Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
255 commits
Select commit Hold shift + click to select a range
0719780
Try to fix build
PawelPeczek-Roboflow May 21, 2026
948c4de
Add support for tensor-native interface for models
PawelPeczek-Roboflow May 21, 2026
0d5e48f
Compress build steps
PawelPeczek-Roboflow May 22, 2026
d46021d
Try to fix build
PawelPeczek-Roboflow May 22, 2026
b33b5cc
Try to fix build
PawelPeczek-Roboflow May 22, 2026
1f01890
workflows: WorkflowImageData hosts tensor_image alongside numpy_image
PawelPeczek-Roboflow May 22, 2026
c88db40
workflows: add tensor-mode (de)serializer siblings for loader swap
PawelPeczek-Roboflow May 22, 2026
5082dc5
workflows: add common/utils_tensor.py mirror for image-shape utils
PawelPeczek-Roboflow May 22, 2026
8a92b2f
inference-models adapters: stop forcing input_color_format on tensor …
PawelPeczek-Roboflow May 22, 2026
3c0ff2d
workflows: pilot tensor-native sibling for object detection v3
PawelPeczek-Roboflow May 22, 2026
9329932
workflows: loader swaps OD v3 + image (de)serializers under tensor flag
PawelPeczek-Roboflow May 22, 2026
24bcbee
WIP - safe commit
PawelPeczek-Roboflow May 22, 2026
f3f0483
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow May 26, 2026
8bd6495
First sketch of the kinds
PawelPeczek-Roboflow May 26, 2026
a189759
Add scratch of changes to tensor-native kinds
PawelPeczek-Roboflow May 26, 2026
f74ffed
Port absolute static crop
PawelPeczek-Roboflow May 26, 2026
b671460
Bring bounding rect block
PawelPeczek-Roboflow May 26, 2026
1aa9abf
Add byte trackers
PawelPeczek-Roboflow May 27, 2026
c009666
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jun 10, 2026
0a9c133
WIP - safe commit
PawelPeczek-Roboflow Jun 10, 2026
1183bcf
Patch classification UQL ops
PawelPeczek-Roboflow Jun 10, 2026
450e21e
Add changes to UQL detection operations
PawelPeczek-Roboflow Jun 11, 2026
6735298
Add detections UQL operations
PawelPeczek-Roboflow Jun 11, 2026
ae3e96f
Add detections UQL operations
PawelPeczek-Roboflow Jun 11, 2026
4639c41
Add changes regarding inference-models iterators and detection-event-log
PawelPeczek-Roboflow Jun 11, 2026
bb514a4
Add changes regarding inference-models iterators and detection-event-log
PawelPeczek-Roboflow Jun 11, 2026
e2eba1c
Add changes to enable tensor-based line-counter
PawelPeczek-Roboflow Jun 11, 2026
757a5ea
Add changes regarding remaining analytics blocks
PawelPeczek-Roboflow Jun 11, 2026
71eec3d
Add tensor-native variants of computer-vision blocks
PawelPeczek-Roboflow Jun 11, 2026
9fa439c
Add alignment of fusion and formatters blocks
PawelPeczek-Roboflow Jun 11, 2026
d5cb5a5
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jun 11, 2026
90b5bd2
Add changes for embedding-related blocks
PawelPeczek-Roboflow Jun 11, 2026
d788377
Migrate trackers and visualisations
PawelPeczek-Roboflow Jun 11, 2026
b8e4410
Add remaining changes
PawelPeczek-Roboflow Jun 11, 2026
e4ca396
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jun 11, 2026
b968d00
Add loader to new blocks
PawelPeczek-Roboflow Jun 11, 2026
638e203
Add CLIP model
PawelPeczek-Roboflow Jun 11, 2026
150ab07
Add CLIP model
PawelPeczek-Roboflow Jun 11, 2026
9a35ea7
Add depth estimation migration
PawelPeczek-Roboflow Jun 11, 2026
3bff2cb
Add new models
PawelPeczek-Roboflow Jun 11, 2026
64185dc
Add florence2 compatibilitry
PawelPeczek-Roboflow Jun 15, 2026
d5c7c39
Add changes in data representation
PawelPeczek-Roboflow Jun 15, 2026
3dd2ac9
Add QWEN model
PawelPeczek-Roboflow Jun 15, 2026
27a5e92
Add more models
PawelPeczek-Roboflow Jun 15, 2026
