Skip to content

Commit

Permalink
chore(release): 0.16.0
Browse files Browse the repository at this point in the history
  • Loading branch information
beniz committed Apr 23, 2021
1 parent bee3183 commit 642f173
Show file tree
Hide file tree
Showing 2 changed files with 19 additions and 17 deletions.
34 changes: 18 additions & 16 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
@@ -1,26 +1,28 @@
# DeepDetect: Open Source Deep Learning Server & API (Changelog)

## [0.15.0](https://github.com/jolibrain/deepdetect/compare/v0.14.0...v0.15.0) (2021-03-26)
## [0.16.0](https://github.com/jolibrain/deepdetect/compare/v0.15.0...v0.16.0) (2021-04-23)


### Features

* **nbeats:** default backcast loss coeff to zero, allows very short forecast length to learn smoothly ([db17a41](https://github.com/jolibrain/deepdetect/commit/db17a41401b037187b5ccf2e54464e3f6647e40d))
* **timeseries:** add MAE and MSE metrics ([847830d](https://github.com/jolibrain/deepdetect/commit/847830d8f6a011be05763b36fbf7240dd6d867e6))
* **timeseries:** do not output per serie metrics as a default, add prefix _all for displaying all metrics ([5b6bc4e](https://github.com/jolibrain/deepdetect/commit/5b6bc4e19274595741e8fd11cbfd326b0497b79f))
* **torch:** model publishing with the platform ([da14d33](https://github.com/jolibrain/deepdetect/commit/da14d33affb362aa869367fb748d5dbac1d73a10))
* **torch:** save last model at training service interruption ([b346923](https://github.com/jolibrain/deepdetect/commit/b34692395ee6c0d03b6a378d2b454a1479e52e76))
* **torch:** SWA for RANGER/torch (https://arxiv.org/abs/1803.05407) ([74cf54c](https://github.com/jolibrain/deepdetect/commit/74cf54cce30b791def7712eabd0c93c31eebf91b))
* **torch/csvts:** create db incrementally ([4336e89](https://github.com/jolibrain/deepdetect/commit/4336e893efe3c41d97b0199d300c5461fde55776))
* **torch:** add confidence threshold for classification ([0e75d88](https://github.com/jolibrain/deepdetect/commit/0e75d88fb949fc2e0e23ed744b5752df8b581d5a))
* **torch:** add more backbones to traced detection models ([f4d05e1](https://github.com/jolibrain/deepdetect/commit/f4d05e1ea9f419832cc4c63c5367f7615ef22b2f))
* **torch:** allow FP16 inference on GPU ([705d3d7](https://github.com/jolibrain/deepdetect/commit/705d3d77c8f325d7707cafb422e948b9cc3ac7f7))
* **torch:** madgrad optimizer ([0657d82](https://github.com/jolibrain/deepdetect/commit/0657d82cd05d575cb6d45c2f122946626d7457a8))
* **torch:** training of detection models on backend torch ([b920999](https://github.com/jolibrain/deepdetect/commit/b9209991a4e44a45d9bacaed182fd7ecacaed369))


### Bug Fixes

* **caffe/detection:** fix rare spurious detection decoding, see bug 1190 ([94935b5](https://github.com/jolibrain/deepdetect/commit/94935b5a6c9a4ab9321cca52d5050f3b520e9ff7))
* **chore:** add opencv imgcodecs explicit link ([8ff5851](https://github.com/jolibrain/deepdetect/commit/8ff585140f8784e2a91a955c53d10fcb0917369d))
* compile flags typo ([8f0c947](https://github.com/jolibrain/deepdetect/commit/8f0c947eefad3bde0defae52f4b85317a0e98f50))
* docker cpu link in readme ([1541dcc](https://github.com/jolibrain/deepdetect/commit/1541dccfdbede08ffd0ce466f7f27a171d6647a9))
* tensorrt tests on Jetson nano ([25b12f5](https://github.com/jolibrain/deepdetect/commit/25b12f573d6894a24d233a30ee85092327e0d96f))
* **nbeats:** make seasonality block work ([d035c79](https://github.com/jolibrain/deepdetect/commit/d035c794822f57be5d2aad57403cc2d7ba06738f))
* **torch:** display msg if resume fails, also fails if not best_model.txt file ([d8c5418](https://github.com/jolibrain/deepdetect/commit/d8c541838713ee2922e3d262c24fcf0bf058ce1a))
* **torch:** throw error if multiple models are provided ([efbd1f9](https://github.com/jolibrain/deepdetect/commit/efbd1f93077791299013e04bef292aba9a09afa5))
* **torch:** default gradient clipping to true when using madgrad ([5979019](https://github.com/jolibrain/deepdetect/commit/5979019c27cb5e84ddcb38f40bbd962c32d7003f))
* remove dirty git flag on builds ([6daa4f5](https://github.com/jolibrain/deepdetect/commit/6daa4f5343fb31afbf0efd7330da7513b652e539))
* services names were not always case insentitive ([bee3183](https://github.com/jolibrain/deepdetect/commit/bee318356c2bf056247073c73f580016970f379f))
* **chains:** cloning of image crops in chains ([2e62b7e](https://github.com/jolibrain/deepdetect/commit/2e62b7e6f3f75d2de08e8c6088c5a2da7b320d39))
* **ml:** refinedet image dimensions configuration via API ([20d56e4](https://github.com/jolibrain/deepdetect/commit/20d56e4ac6ab4691c32187137b520996160c8d59))
* **TensorRT:** fix some memory allocation weirdness in trt backend ([4f952c3](https://github.com/jolibrain/deepdetect/commit/4f952c3fbc2f8da03ebc66644e125576d9b12fee))
* **timeseries:** throw if no data found ([a95e7f9](https://github.com/jolibrain/deepdetect/commit/a95e7f936c35cbe5cf24779fe5e899667b7f6e6c))
* **torch:** allow partial or mismatching weights loading only if finetuning ([23666ea](https://github.com/jolibrain/deepdetect/commit/23666ea49ece302477e1f2d8f88edc41366ff213))
* **torch:** Fix underflow in CSVTS::serialize_bounds ([c8b11b6](https://github.com/jolibrain/deepdetect/commit/c8b11b66b4b264ac16b3b2357fbd66293c01f99d))
* **torch:** fix very long ETA with iter_size != 1 ([0c716a6](https://github.com/jolibrain/deepdetect/commit/0c716a60b2742c70ad715ad4e9b23a3f4d035a77))
* **torch:** parameters are added only once to solver during traced model training ([86cbcf5](https://github.com/jolibrain/deepdetect/commit/86cbcf5f41f868a6472bb3df46015db34b61f1a2))
* **torch:** select correct mltype ([e674fcc](https://github.com/jolibrain/deepdetect/commit/e674fcce4812cf499a6e2eaaa8f98b34d35d5ffa))
2 changes: 1 addition & 1 deletion package.json
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
{
"name": "deepdetect",
"description": "Open Source Deep Learning Server & API",
"version": "0.15.0",
"version": "0.16.0",
"private": true,
"devDependencies": {
"standard-version": "^9.0.0"
Expand Down

0 comments on commit 642f173

Please sign in to comment.