Changes related to Inference

• Use direct UNIX socket for container stats by @sberan in #1463
• Fix Moondream2 by @capjamesg in #1445
• Fix/nvidia t4 parallel by @grzegorz-roboflow in #1467

Changes related to Inference-Experimental

• inference-exp docker builds CI refinements by @grzegorz-roboflow in #1451
• fixes due to updates to Roboflow-GPU-VM-Runner by @grzegorz-roboflow in #1465
• Add changes to align inference-exp with refrence RFDetr implementation by @PawelPeczek-Roboflow in #1459

Full Changelog: v0.51.9...v0.51.10

Inference

• Expose processing time header by @stellasphere in #1449
• Fix countinference issue by @bigbitbus in #1452
• Fix rfdetr eps by @probicheaux in #1454
• Install inference-cli before forcing numpy to prevent overwriting with numpy 2 by @grzegorz-roboflow in #1441
• Fix/force numpy 1.26.4 on jetson dockers by @grzegorz-roboflow in #1457

inference-exp

• Fix post-processing for instance-segmentation in inference-exp by @PawelPeczek-Roboflow in #1440
• Publish intefence exp dockers by @grzegorz-roboflow in #1443
• inference-exp: Florence 2 LoRA by @hansent in #1442

Full Changelog: v0.51.7...v0.51.9

What's Changed

• Add new model size to rf-detr by @Matvezy in #1446

Full Changelog: v0.51.6...v0.51.7

Changes to Inference:

• Webrtc latency for heavy workflows by @grzegorz-roboflow in #1444
• Export PYTHONPATH so inference is picked up correctly by inference-cli by @grzegorz-roboflow in #1439

Changes to Inference-Experimental:

• Improve readme to nicely list registered models by @PawelPeczek-Roboflow in #1437
• Apply changes such that MPS is correctly supported for rfdetr, yet still remain non-default for MacOS by @PawelPeczek-Roboflow in #1436

Full Changelog: v0.51.5...v0.51.6

What's Changed

Changes to Inference:

• Add inference cli to jetson dockers by @grzegorz-roboflow in #1429
• Make inference API urls configurable for workflow remote execution by @bigbitbus in #1428
• Set TTL for validated models to 1 hour by @bigbitbus in #1426

Changes to Inference-Experimental:

• Fix issue with auto-loader, adjust project.toml to use new supervision and kick-off readme by @PawelPeczek-Roboflow in #1423
• rename from perception_encoder to perception-encoder by @hansent in #1427
• add CLIP w/ TORCH backend to inference_experimental by @hansent in #1415
• Improve README in terms of installation instructions and clarity by @PawelPeczek-Roboflow in #1431
• Fix/add cpu only tests marker by @grzegorz-roboflow in #1432
• fix CUDA issue on inference-experimental GPU CI by @grzegorz-roboflow in #1419
• Add improvements to auto-negotiation errors clarity by @PawelPeczek-Roboflow in #1434

Full Changelog: v0.51.4...v0.51.5

What's Changed

• add upper bound on supervision version at 0.26.0 by @grzegorz-roboflow in #1424

Full Changelog: v0.51.3...v0.51.4

What's Changed

• Add webrtc data channel by @grzegorz-roboflow in #1406
• Fix OSX app build / windows installer dependency issue introduced in 0.51 by @hansent in #1413
• Add x-processing-time header to response when running on GCP_SERVERLESS by @grzegorz-roboflow in #1416
• When wildcard output is specified, None is sent as stream output - pick first video output by @grzegorz-roboflow in #1414
• Unit tests ci/inference v1 models by @grzegorz-roboflow in #1411
• Test most current fastapi and typer pins by @grzegorz-roboflow in #1407
• add perception encoder to inference_experimental by @hansent in #1399
• Add first scratches of new interface by @PawelPeczek-Roboflow in #1250

Full Changelog: v0.51.2...v0.51.3

What's Changed

• Improve artifact download reliability by @sberan in #1397
• Add additional content length checks for file downloads by @sberan in #1405

Full Changelog: v0.51.1...v0.51.2

⚡ Speed-ups (with https://github.com/codeflash-ai)

• ⚡️ Speed up method RFDETRObjectDetection.postprocess by 257% by @misrasaurabh1 in #1381

💪 Added

• Add variable speed simulation by @lou-roboflow in #1392
• Add new Moondream2 model by @capjamesg in #1388
• Allow dataset upload tags from list selector by @NickHerrig in #1364

🧑‍🏭 Maintanence

• Add fix for the problem with JSON object as default vaule for WorkflowParameter by @PawelPeczek-Roboflow in #1379
• Bump peft with fixed dependencies by @probicheaux in #1387
• Feat/pass countinference to serverless getweights by @bigbitbus in #1373
• Add tests covering rfdetr sigmoid_stable by @grzegorz-roboflow in #1394
• Install libsqlite3-dev before building python3.9 on jetson 4.6.1 by @grzegorz-roboflow in #1383
• Suppress missing dependencies warnings when starting jetson 4.6.1 by @grzegorz-roboflow in #1389

Full Changelog: v0.51.0...v0.51.1

What's Changed

• Add perception encoder workflow blocks by @hansent in #1350

• Feat/stability ai outpainting by @grzegorz-roboflow in #1367

• Respect image parent_id property when deserializing by @brunopicinin in #1363

• Fix zone filter anchor extraction for non-center detection reference points by @ryanjball in #1372

• Fix 'Invalid date error' when creating new workflow by @grzegorz-roboflow in #1365

• Fix velocity block - sv.Detections[data] items must be np.array elements by @grzegorz-roboflow in #1368

• communicate failed pipeline over webrtc by @grzegorz-roboflow in #1376

• Update owlv2.py by @stellasphere in #1375

• docs: Add installation and a quickstart code snippet in the getting-started page in docs by @soumik12345 in #1362

• Change ONVIF Block Name by @lou-roboflow in #1371

New Contributors

• @soumik12345 made their first contribution in #1362

Full Changelog: v0.50.5...v0.51.0