b79231a
Add more models
PawelPeczek-Roboflow Jun 15, 2026
29f88f2
Add SAM3 worklflow blocks
PawelPeczek-Roboflow Jun 16, 2026
2b746e3
Add adjustment to SAM3 and deprecate YoloWorld
PawelPeczek-Roboflow Jun 16, 2026
15c4b32
Add adjustments for SAM3
PawelPeczek-Roboflow Jun 16, 2026
33a5063
Add 3rd party blocks
PawelPeczek-Roboflow Jun 16, 2026
174eb42
Add parity for RF models blocks
PawelPeczek-Roboflow Jun 16, 2026
0239fe5
Add remaining blocks
PawelPeczek-Roboflow Jun 16, 2026
ba02365
Add fixes
PawelPeczek-Roboflow Jun 16, 2026
6aa2398
Merge with main
PawelPeczek-Roboflow Jun 16, 2026
4a77593
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jun 17, 2026
bbc8af7
Add changes to be able to disable unwanted integration tests
PawelPeczek-Roboflow Jun 17, 2026
8f4009b
Make linters happy
PawelPeczek-Roboflow Jun 17, 2026
6e99919
Add fixes for unit tests
PawelPeczek-Roboflow Jun 17, 2026
aa581b9
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jun 17, 2026
8a5aa56
Make linters happy
PawelPeczek-Roboflow Jun 17, 2026
0d1b4c7
WIP - safe commit
PawelPeczek-Roboflow Jun 17, 2026
b4894d9
WIP - safe commit
PawelPeczek-Roboflow Jun 17, 2026
95d10e2
WIP - safe commit
PawelPeczek-Roboflow Jun 17, 2026
d80faef
WIP - safe commit
PawelPeczek-Roboflow Jun 17, 2026
e6359e6
WIP - safe commit
PawelPeczek-Roboflow Jun 17, 2026
0af323f
Establish tests in ci
PawelPeczek-Roboflow Jun 17, 2026
e6d357d
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jun 17, 2026
8adb0d8
Make linters happy
PawelPeczek-Roboflow Jun 18, 2026
88d567a
Bump inference-models version
PawelPeczek-Roboflow Jun 18, 2026
2c2ecae
Bump inference-models in inference deps
PawelPeczek-Roboflow Jun 18, 2026
e46082b
Adjust changelog
PawelPeczek-Roboflow Jun 18, 2026
5191de4
Bump inference version
PawelPeczek-Roboflow Jun 18, 2026
565a2e4
Fix issue with EasyOCR results parity
PawelPeczek-Roboflow Jun 18, 2026
1269fa3
Alihgn SAGI
PawelPeczek-Roboflow Jun 18, 2026
ff8dcc6
Align SAM2
PawelPeczek-Roboflow Jun 18, 2026
6076946
Align integration tests
PawelPeczek-Roboflow Jun 18, 2026
77c0874
Make linters happy
PawelPeczek-Roboflow Jun 18, 2026
96aaa11
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jun 18, 2026
f160155
Add benchmark scripts
PawelPeczek-Roboflow Jun 18, 2026
e16a3bf
Porvide soft-torch tensor enforcement
PawelPeczek-Roboflow Jun 18, 2026
631b07a
Extend coverage for tensor inputs
PawelPeczek-Roboflow Jun 18, 2026
40e3b62
Add test coverage for torch.Tensor inputs
PawelPeczek-Roboflow Jun 18, 2026
46f63b8
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jun 18, 2026
3bf97b4
Add extension to inference pipeline
PawelPeczek-Roboflow Jun 19, 2026
258c73b
WIP
PawelPeczek-Roboflow Jun 19, 2026
5c900de
Add changes to align tensor-native processing with crops and stitches
PawelPeczek-Roboflow Jun 22, 2026
295029b
Add pynvvideocodec to requirements
PawelPeczek-Roboflow Jun 22, 2026
aed333a
Add fix
PawelPeczek-Roboflow Jun 22, 2026
e315022
Add Compact Mask support in visualisations blocks
PawelPeczek-Roboflow Jun 23, 2026
fa3f1f7
Remove comments
PawelPeczek-Roboflow Jun 24, 2026
2486e55
Upgrade Jetpack 5 build
PawelPeczek-Roboflow Jun 24, 2026
bb66655
Fix image layout bug
PawelPeczek-Roboflow Jun 25, 2026
05051ac
Modify script to enable recording
PawelPeczek-Roboflow Jun 25, 2026
aa5d895
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jul 7, 2026
e9f142f
Bump inference-models dependencieS
PawelPeczek-Roboflow Jul 7, 2026
b15e77e
Add coverage for missing blocks
PawelPeczek-Roboflow Jul 7, 2026
394dbf6
Resolve conflicts with main
