pullrequests Search Results · bucket language:JavaScript language:JavaScript language:Java language:Python language:JavaScript
Filter by
1M results
... (#574) already moved the agent docs to tigris buckets lifecycle create, the plural set-cors flags, the five
set-migration flags and the mk + set-migration --write-through --region auto flow for R2. The ...
Motivation
get_batch_sizes_to_capture drops the top decode bucket when --cuda-graph-max-bs is below the request pool and not a
multiple of the attention-TP alignment. Example: DP attention with attn ...
blackwell
lora
Flat per-bucket capturable Adam CUDA kernel at the DRAM rate (replaces 150 multi_tensor_apply launches).
Tests: see commits
Stacked on #8135 (which is stacked on #8121): the first commits in this ...
Summary
- gmm_v2 computes wrong rows when a tiling splits the contraction (tile_k K) and a group ends inside a row tile (a row
bucket below tile_m, which happens whenever tile_m bucket ...
... actually purge from data still held by snapshots. HDDS-17502 added the
GetBucketDeletedBytes OM API, which returns the purgeable vs. snapshot-held split per bucket. This PR uses that API to
report OM ...
recon
...
How it solves it:
- The window and pools are keyed by provider bucket: the organization of the deployment that serves the call plus its
model class
- Admission moves to the router stage that ...
:earth_americas: Summary
draw_chart gains four optional parameters so it can draw a per-subject observations over time chart: one stacked,
scrolling panel per category value, empty time buckets ...
🚨 Severity: HIGH
💡 Vulnerability: get_user_id used get_unverified_claims to extract the user ID for rate limiting, allowing an attacker
to spoof tokens and poison other users rate limit buckets ...
Summary
- enforce a hard maxEntries bound for in-memory rate-limit buckets
- evict expired buckets first, then the oldest bucket when admitting a new client key
- validate invalid limiter ...