From 647b30f7072e6c3a9240304d9100837cb7cca272 Mon Sep 17 00:00:00 2001 From: Kurt Heiss Date: Tue, 19 May 2026 11:13:14 -0700 Subject: [PATCH 1/5] Update release notes for 26.05 --- docs/docs/extraction/releasenotes.md | 48 +++++++++++++--------------- 1 file changed, 23 insertions(+), 25 deletions(-) diff --git a/docs/docs/extraction/releasenotes.md b/docs/docs/extraction/releasenotes.md index 9adb164a9e..ddb4992b8e 100644 --- a/docs/docs/extraction/releasenotes.md +++ b/docs/docs/extraction/releasenotes.md @@ -1,37 +1,35 @@ # Release Notes for NeMo Retriever Library -This documentation contains the release notes for [NeMo Retriever Library](overview.md). +This documentation contains the 26.05 Release Notes (26.5.0) for [NeMo Retriever Library](overview.md). -## 26.03 Release Notes (26.3.0) - -NVIDIA® NeMo Retriever Library version 26.03 adds broader hardware and software support along with many pipeline, evaluation, and deployment enhancements. +NVIDIA® NeMo Retriever Library version 26.05 builds on the 26.03 foundation with a graph-based ingest architecture, expanded multimodal and tabular capabilities, production-oriented service deployment, and documentation aligned to a Helm-first supported path. To upgrade the Helm charts for this release, refer to the [NeMo Retriever Library Helm Charts](https://github.com/NVIDIA/NeMo-Retriever/blob/main/nemo_retriever/helm/README.md). -Highlights for the 26.03 release include: - -- Legacy ingestion repository consolidated under NeMo-Retriever -- NeMo Retriever Extraction pipeline renamed to NeMo Retriever Library -- NeMo Retriever Library now supports two deployment options: - - A new no-container, pip-installable in-process library for development (available on PyPI) - - Existing production-ready Helm chart with NIMs -- Added documentation notes on Air-gapped deployment support -- Added documentation notes on OpenShift support -- Added support for RTX4500 Pro Blackwell SKU -- Added support for llama-nemotron-embed-vl-v2 in text and text+image modes -- New extract methods `pdfium_hybrid` and `ocr` target scanned PDFs to improve text and layout extraction from image-based pages -- VLM-based image caption enhancements: - - Infographics can be captioned - - Reasoning mode is configurable -- Enabled hybrid search with Lancedb -- Added retrieval_bench subfolder with generalizable agentic retrieval pipeline -- The project now uses UV as the primary environment and package manager instead of Conda, resulting in faster installs and simpler dependency handling -- Default TTL for long-running pipeline job state increased from 1–2 hours to 48 hours so long-running jobs (for example, VLM captioning) do not expire before completion -- NeMo Retriever Library currently does not support image captioning via VLM; this feature will be added in the next release -- Documentation: multimodal extraction is covered on one page with an in-page table of contents and redirects from the former per-topic URLs +Highlights for the 26.05 release include: + +- **Graph-based ingest pipeline** — `graph_pipeline` and the graph stage registry are the canonical ingestion path; mode-specific example scripts are consolidated around this model +- **Root CLI** — `retriever ingest` and `retriever query` with NIM URL flags, batch tuning, and LanceDB controls (overwrite/append) +- **Retriever Service v2** — scalable multi-pod architecture with gateway, process isolation, and VectorDB integration +- **Nemotron OCR v2** — default OCR engine with CLI language selectors and unified OCR actors +- **VLM image captioning** — image captioning via vLLM (including Omni caption model profiles); addresses the capability deferred in 26.03 +- **vLLM inference stack** — vLLM-backed text and vision-language embedders, multimodal VL reranker, and torch 2.11 stack for local GPU installs +- **Video retrieval pipeline** — frame extraction, OCR, audio-visual fusion, and text deduplication for video corpora +- **Text-to-SQL** — agent graph and tabular tooling for structured data retrieval +- **Live RAG SDK** — `Retriever.answer()` and optional batch operator graph via LiteLLM (`[llm]` extra) +- **Vector database** — VDB operators integrated directly in the pipeline; custom metadata support; LanceDB hybrid search guidance updated +- **Evaluation** — BEIR-centric evaluation overhaul; `retriever skill-eval` benchmark CLI for the NeMo Retriever skill +- **Packaging** — optional install extras (`[local]`, `[multimedia]`, `[llm]`, `[tabular]`, `[nemotron-parse]`, `[service]`, and others) including slim remote/NIM-only installs on Mac and Windows +- **Audio** — long-audio Parakeet chunking with time-aligned segments; punctuation-based audio segmenting +- **`allow_no_gpu`** — option to skip GPU requirement during ingest for CPU-only experimentation +- **Chunking API** — text splitting moved into `.extract(split_config=...)