PawelPeczek-Roboflow Jul 7, 2026
d083a8a
Bump dependency to inference-models
PawelPeczek-Roboflow Jul 7, 2026
a933509
Add fixes for problems
PawelPeczek-Roboflow Jul 7, 2026
adb1ad9
Step 1 of custom-python migration
PawelPeczek-Roboflow Jul 7, 2026
07b264a
Step 2 of custom-python migration
PawelPeczek-Roboflow Jul 7, 2026
6d72fa2
Step 3 of custom-python migration
PawelPeczek-Roboflow Jul 7, 2026
ce3c05a
Step 4 of custom-python migration
PawelPeczek-Roboflow Jul 7, 2026
22bc520
Step 5 of custom-python migration
PawelPeczek-Roboflow Jul 7, 2026
3dc5f15
Step 6 of custom-python migration
PawelPeczek-Roboflow Jul 7, 2026
cada1a5
Step 7 of custom-python migration
PawelPeczek-Roboflow Jul 7, 2026
ee872d6
Make linters happy
PawelPeczek-Roboflow Jul 8, 2026
9a90b8c
Add fixes for CI issues
PawelPeczek-Roboflow Jul 8, 2026
2bfae89
Fix for CI bug
PawelPeczek-Roboflow Jul 8, 2026
0bc6051
Try to produce missing JP builds
PawelPeczek-Roboflow Jul 9, 2026
12a43fb
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jul 9, 2026
36f6072
Try to fix JP 7 build
PawelPeczek-Roboflow Jul 9, 2026
a02b456
Fix to docker build of JP6 and 7
PawelPeczek-Roboflow Jul 9, 2026
9e6b908
Fix to docker build of JP7
PawelPeczek-Roboflow Jul 9, 2026
a54f6a8
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jul 13, 2026
ca1b4d7
Align uv lock
PawelPeczek-Roboflow Jul 13, 2026
6121517
Fix the test
PawelPeczek-Roboflow Jul 13, 2026
94cfaad
Adjust CI
PawelPeczek-Roboflow Jul 13, 2026
2d663db
Upgrade test script and fix requirements
PawelPeczek-Roboflow Jul 13, 2026
94f1a8a
Add fixes to the CR comments
PawelPeczek-Roboflow Jul 13, 2026
bbb60fb
Try to fix JP 7 builds
PawelPeczek-Roboflow Jul 13, 2026
59ab92f
Add accelerated tensor-native media runtimes (#2648)
alexnorell Jul 14, 2026
7021ed2
Add changes to data source to materliasied more aligned to setup in env
PawelPeczek-Roboflow Jul 14, 2026
974d4f7
Adjust mutation behaviour of WorkflowImageData
PawelPeczek-Roboflow Jul 14, 2026
4752a59
Propagete the change to old blocks
PawelPeczek-Roboflow Jul 14, 2026
7bb6000
Align elements selection
PawelPeczek-Roboflow Jul 14, 2026
f11f1f9
Upgrade velocity block
PawelPeczek-Roboflow Jul 14, 2026
fa6c4aa
Fix reference path
PawelPeczek-Roboflow Jul 14, 2026
0df04b2
Add tensor parity for contrast enhancement
PawelPeczek-Roboflow Jul 14, 2026
6aad370
Add follow up changes
PawelPeczek-Roboflow Jul 14, 2026
fd5cd42
Add contrast equalisation
PawelPeczek-Roboflow Jul 14, 2026
1efe15f
Add more tensor-native blocks
PawelPeczek-Roboflow Jul 14, 2026
47ac1b6
Fix flaky tests
PawelPeczek-Roboflow Jul 14, 2026
b4c421d
Fix tests
PawelPeczek-Roboflow Jul 14, 2026
1412e2b
Make linters happy
PawelPeczek-Roboflow Jul 14, 2026
489883f
Optimize tensor-native blocks: analytics, overlap, consensus, stitch …
Erol444 Jul 14, 2026
6cbad52
GPU-native bounding box visualization for the tensor pipeline (3-10x …
Erol444 Jul 14, 2026
6c4671a
Fix Jetson OpenCV NumPy 2 ABI
alexnorell Jul 14, 2026
336b631
Generate PyTorch AOT shims in Jetson build
alexnorell Jul 14, 2026
14898ae
Set PyTorch code generator import path
alexnorell Jul 14, 2026
8a5b54d
Reset stale Jetson PyTorch build cache
alexnorell Jul 14, 2026
fb45683
Build Jetson OpenCV with Python limited API
alexnorell Jul 14, 2026
36391a1
Use NumPy 2 compatible OpenCV binding on Jetson 6
alexnorell Jul 14, 2026
85c50f3
Add Jetson NVJPEG runtime dependency
alexnorell Jul 14, 2026
f0f4134
Replace Jetson 7.1 server image with JetPack 7.2
alexnorell Jul 14, 2026
025ef64
Publish full JetPack 7.2 inference server
