Koushik Dutta
|
d2f1c69e98
|
predict: shuffle common files out of tflite
|
2024-06-05 08:30:39 -07:00 |
|
Koushik Dutta
|
2a2f96a771
|
predict: use smaller image for face detect
|
2024-06-04 22:15:43 -07:00 |
|
Koushik Dutta
|
2f19866f05
|
predict: relax face threshold
|
2024-06-03 23:34:01 -07:00 |
|
Koushik Dutta
|
9a770e9dc9
|
predict: update models
|
2024-06-03 15:08:58 -07:00 |
|
Koushik Dutta
|
5eac8d0ab9
|
predict: lock opencv version,
roll back to 9c flt
|
2024-06-03 10:38:36 -07:00 |
|
Koushik Dutta
|
83a3352862
|
predict: extract rough text scores
|
2024-06-02 13:33:54 -07:00 |
|
Koushik Dutta
|
d64f90c0c8
|
predict: republish with smaller plate/face models. fix openvino thread bugs
|
2024-05-31 10:49:08 -07:00 |
|
Koushik Dutta
|
d5ba6f34d6
|
onnx: cleanup
|
2024-05-13 10:00:53 -07:00 |
|
Koushik Dutta
|
820ef70033
|
predict plugins: refactor recog, add onnx, fix spurious model leaks
|
2024-05-12 21:00:48 -07:00 |
|
Koushik Dutta
|
f662bd7de4
|
openvino/coreml: fix text threshold
|
2024-04-23 13:09:55 -07:00 |
|
Koushik Dutta
|
0560d857c1
|
text recognition: improve skew calc
|
2024-04-23 11:56:14 -07:00 |
|
Koushik Dutta
|
4ee72cd074
|
text: fixup skew angle calculation
|
2024-04-23 09:25:44 -07:00 |
|
Koushik Dutta
|
4d98ccf86b
|
predict: move face/text recognition into separate models
|
2024-04-23 08:33:09 -07:00 |
|
Koushik Dutta
|
ff2d1d5f97
|
predict: fix text skews
|
2024-04-22 20:50:52 -07:00 |
|
Koushik Dutta
|
1294fc291a
|
openvino/coreml: wip refactor text recognition
|
2024-04-22 13:07:22 -07:00 |
|
Koushik Dutta
|
a5a1959bd0
|
predict: fix blank detections in dropdown
|
2024-04-19 11:48:44 -07:00 |
|
Koushik Dutta
|
62e23880fd
|
coreml: handle batching hint failures
|
2024-04-19 10:33:37 -07:00 |
|
Koushik Dutta
|
ffc9ca14b5
|
predict: fix double periodic restarts
|
2024-04-19 07:37:40 -07:00 |
|
Koushik Dutta
|
3897e78bdc
|
coreml/sdk: batching support
|
2024-04-17 10:59:54 -07:00 |
|
Koushik Dutta
|
ad9e9f2d1d
|
predict: publish betas
|
2024-04-16 23:24:00 -07:00 |
|
Koushik Dutta
|
7192c5ddc2
|
openvino: fix potential thread safety.
coreml/openvino: more recognition wip
|
2024-04-16 15:48:40 -07:00 |
|
Koushik Dutta
|
c1c9fec62f
|
openvino: working lpr and face recog;
|
2024-04-12 22:12:07 -07:00 |
|
Koushik Dutta
|
27a1c5269a
|
coreml: fixup detection test
|
2024-04-12 22:04:55 -07:00 |
|
Koushik Dutta
|
1dae1834ad
|
predict: missing files
|
2024-04-12 20:45:31 -07:00 |
|
Koushik Dutta
|
10a7877522
|
tensorflow-lite: fix prediction crash
|
2024-04-08 09:29:52 -07:00 |
|
Koushik Dutta
|
95ac72c5c8
|
coreml: encode embedding
|
2024-04-02 22:07:13 -07:00 |
|
Koushik Dutta
|
e37295fb20
|
coreml: move vision framework into coreml
|
2024-04-01 20:45:31 -07:00 |
|
Koushik Dutta
|
2e72366d41
|
vision-framework: initial release
|
2024-04-01 10:20:49 -07:00 |
|
Koushik Dutta
|
c1f1e96109
|
predict: cleanups
|
2023-08-15 21:29:18 -07:00 |
|
Koushik Dutta
|
62c4a8b240
|
detection plugins: remove image splitting logic, let upstream handle that. switch to yolov8_320 as default.
|
2023-07-31 14:12:56 -07:00 |
|
Koushik Dutta
|
4feeeda904
|
openvino: restart if detection times out
|
2023-07-23 18:48:07 -07:00 |
|
Koushik Dutta
|
0c05b59121
|
tensorflow-lite: temp hack to fix wide angle detection
|
2023-07-20 19:53:43 -07:00 |
|
Koushik Dutta
|
1b5565b5b2
|
openvino: choose better defaults for precision
|
2023-06-11 08:48:55 -07:00 |
|
Koushik Dutta
|
9292ebbe48
|
tensorflow-lite: fix missing settings, add python version hints
|
2023-06-08 07:54:41 -07:00 |
|
Koushik Dutta
|
205fdb0222
|
fix per frame rpc gc churn
|
2023-05-12 20:26:18 -07:00 |
|
Koushik Dutta
|
5f6adc9449
|
predict: publish
|
2023-04-27 21:53:50 -07:00 |
|
Koushik Dutta
|
ffee1c5cc2
|
predict: publish
|
2023-04-26 22:51:28 -07:00 |
|
Koushik Dutta
|
0b6ffc2b87
|
predict: strip out allow list
|
2023-04-16 20:50:49 -07:00 |
|
Koushik Dutta
|
0d703c2aff
|
predict: remove filter options
|
2023-04-15 09:33:10 -07:00 |
|
Koushik Dutta
|
645de2e5fd
|
predict: improve input aspect ratio match fast path
|
2023-04-07 10:17:05 -07:00 |
|
Koushik Dutta
|
6a5d6e6617
|
predict: cleanups
|
2023-03-28 10:26:43 -07:00 |
|
Koushik Dutta
|
b55e7cacb3
|
predict: remove old pipline code
|
2023-03-27 11:14:53 -07:00 |
|
Koushik Dutta
|
5472d90368
|
opencv: beta
|
2023-03-26 19:21:22 -07:00 |
|
Koushik Dutta
|
6cd412de88
|
detect: add format hints
|
2023-03-23 19:59:26 -07:00 |
|
Koushik Dutta
|
001918d613
|
predict: fix detections from webui
|
2023-03-16 23:58:45 -07:00 |
|
Koushik Dutta
|
2bce019677
|
predict: make models a separate download
|
2023-03-16 23:29:02 -07:00 |
|
Koushik Dutta
|
aedb985941
|
detect: support motion on new pipeline
|
2023-03-16 20:37:12 -07:00 |
|
Koushik Dutta
|
38ba31ca7d
|
tensorflow-lite: use multiple tpu
|
2023-03-15 23:08:48 -07:00 |
|
Koushik Dutta
|
f6c7b00562
|
tensorflow-lite: fix numpy serialization issue
|
2023-03-14 23:41:55 -07:00 |
|
Koushik Dutta
|
ed341a12b1
|
predict: rgba to rgb conversion
|
2023-03-14 14:50:28 -07:00 |
|