The official source reports this update: nightly-last-green: Reduce chunked cross-entropy memory test footprint (#8824). The original 8,192-token by 151,936-vocabulary test needs 4.64 GiB for each full FP32 logits-sized tensor. In the reproduced CI stall, a long-lived pytest worker held 28,832 MiB (28.2 GiB) of GPU memory. This PR reduces the token count from 8,192 to 128 while retaining the 151,936-entry vocabulary: each full FP32 tensor becomes 74.2 MiB, a 64-fold reduction. Preserve loss/gradient tolerances and the two-FP32-tensor memory-savings requirement, and release test-owned CUDA references and cache state on exit.
Canonical AI event
Nightly-last-green reduces memory footprint of cross-entropy test by 64-fold.
The memory footprint of the cross-entropy test was reduced from 4.64 GiB to 74.2 MiB per tensor.
15 Oct 20261 verified claims1 sources1 observations
This official update documents a development concerning nightly-last-green: Reduce chunked cross-entropy memory test footprint (#8824). Its practical significance depends on the scope and evidence stated by the source.
Read the official source update and verify its stated scope, evidence and timing before acting on it.
Connected knowledge
Entities affected by this event
Continue this topic
Related verified updates
Evidence trail
Sources behind the event
Editorial presentation
Open story