alexnorell Jul 14, 2026
f095f33
Build native Orin attention kernels
alexnorell Jul 14, 2026
66abfbd
Reject pip CUDA runtimes in Jetson image
alexnorell Jul 14, 2026
a533235
Fix FlashAttention SM87 patch whitespace
alexnorell Jul 14, 2026
9d811e7
Fix review findings in tensor media runtimes
alexnorell Jul 14, 2026
99fab84
Install Python and pip in Jetson 6.2 runtime
alexnorell Jul 14, 2026
310dab2
Copy pip launchers into Jetson 6.2 runtime
alexnorell Jul 14, 2026
bf6e2bb
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jul 14, 2026
5106b43
Vump timoeut
PawelPeczek-Roboflow Jul 15, 2026
b362741
fix jp6 build
PawelPeczek-Roboflow Jul 15, 2026
7afc752
Make linters happy
PawelPeczek-Roboflow Jul 15, 2026
7178a08
Fix JP6.2 build
PawelPeczek-Roboflow Jul 15, 2026
cdac588
Try to Fix JP6.2 build
PawelPeczek-Roboflow Jul 15, 2026
a23ec78
Bump opencv requirements
PawelPeczek-Roboflow Jul 15, 2026
83ced04
Bump opencv requirements
PawelPeczek-Roboflow Jul 15, 2026
97f0f08
CR PR2357: make torch/inference_models import-optional off-flag (FQ-3)
PawelPeczek-Roboflow Jul 15, 2026
c27c736
CR PR2357: D1 polygon-serialisation flag + D4 vlm classification byte…
PawelPeczek-Roboflow Jul 15, 2026
220cc42
CR PR2357: PP-OCR tensor sibling + classification/embedding/tensor de…
PawelPeczek-Roboflow Jul 15, 2026
aef30d7
CR PR2357: viz tensor byte-parity + mask-materialisation perf
PawelPeczek-Roboflow Jul 15, 2026
55d10f0
CR PR2357: media bridge Python halves - bounded grab + interrupt/clos…
PawelPeczek-Roboflow Jul 15, 2026
6fe1a15
CR PR2357: address review - drop Owner-1 import deferrals + fork/spaw…
PawelPeczek-Roboflow Jul 15, 2026
af33861
CR PR2357 Wave2: fix Florence-2 tensor task-dict unwrap (BLOCKER)
PawelPeczek-Roboflow Jul 15, 2026
75c8f10
CR PR2357 Wave2: align semseg tensor exclusion set to numpy + restore…
PawelPeczek-Roboflow Jul 15, 2026
d4fc888
CR PR2357 Wave2: instance-seg tensor REMOTE polygon fallback (older s…
PawelPeczek-Roboflow Jul 15, 2026
c09a239
CR PR2357 Wave2: SAM enforce-RLE + revert collapse_to_most_confident
PawelPeczek-Roboflow Jul 15, 2026
9a9b67b
CR PR2357 Wave2: fusion class_id-collision flag + consensus mask devi…
PawelPeczek-Roboflow Jul 15, 2026
6f8dc66
CR PR2357 Wave2: pick_detections_by_parent_class - do not raise on in…
PawelPeczek-Roboflow Jul 15, 2026
83f3951
CR PR2357 Wave2: model_monitoring tensor telemetry parity (D5)
PawelPeczek-Roboflow Jul 15, 2026
75dd59f
CR PR2357 Wave2: perspective_correction keypoint parity + CENTER_OF_M…
PawelPeczek-Roboflow Jul 15, 2026
0939c44
CR PR2357 Wave2: supervision floor, iter-order tests, easy_ocr text, …
PawelPeczek-Roboflow Jul 15, 2026
fee2ab2
CR PR2357 Wave2 fix: SAM manifest parity — mirror numpy output field,…
PawelPeczek-Roboflow Jul 15, 2026
084c7e1
CR PR2357 Wave2 fix: instance-seg v4 REMOTE polygon fallback (parity …
PawelPeczek-Roboflow Jul 15, 2026
5616ca6
CR PR2357 lane 1b: explicit CLASSIFICATION_STYLE_KEY replaces the sha…
PawelPeczek-Roboflow Jul 15, 2026
0ad1efb
CR PR2357: black-format deserializers_tensor.py (static_code_analysis…
PawelPeczek-Roboflow Jul 15, 2026
21a25b7
CR PR2357: black + isort the tensor-branch style debt (static_code_an…
PawelPeczek-Roboflow Jul 15, 2026
cf30fe9
Make linters happy with tests modukes
PawelPeczek-Roboflow Jul 15, 2026
ddd155b
CR PR2357: guard tensor-only test_base test under flag-OFF (unit_test…
PawelPeczek-Roboflow Jul 15, 2026
c5d6060
media: fix GPU build CC floor + HW decoder routing
PawelPeczek-Roboflow Jul 15, 2026
03b15c7
test(pp_ocr): tensor-native parity for the extracts-text integration …
