v1.3.6

# 🚀 Added


## 🧠 NVIDIA Cosmos 3 Edge — initial preview

<img width="977" height="449" alt="image" src="https://github.com/user-attachments/assets/73fc673b-afa8-4f7d-b1f1-37c38d5921cc" />


This release ships an initial preview of the [NVIDIA Cosmos 3 Edge](https://huggingface.co/nvidia/Cosmos3-Edge) model family (https://github.com/roboflow/inference/pull/2675):

- **Reasoning stack in Workflows.** The new `roboflow_core/cosmos3_edge@v1`  block exposes the Cosmos 3 Edge reasoner (VLM) in Workflows — single-image  and multi-frame reasoning emitting `language_model_output`, so it chains  directly into the existing VLM tooling (e.g. *VLM as Detector*). The block  is GPU-gated.
- **Available on the Roboflow platform.**
- **Standalone GPU build available as a trial:**

  ```bash
  docker pull roboflow/roboflow-inference-server-gpu:1.3.6-cosmos3
  ```

The standard release images ship without the Cosmos 3 dependency stack — the trial build above is the way to run it self-hosted during the preview. Under the hood the preview also lands the Cosmos 3 Edge world-model surface (image-to-video generation, forward/inverse dynamics) in `inference_models`, ahead of an HTTP/Workflows surface for generative outputs in a future release.


## ⚡ RF-DETR object detection — TensorRT pre/post-processing speed-up

RF-DETR object detection on the TensorRT backend now runs GPU-accelerated (Triton-kernel) image pre-processing and fused post-processing, removing the CPU bottleneck around the TensorRT forward pass.

Measured end-to-end latency (Orin AGX, JetPack 6.2, TRT fp16 package):

| Scenario | before (mean) | after (mean) | speed-up |
|---|---|---|---|
| 3840×2160, batch 4 | 436.8 ms | 36.0 ms | ~12× |
| 640×480, batch 1 | 17.3 ms | 7.4 ms | ~2.3× |

**Where it applies:**

- **GPU builds and Jetson JetPack 6+.** CPU builds and older JetPacks  (JetPack 5 and below) keep the existing pipeline, unchanged.
- **Selected pre-processing pathways only.** The accelerated path engages for  model packages using plain **stretch resize** with standard 3-channel input  and per-channel normalization. Models configured with other resize modes  (letterbox / fit), dataset-version resize dimensions, or additional image  transforms (static crop, contrast, grayscale, auto-orient) automatically and  transparently fall back to the existing base implementation — no behavior  change for those models.

Selection is automatic with conservative compatibility checks and produces identical predictions to the base pipeline. For explicit control (pinning or disabling per deployment) use the environment variables `INFERENCE_MODELS_RFDETR_PREPROCESSOR` and `INFERENCE_MODELS_RFDETR_POSTPROCESSOR` (values: `base`, `threaded-exact-v1`, `triton-universal-v1` / `triton-fused-v1`).

## 🔥 New workflow blocks

- **Auto Rotate on Edges** (`roboflow_core/auto_rotate_on_edges@v1`) — rotates  an image so its dominant straight lines become vertical, horizontal, or the  nearest axis (@jeku46, https://github.com/roboflow/inference/pull/2655).  Built for line-dominated inputs that arrive skewed — industrial X-rays,  documents, labels, shelves — with a single-pass gradient-histogram estimate,  sub-degree refinement, and identity-passthrough guards for flat or  orientation-ambiguous images. Outputs the rotated image and the applied  angle.
- **Frame Delay** (`roboflow_core/frame_delay@v1`) — returns any workflow  value (detections, numbers, images, …) as it was `|offset|` frames ago on  the same video stream, enabling cross-frame comparison and temporal  alignment (@rafel-roboflow, https://github.com/roboflow/inference/pull/2668).  Memory-bounded per-stream ring buffer; past-only by design; works in every  execution context including the WebRTC video path.

## 🔭 Observability

- **CUDA allocator memory breakdown** on `/model/registry` — live tensor  allocations vs. PyTorch-reserved vs. allocator cache vs. non-PyTorch device  memory, so production incidents can tell allocator growth from real model  memory (@hansent, https://github.com/roboflow/inference/pull/2657).
- **Per-pipeline stream session id in usage tracking** — each  `InferencePipeline` now carries a stable stream session identity (callers  may supply their own, e.g. `DEVICE_ID:stream_name`), so concurrent pipelines  under one API key and workflow no longer merge in usage aggregation  (@sberan, https://github.com/roboflow/inference/pull/2634).

# 🔧 Fixed

- **TrOCR now served through the `inference_models` adapter by default** — the  original HuggingFace package path hit tokenizer-compatibility issues with  newer `transformers`; the adapter implementation is compatible and is now  the default (@PawelPeczek-Roboflow,  https://github.com/roboflow/inference/pull/2681).
- **SAM3 concept-path post-processing memory bounded** — the upstream  post-processor interpolated every kept mask to full resolution in one batch  and applied the detection cap last, spiking multi-GiB host/GPU peaks on  large images (a direct contributor to a serverless OOM crash loop). The  pipeline is now chunked and cap-first, bounding peak memory regardless of  instance count (@bigbitbus, https://github.com/roboflow/inference/pull/2670).
- **OWLv2 embeddings cache fixes** — re-signed image URLs no longer defeat the  cache (https://github.com/roboflow/inference/pull/2659), and the cache is  consulted before reference images are materialized, skipping redundant  downloads entirely (@bigbitbus,  https://github.com/roboflow/inference/pull/2660).
- **Secure-gateway routing audit** — every outbound HTTP call in `inference/`  and `inference_models/` was audited for `SECURE_GATEWAY` compatibility;  three gaps fixed, including the GitHub version check stalling server startup  behind a gateway (now timeout-bounded and auto-disabled) and workflow remote  step execution dead-ending against hosted endpoints (@alexnorell,
  https://github.com/roboflow/inference/pull/2658).
- **Inner-workflow validation errors return HTTP 400, not 500** — compile-time  child-workflow problems (stale parameter bindings, invalid nesting, cycles)  are now correctly classified as client errors in both sync and async route  handlers (@dkosowski87, https://github.com/roboflow/inference/pull/2645).
- **Pre-release security patches** rolled up ahead of the release  (@PawelPeczek-Roboflow, https://github.com/roboflow/inference/pull/2688).

# 🚧 Maintenance

- **JetPack 7.2 build workflow** — groundwork for upcoming JetPack 7.2 server  images (@alexnorell, https://github.com/roboflow/inference/pull/2654).
- **RF-DETR server integration tests** — all RF-DETR model aliases (detection  + segmentation) now covered against both legacy and v1 server endpoints, in  both `USE_INFERENCE_MODELS` modes (@PawelPeczek-Roboflow,  https://github.com/roboflow/inference/pull/2673).
- **New unit-test CI workflow** for the model manager and inference server  (@grzegorz-roboflow, https://github.com/roboflow/inference/pull/2669), and  right-sized CI runners for the dev-test workflow (@iurisilvio,  https://github.com/roboflow/inference/pull/2665).
---

**Full changelog:** [`v1.3.5...v1.3.6`](https://github.com/roboflow/inference/compare/v1.3.5...v1.3.6)