Backfill issue -436 — The AI Toolchain

EDITIONS tail grep head diff uniq

The daily firehose — everything the toolchain shipped today, already filtered.

// HOW THIS ISSUE IS MADE

We read every release from the 174 tools on our watchlist at the source — GitHub and GitLab release notes, vendor release pages and changelogs, project blogs and feeds, vendor press releases, and the source code behind the tag. Bug-fix-only releases and non-product newsroom noise are dropped; what's left is summarized down to the new capability, how to try it, and any screenshots or videos the release itself published. Every entry links to the sources it was built from.

VIEW

ISSUE VIEW full issue

Do you prefer this view?

$ tct list   # 2 tools matched

AI & LLM Tooling

◆ AI Model & Data Infrastructure

NVIDIA Triton Inference Server

v1.4.0 NOTES

Triton v1.4.0 adds PyTorch/TorchScript backend, Custom Backend SDK, and richer perf_client latency reporting.

└──▷ GET THIS VERSION

$ git clone --branch v1.4.0 https://github.com/triton-inference-server/server.git
# already have the repo? check out this version:
$ git checkout v1.4.0

›Adds --strict-model-config=false improvements to automatic model configuration creation, reducing manual config overhead.
›New libtorch backend enables running PyTorch models decorated or traced to TorchScript directly in the inference server.
›Custom Backend SDK now available with improved documentation to simplify building custom backends, distributed as v1.4.0_ubuntu1604.custombackend.tar.gz and v1.4.0_ubuntu1804.custombackend.tar.gz.
›perf_client now accepts user-supplied input data instead of being limited to random or zero values.
›perf_client now reports latency at multiple confidence percentiles (p50, p90, p95, p99) plus a user-supplied percentile that also drives latency stabilization.

+3 moreshow less

›C++ and Python client libraries now support specifying additional HTTP headers when using the HTTP protocol.
›Improved AsyncRun API in C++ and Python client libraries.
›Build system migrated from Bazel to CMake for improved portability and modularity.

Was this useful?

Other / Uncategorized

◆ VECTOR DB RAG

Weaviate

0.18.1 3 RELEASES · 2019-07-09 → 2019-07-30 NOTES STABLE

Weaviate 0.18.1 lets you combine vector explore and structured where filters in a single Get query.

└──▷ GET THIS VERSION

$ git clone --branch 0.18.1 https://github.com/weaviate/weaviate.git
# already have the repo? check out this version:
$ git checkout 0.18.1

›Enables combining the explore argument with a where argument inside a Get retrieval query, allowing vector-based ranking alongside exact string, keyword, or geo-spatial filtering in a single request.

2 more releases in this issue · 2019-07-09 → 2019-07-30

0.17.0 NOTES STABLE

Weaviate 0.17.0 adds vector indexing at import time and two new GQL vector-search surfaces: { Local { Explore }} and explore() in { Local { Get }}.

└──▷ GET THIS VERSION

$ git clone --branch 0.17.0 https://github.com/weaviate/weaviate.git
# already have the repo? check out this version:
$ git checkout 0.17.0

›Adds new GQL field { Local { Explore }} for concept (vector-based) search across Weaviate's vector search backend.
›Extends { Local { Get }} with a new explore() argument enabling vector-based concept search alongside the existing where() structured search.
›Imports through non-batch paths now generate and store a vector representation in Weaviate's vector database at import time.

└──▷ BREAKING ON UPGRADE

!Vector-based concept search requires a new esvector backend — see docker-compose/runtime/docker-compose.yml for a reference configuration.
!Vectors are only created at import time with no reindex capability, so upgrading to 0.17.x requires setting up a fresh Weaviate installation and reimporting all concepts.

0.16.0 NOTES STABLE

Weaviate 0.16.0 lets users supply their own UUIDs on POST and batch endpoints, simplifying cross-reference imports.

└──▷ GET THIS VERSION

$ git clone --branch 0.16.0 https://github.com/weaviate/weaviate.git
# already have the repo? check out this version:
$ git checkout 0.16.0

└──▷ TRY IT

Pre-assign a UUID when creating a thing so you can wire up cross-references without waiting for a server-generated ID.

$ curl -X POST http://localhost:8080/things \
  -H 'Content-Type: application/json' \
  -d '{"id": "a7e10b51-1f3e-4f5a-8d2e-000000000001", "class": "Article", "schema": {"title": "Example"}}'

›Adds user-specified UUID support: set the id field on POST /things, POST /actions, POST /batching/things, and POST /batching/actions to assign your own UUIDs instead of letting Weaviate generate them, making cross-reference preparation possible without round-tripping for server-assigned IDs.

Was this useful?

// END OF TRANSMISSION

The AI Toolchain · curated daily · July 30, 2019
Releases & features from the open-source and commercial AI tools that practitioners actually run.

The AI Toolchain — issue -436, July 30, 2019

NVIDIA Triton Inference Server

Weaviate