PawelPeczek-Roboflow Jul 15, 2026
29f544e
Increase timeout
PawelPeczek-Roboflow Jul 15, 2026
38cefb4
fix(docker): restore builder-stage python symlink in onnx.gpu image
PawelPeczek-Roboflow Jul 15, 2026
f108737
fix(docker): strip leftover ninja/cmake/meson from onnx.gpu runtime PATH
PawelPeczek-Roboflow Jul 15, 2026
5f833c6
jp72: split framework wheel builds into a component image
alexnorell Jul 16, 2026
7df23d8
fix(jp72): install cmake and ninja in wheels toolchain
alexnorell Jul 16, 2026
42212ff
fix(docker): assert the tensor bridge lands in the JP6.2 runtime image
PawelPeczek-Roboflow Jul 16, 2026
22e3d2c
fix(docker): install kmod in the JP6.2 runtime image
PawelPeczek-Roboflow Jul 16, 2026
1da4055
fix(docker): assert lsmod/modprobe exist in the JP6.2 runtime image
PawelPeczek-Roboflow Jul 16, 2026
523794a
dev: print the watchdog latency split in run_workflow_on_video summary
PawelPeczek-Roboflow Jul 16, 2026
c18da59
Revert "dev: print the watchdog latency split in run_workflow_on_vide…
PawelPeczek-Roboflow Jul 16, 2026
b03289a
fix(jetson): move tensor-native workflow outputs to CPU before stream…
alexnorell Jul 16, 2026
4282e52
fix(media): ignore non-video RTSP tracks in uridecodebin producers
PawelPeczek-Roboflow Jul 16, 2026
ff63c51
Revert "fix(media): ignore non-video RTSP tracks in uridecodebin prod…
PawelPeczek-Roboflow Jul 16, 2026
f634334
fix(media): explicit video-only RTSP pipeline on Jetson (jetson-utils…
PawelPeczek-Roboflow Jul 16, 2026
c926a51
perf(stream): stop materialising tensor frames at sink dispatch
PawelPeczek-Roboflow Jul 16, 2026
2e6d36a
fix(media): report bus errors before EOS in the Jetson bridge grab
PawelPeczek-Roboflow Jul 16, 2026
a20f56d
perf(media): jetson-utils consume model for the Jetson tensor bridge …
PawelPeczek-Roboflow Jul 16, 2026
6f7d285
perf(media): per-phase timing counters in the Jetson tensor bridge (v5)
PawelPeczek-Roboflow Jul 28, 2026
b6c7a5c
feat(stream): video processing modes with self-tuning auto collection
PawelPeczek-Roboflow Jul 29, 2026
4b349fb
feat(media): lossless file handoff in the Jetson tensor bridge (v6)
PawelPeczek-Roboflow Jul 29, 2026
fcdc355
perf(media): per-fd EGL registration cache in the Jetson tensor bridg…
PawelPeczek-Roboflow Jul 29, 2026
3b9e8f1
feat(stream): explicit 'legacy' opt-out from the flag-driven AUTO def…
PawelPeczek-Roboflow Jul 29, 2026
b2cd98b
fix(media): route local files to the Jetson producer + log constructi…
PawelPeczek-Roboflow Jul 29, 2026
24676da
feat(media): build the Jetson tensor bridge into the JP5 (5.1.1) image
PawelPeczek-Roboflow Jul 29, 2026
181c5e9
fix(media): preload libGLdispatch in the JP5 image (static TLS exhaus…
PawelPeczek-Roboflow Jul 29, 2026
af527f2
feat(stream): rate-matched adaptive collection window in AUTO mode
PawelPeczek-Roboflow Jul 29, 2026
249509a
perf(workflows): empty-detections tensor passthrough in bbox visualiz…
PawelPeczek-Roboflow Jul 30, 2026
b791dc5
Go for new inference-models package release
PawelPeczek-Roboflow Jul 30, 2026
291500f
Bump requirements
PawelPeczek-Roboflow Jul 30, 2026
ffa89fe
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Jul 30, 2026
f59da19
fix(workflows): tensor-sibling parity for the main merge (ffa89fec7)
PawelPeczek-Roboflow Jul 30, 2026
3c43507
fix(workflows): re-port depth_estimation v1_tensor to post-merge cont…
PawelPeczek-Roboflow Jul 30, 2026
023dc26
style: black collection_policy.py + base_tensor.py (CI gate)
PawelPeczek-Roboflow Jul 30, 2026
af863d9
feat(workflows): frame_delay tensor sibling — spill image payloads to…