` +- **Documentation** — Helm-first deployment story; [Docker Compose for local development](https://github.com/NVIDIA/NeMo-Retriever/blob/main/nemo_retriever/docker.md) documented as **unsupported** developer tooling (not a production NIM deployment path) +- **Documentation** — duplicate user-defined stages page removed; UDF and custom stages guidance consolidated in the [graph README](https://github.com/NVIDIA/NeMo-Retriever/tree/main/nemo_retriever/src/nemo_retriever/graph#nemo-retriever-graph) +- **Documentation** — consolidated extraction concepts, ingest workflow, embeddings, and audio/video guides ## Release Notes for Previous Versions +| [26.03](https://docs.nvidia.com/nemo/retriever/26.3.0/extraction/releasenotes/) | [26.1.2](https://docs.nvidia.com/nemo/retriever/26.1.2/extraction/releasenotes/) | [26.1.1](https://docs.nvidia.com/nemo/retriever/26.1.1/extraction/releasenotes/) | [25.9.0](https://docs.nvidia.com/nemo/retriever/25.9.0/extraction/releasenotes/) From afb2668360c8934aa358c3e71ab0c9ed880a2a6f Mon Sep 17 00:00:00 2001 From: Kurt Heiss Date: Tue, 19 May 2026 11:17:03 -0700 Subject: [PATCH 2/5] Update RNs following greptile review --- docs/docs/extraction/releasenotes.md | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/docs/docs/extraction/releasenotes.md b/docs/docs/extraction/releasenotes.md index ddb4992b8e..a3d6038722 100644 --- a/docs/docs/extraction/releasenotes.md +++ b/docs/docs/extraction/releasenotes.md @@ -1,6 +1,8 @@ # Release Notes for NeMo Retriever Library -This documentation contains the 26.05 Release Notes (26.5.0) for [NeMo Retriever Library](overview.md). +This documentation contains the 26.05 Release Notes (26.5.0) for [NeMo Retriever Library](overview.md). + +## 26.05 Release Notes (26.5.0) NVIDIA® NeMo Retriever Library version 26.05 builds on the 26.03 foundation with a graph-based ingest architecture, expanded multimodal and tabular capabilities, production-oriented service deployment, and documentation aligned to a Helm-first supported path. From 98c4a24810dc3f6c921985b85ee06e6bfd9665bf Mon Sep 17 00:00:00 2001 From: Kurt Heiss Date: Tue, 19 May 2026 11:19:20 -0700 Subject: [PATCH 3/5] Update docs/docs/extraction/releasenotes.md Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> --- docs/docs/extraction/releasenotes.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/docs/extraction/releasenotes.md b/docs/docs/extraction/releasenotes.md index a3d6038722..000679a257 100644 --- a/docs/docs/extraction/releasenotes.md +++ b/docs/docs/extraction/releasenotes.md @@ -4,6 +4,8 @@ This documentation contains the 26.05 Release Notes (26.5.0) for [NeMo Retriever ## 26.05 Release Notes (26.5.0) +## 26.05 Release Notes (26.5.0) + NVIDIA® NeMo Retriever Library version 26.05 builds on the 26.03 foundation with a graph-based ingest architecture, expanded multimodal and tabular capabilities, production-oriented service deployment, and documentation aligned to a Helm-first supported path. To upgrade the Helm charts for this release, refer to the [NeMo Retriever Library Helm Charts](https://github.com/NVIDIA/NeMo-Retriever/blob/main/nemo_retriever/helm/README.md). From a24215a30887e2423faff5303541b7d78efd39bc Mon Sep 17 00:00:00 2001 From: Kurt Heiss Date: Tue, 19 May 2026 11:22:56 -0700 Subject: [PATCH 4/5] Updates per Greptile to 26.05 RNs --- docs/docs/extraction/releasenotes.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/docs/extraction/releasenotes.md b/docs/docs/extraction/releasenotes.md index a3d6038722..6c696269e0 100644 --- a/docs/docs/extraction/releasenotes.md +++ b/docs/docs/extraction/releasenotes.md @@ -1,6 +1,6 @@ # Release Notes for NeMo Retriever Library -This documentation contains the 26.05 Release Notes (26.5.0) for [NeMo Retriever Library](overview.md). +This documentation contains the release notes for [NeMo Retriever Library](overview.md). ## 26.05 Release Notes (26.5.0) From 015586193c117a2e67a817eb993964fbe15ec3d6 Mon Sep 17 00:00:00 2001 From: Kurt Heiss Date: Tue, 19 May 2026 11:26:44 -0700 Subject: [PATCH 5/5] again --- docs/docs/extraction/releasenotes.md | 2 -- 1 file changed, 2 deletions(-) diff --git a/docs/docs/extraction/releasenotes.md b/docs/docs/extraction/releasenotes.md index 18657da368..6c696269e0 100644 --- a/docs/docs/extraction/releasenotes.md +++ b/docs/docs/extraction/releasenotes.md @@ -4,8 +4,6 @@ This documentation contains the release notes for [NeMo Retriever Library](overv ## 26.05 Release Notes (26.5.0) -## 26.05 Release Notes (26.5.0) - NVIDIA® NeMo Retriever Library version 26.05 builds on the 26.03 foundation with a graph-based ingest architecture, expanded multimodal and tabular capabilities, production-oriented service deployment, and documentation aligned to a Helm-first supported path. To upgrade the Helm charts for this release, refer to the [NeMo Retriever Library Helm Charts](https://github.com/NVIDIA/NeMo-Retriever/blob/main/nemo_retriever/helm/README.md).