with torch.compile and its distributed stack (FSDP). Born at Meta and governed since 2022 by the PyTorch Foundation under the Linux Foundation, it is the default in which GPT, Llama, Claude-class and most frontier models are implemented — and the reason new chips must run PyTorch before anyone buys them.","attrs":[{"g":"Overview","l":"Origin","v":"Meta AI (FAIR), 2016"},{"g":"Overview","l":"Governance","v":"PyTorch Foundation (Linux Foundation) since 2022"},{"g":"Overview","l":"Key pieces","v":"torch.compile, FSDP distributed training, TorchInductor"}],"kf":{"maker":"Meta → PyTorch Foundation","since":"2016","license":"BSD"},"aliases":["torch"],"ticker":null,"exchange":null,"hq":null,"founded":null,"website":null,"status":"Active","region":null,"asof":"2026-01-15","sources":[{"url":"https://pytorch.org/","label":"PyTorch"}]},"meta":{"found":true,"count":1,"served_from":"site cache (no live upstream calls)","as_of":"2026-10-14T06:34:13+00:00","disclaimer":"Research and information only — not investment advice. Structured records carry basis labels (vendor spec / reported / estimate); market quotes are delayed/indicative.","attribution":"Records compiled from vendor documentation, official filings and reputable public reporting; market quotes via financial data providers. See /data-sources."}}