PawelPeczek-Roboflow Jul 30, 2026
643cd50
style: isort loader.py (CI gate — ordering broke in the main merge)
PawelPeczek-Roboflow Jul 30, 2026
9ca0675
test(workflows): tensor-native parity for nearest-neighbor integratio…
PawelPeczek-Roboflow Jul 30, 2026
abe6d81
Make linters happy
PawelPeczek-Roboflow Jul 30, 2026
0629a7b
Resolve conflicts with main
PawelPeczek-Roboflow Aug 3, 2026
5a25956
Make lintesr happy and bump inference-models
PawelPeczek-Roboflow Aug 3, 2026
1342138
Add env flag to disable GSTREAMER sources in old infenece
PawelPeczek-Roboflow Aug 3, 2026
564a0bc
Align function names after merge
PawelPeczek-Roboflow Aug 3, 2026
918d93f
Add parity for depended resources to tensor-native blocks
PawelPeczek-Roboflow Aug 3, 2026
3915dd8
Extend timeout for tests
PawelPeczek-Roboflow Aug 3, 2026
b2a2f21
Extend timeout for tests
PawelPeczek-Roboflow Aug 3, 2026
a3b3ec8
Fix tests
PawelPeczek-Roboflow Aug 3, 2026
a02ca01
Merge branch 'main' into mvp/new-inference-pipeline
PawelPeczek-Roboflow Aug 3, 2026
1db5031
Bring back timeout
PawelPeczek-Roboflow Aug 3, 2026
4be8e86
feat(workflows): GPU-native mask visualization compositor (tensor pip…
Erol444 Aug 4, 2026
5d501a4
Add mask visualisation pass-through
PawelPeczek-Roboflow Aug 4, 2026
6e0d584
Speed-up label visualisation block
PawelPeczek-Roboflow Aug 4, 2026
7486b97
Add env variable to enforce dense masks
PawelPeczek-Roboflow Aug 4, 2026
325c10a
true-tensorification of mask visualisation block
PawelPeczek-Roboflow Aug 4, 2026
c0c47e9
true-tensorification of mask visualisation block
PawelPeczek-Roboflow Aug 4, 2026
40e0d59
Attempt to remove default-cuda-stream congestion
PawelPeczek-Roboflow Aug 4, 2026
c9a115e
Try to fix label vis
PawelPeczek-Roboflow Aug 4, 2026
c86b4db
Applin quick fixes for detecteions stitch and image slicer performance
PawelPeczek-Roboflow Aug 4, 2026
3160d4f
Route RTSP producer options per video source
alexnorell Aug 5, 2026
2910baf
fix(jp72): support tensor-native SDES SRTP
alexnorell Jul 20, 2026
3dce375
fix(media): move Jetson conversion off decoder callbacks
alexnorell Aug 5, 2026
f62dfe4
Make Jetson TensorRT download resumable
alexnorell Aug 5, 2026
0f68bc4
Bound Jetson dependency download time
alexnorell Aug 5, 2026
71a36a5
Vendor TensorRT package into Jetson build context
alexnorell Aug 5, 2026
bfd9133
Use checked-out context for Jetson builds
alexnorell Aug 5, 2026
98ac580
Repair Jetson RTSP startup
alexnorell Aug 5, 2026
e505718
Expose source options and restore local JP62 builds
alexnorell Aug 5, 2026
aed76b2
Test aligned video source options
alexnorell Aug 5, 2026
d9c65cf
Preserve positional pipeline compatibility
alexnorell Aug 5, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
The table of contents is too big for display.
Diff view
Diff view
  •  
  •  
  •  
3 changes: 3 additions & 0 deletions .github/workflows/docker.jetson.6.2.0.yml
Original file line number Diff line number Diff line change
Expand Up @@ -52,11 +52,14 @@ jobs:
base_image: ${{ env.BASE_IMAGE }}
force_push: ${{ github.event.inputs.force_push }}
token: ${{ secrets.GITHUB_TOKEN }}
- name: Fetch TensorRT local repository
run: docker/scripts/fetch_jetson_6_2_tensorrt.sh
- name: Set up Depot CLI
uses: depot/setup-action@v1
- name: Build and Push
uses: depot/build-push-action@v1
with:
context: .
push: ${{ github.event_name == 'release' || (github.event.inputs.force_push == 'true')}}
project: 2rp7mfjw7q
tags: ${{ steps.tags.outputs.image_tags }}
Expand Down
61 changes: 0 additions & 61 deletions .github/workflows/docker.jetson.7.1.0.yml

This file was deleted.

2 changes: 1 addition & 1 deletion .github/workflows/docker.jetson.7.2.0.yml
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@ jobs:
runs-on:
labels: depot-ubuntu-24.04-4
group: public-depot
timeout-minutes: 360
timeout-minutes: 1440
permissions:
id-token: write
contents: read
Expand Down
74 changes: 74 additions & 0 deletions .github/workflows/docker.media.jetson.7.2.0.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,74 @@
name: Build and Push Jetson 7.2.0 Media Container

permissions:
contents: read

on:
release:
types: [created]
push:
branches: [main]
paths:
- ".github/workflows/docker.media.jetson.7.2.0.yml"
- "docker/dockerfiles/Dockerfile.media.jetson.7.2.0"
- "docker/native/jetson_tensor_bridge/**"
- "docker/scripts/**"
- "LICENSE"
- "LICENSE.core"
workflow_dispatch:
inputs:
force_push:
type: boolean
description: "Do you want to push the image after building?"
default: false
custom_tag:
type: string
description: "Custom tag to use for the image"
default: ""

env:
VERSION: "0.0.0"
BASE_IMAGE: "roboflow/roboflow-inference-media-jetson-7.2.0"

jobs:
docker:
runs-on:
labels: depot-ubuntu-24.04-4
group: public-depot
timeout-minutes: 180
permissions:
id-token: write
contents: read
steps:
- name: Login to Docker Hub
if: ${{ github.event_name == 'release' || inputs.force_push }}
uses: docker/login-action@v4
with:
username: ${{ secrets.DOCKERHUB_USERNAME }}
password: ${{ secrets.DOCKERHUB_TOKEN }}
- name: Checkout
uses: actions/checkout@v6
- name: Read version from file
run: echo "VERSION=$(DISABLE_VERSION_CHECK=true python ./inference/core/version.py)" >> "$GITHUB_ENV"
- name: Determine image tags
id: tags
uses: ./.github/actions/determine-tags
with:
custom_tag: ${{ inputs.custom_tag }}
version: ${{ env.VERSION }}
base_image: ${{ env.BASE_IMAGE }}
force_push: ${{ inputs.force_push }}
token: ${{ secrets.GITHUB_TOKEN }}
- name: Set up Depot CLI
uses: depot/setup-action@v1
- name: Build image
uses: depot/build-push-action@v1
with:
project: v1xzfwkc4b
context: .
file: ./docker/dockerfiles/Dockerfile.media.jetson.7.2.0
platforms: linux/arm64
push: ${{ github.event_name == 'release' || inputs.force_push }}
save: ${{ github.event_name != 'release' && !inputs.force_push }}
save-tag: ${{ github.event_name != 'release' && !inputs.force_push && format('media-jp72-{0}', github.sha) || '' }}
tags: ${{ steps.tags.outputs.image_tags }}
108 changes: 108 additions & 0 deletions .github/workflows/docker.wheels.jetson.7.2.0.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,108 @@
name: Build and Push Jetson 7.2.0 Wheels Container

permissions:
contents: read

on:
release:
types: [created]
push:
branches:
- main
# Bootstrap trigger while the MVP branch is in flight; remove after merge.
- mvp/new-inference-pipeline
paths:
- ".github/workflows/docker.wheels.jetson.7.2.0.yml"
- "docker/dockerfiles/Dockerfile.wheels.jetson.7.2.0"
- "docker/docker-bake.jetson.7.2.0.hcl"
- "docker/patches/**"
workflow_dispatch:
inputs:
force_push:
type: boolean
description: "Do you want to push the image after building?"
default: false
custom_tag:
type: string
description: "Custom tag to use for the image"
default: ""

env:
VERSION: "0.0.0"
BASE_IMAGE: "roboflow/roboflow-inference-wheels-jetson-7.2.0"

jobs:
# The expensive compiles fan out to separate Depot builders; each populates the
# shared project cache, and the assemble job below cache-hits all of them.
components:
strategy:
fail-fast: false
matrix:
include:
- name: torch-stack
targets: jetson-wheels-torch-stack-jp72
- name: ort-stack
targets: jetson-wheels-ort-stack-jp72
name: Build ${{ matrix.name }}
runs-on:
labels: depot-ubuntu-24.04-4
group: public-depot
timeout-minutes: 1440
permissions:
id-token: write
contents: read
steps:
- name: Checkout
uses: actions/checkout@v6
- name: Set up Depot CLI
uses: depot/setup-action@v1
- name: Build component
uses: depot/bake-action@v1
with:
files: ./docker/docker-bake.jetson.7.2.0.hcl
targets: ${{ matrix.targets }}
project: v1xzfwkc4b

assemble:
name: Assemble wheels image
needs: components
runs-on:
labels: depot-ubuntu-24.04-4
group: public-depot
timeout-minutes: 180
permissions:
id-token: write
contents: read
steps:
- name: Login to Docker Hub
if: ${{ github.event_name == 'release' || inputs.force_push }}
uses: docker/login-action@v4
with:
username: ${{ secrets.DOCKERHUB_USERNAME }}
password: ${{ secrets.DOCKERHUB_TOKEN }}
- name: Checkout
uses: actions/checkout@v6
- name: Read version from file
run: echo "VERSION=$(DISABLE_VERSION_CHECK=true python ./inference/core/version.py)" >> "$GITHUB_ENV"
- name: Determine image tags
id: tags
uses: ./.github/actions/determine-tags
with:
custom_tag: ${{ inputs.custom_tag }}
version: ${{ env.VERSION }}
base_image: ${{ env.BASE_IMAGE }}
force_push: ${{ inputs.force_push }}
token: ${{ secrets.GITHUB_TOKEN }}
- name: Set up Depot CLI
uses: depot/setup-action@v1
- name: Build image
uses: depot/bake-action@v1
env:
JETSON_WHEELS_TAGS: ${{ steps.tags.outputs.image_tags }}
with:
files: ./docker/docker-bake.jetson.7.2.0.hcl
targets: jetson-wheels-jp72
project: v1xzfwkc4b
push: ${{ github.event_name == 'release' || inputs.force_push }}
save: ${{ github.event_name != 'release' && !inputs.force_push }}
save-tag: ${{ github.event_name != 'release' && !inputs.force_push && format('wheels-jp72-{0}', github.sha) || '' }}
Original file line number Diff line number Diff line change
@@ -0,0 +1,64 @@
name: INTEGRATION TESTS - workflows (tensor-native)
permissions:
contents: read
on:
pull_request:
branches: [main]
push:
branches: [main]
workflow_dispatch:
inputs:
python-version:
description: "Python version to run (or all)"
type: choice
default: all
options:
- all
- "3.10"
- "3.11"
- "3.12"

concurrency:
group: ${{ github.workflow }}-${{ github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}

jobs:
call_is_mergeable:
uses: ./.github/workflows/check_if_branch_is_mergeable.yml
secrets: inherit
build-dev-test:
needs: call_is_mergeable
if: ${{ (github.event.pull_request.head.repo.fork == false) && (github.event_name != 'pull_request' || needs.call_is_mergeable.outputs.mergeable_state != 'not_clean') }}
runs-on:
labels: depot-ubuntu-22.04-16
group: public-depot
strategy:
matrix:
python-version: ${{ fromJSON((github.event.inputs.python-version || 'all') == 'all' && '["3.10","3.11","3.12"]' || format('["{0}"]', github.event.inputs.python-version)) }}
timeout-minutes: 15
steps:
- name: 🛎️ Checkout
uses: actions/checkout@v6
- name: 🐍 Set up Python ${{ matrix.python-version }}
uses: actions/setup-python@v6
with:
python-version: ${{ matrix.python-version }}
check-latest: true
- name: 📦 Cache Python packages
uses: actions/cache@v5
with:
path: ~/.cache/pip
key: ${{ runner.os }}-pip-${{ matrix.python-version }}-${{ hashFiles('requirements/**') }}
restore-keys: |
${{ runner.os }}-pip-${{ matrix.python-version }}-
- name: 🚧 Install GDAL
run: sudo apt-get update && sudo apt-get install -y libgdal-dev
- name: 📦 Install dependencies
run: |
python -m pip install --upgrade pip
pip install --upgrade setuptools
pip install --extra-index-url https://download.pytorch.org/whl/cpu -r requirements/_requirements.txt -r requirements/requirements.sam.txt -r requirements/requirements.cpu.txt -r requirements/requirements.http.txt -r requirements/requirements.test.unit.txt -r requirements/requirements.doctr.txt -r requirements/requirements.yolo_world.txt -r requirements/requirements.transformers.txt -r requirements/requirements.sdk.http.txt -r requirements/requirements.easyocr.txt
- name: 🧪 Integration Tests of Workflows (tensor-native)
# Tensor-native path only: ENABLE_TENSOR_DATA_REPRESENTATION is AND-gated with
# USE_INFERENCE_MODELS, which defaults to True on Linux runners, so it is not set here.
run: ENABLE_TENSOR_DATA_REPRESENTATION=True ROBOFLOW_API_KEY=${{ secrets.API_KEY }} DEEP_LAB_V3_API_KEY=${{ secrets.DEEP_LAB_V3_API_KEY }} MODAL_TOKEN_ID=${{ secrets.MODAL_TOKEN_ID }} MODAL_TOKEN_SECRET=${{ secrets.MODAL_TOKEN_SECRET }} MODAL_ALLOW_ANONYMOUS_EXECUTION=True MODAL_ANONYMOUS_WORKSPACE_NAME=github-test MODAL_SKIP_FLORENCE2_TEST=FALSE LOAD_ENTERPRISE_BLOCKS=TRUE python -m pytest tests/workflows/integration_tests
30 changes: 29 additions & 1 deletion .github/workflows/test.nvidia_t4.yml
Original file line number Diff line number Diff line change
Expand Up @@ -56,7 +56,35 @@ jobs:
- name: �🔨 Build and Push Test Docker - GPU
run: |
docker build -t roboflow/roboflow-inference-server-gpu:test -f docker/dockerfiles/Dockerfile.onnx.gpu .


- name: Verify CUDA OpenCV - GPU
if: ${{ github.event.inputs.test_name == '' }}
run: |
docker run --rm --gpus all --entrypoint sh \
-e OPENCV_REQUIRE_CUDA_BUILD=true \
-e OPENCV_REQUIRE_CUDA_RUNTIME=true \
-v "$PWD/docker/scripts/verify_opencv.sh:/tmp/verify_opencv:ro" \
roboflow/roboflow-inference-server-gpu:test \
/tmp/verify_opencv

- name: Verify NVIDIA GStreamer codecs - GPU
if: ${{ github.event.inputs.test_name == '' }}
run: |
docker run --rm --gpus all --entrypoint sh \
-e GSTREAMER_REQUIRE_NVCODEC=true \
-e GSTREAMER_REQUIRE_NVCODEC_RUNTIME=true \
-v "$PWD/docker/scripts/verify_gstreamer.sh:/tmp/verify_gstreamer:ro" \
roboflow/roboflow-inference-server-gpu:test \
/tmp/verify_gstreamer

- name: Verify CUDA GStreamer tensor bridge - GPU
if: ${{ github.event.inputs.test_name == '' }}
run: |
docker run --rm --gpus all --entrypoint python3 \
-v "$PWD/docker/scripts/verify_gstreamer_cuda_tensor_runtime.py:/tmp/verify_gstreamer_cuda_tensor_runtime.py:ro" \
roboflow/roboflow-inference-server-gpu:test \
/tmp/verify_gstreamer_cuda_tensor_runtime.py

- name: 🔋 Start Test Docker without Torch Preprocessing - GPU
if: ${{ github.event.inputs.test_name == '' || github.event.inputs.test_name == 'regression_without_torch' }}
run: |
Expand Down
Loading