f07ed6bcc0
[Fix] Fix deform_conv ops on Ascend NPU ( #2805 )
2023-06-15 16:46:07 +08:00
9c0f9cb98c
Revert "[Fix] Fix arf op's write conflict when num_orientations is not 1 ( #2822 )" ( #2826 )
2023-06-10 23:12:28 +08:00
e43982597c
[Fix] Fix arf op's write conflict when num_orientations is not 1 ( #2822 )
2023-06-09 10:32:39 +08:00
16d52da1e0
[Fix] keep the shape of iou op's parameter 2 smaller than parameter 1 ( #2820 )
2023-06-09 10:30:12 +08:00
e5dcfba13e
[Refactor] Simplify the logic of sparse_conv ( #2683 )
2023-06-01 14:16:51 +08:00
58e7c8ce7a
[Refactor] Repalce the implementation of roi_align_rotated with mlu-ops ( #2657 )
2023-06-01 00:59:26 +08:00
fd3fbfe197
[Refactor] Repalce the implementation of rotated_feature_align with mlu_ops ( #2659 )
2023-06-01 00:58:29 +08:00
515d5416a0
[Refactor] Replace the implementation of psa_mask with mlu-ops. ( #2756 )
2023-06-01 00:57:30 +08:00
8ca930cc72
[Refactor] Replace carafe op of MLU backend with mlu-ops ( #2818 )
2023-06-01 00:48:48 +08:00
445910eff6
DvcliveLoggerHook: use `by_epoch` for determining when to log ( #2149 )
2023-05-22 15:10:11 +08:00
8725e683b4
[Refactor] Replace the MLU ops implementation with mlu-ops ( #2750 )
2023-05-18 11:23:41 +08:00
e197effeca
[Feature] Add the support of arf op for ascend device ( #2789 )
2023-05-11 15:47:08 +08:00
883d339cfb
Replace the ubuntu-18.04 with 22.04 in CI ( #2793 )
2023-05-11 14:03:24 +08:00
595f8fdff3
[Feature] Add the MLU support for box_iou_rotated op ( #2703 )
2023-04-20 19:31:08 +08:00
8d01bf6a6c
[Fix] Fix torch_npu may not contain optim ( #2766 )
...
* fix npu optimizer
* fix npu optimizer
2023-04-19 10:17:18 +08:00
81163d5cb1
[Feature] Add torch_npu optimizers ( #2753 )
...
* add npu optimizer
* add npu optimizer
* add npu optimizer
2023-04-13 23:08:17 +08:00
bfdd1f9c42
[Fix] Fix boxes.scalar_type in nms_npu ( #2731 )
2023-04-04 16:05:42 +08:00
bc727f7132
[Enhancement] Repalce the implementation of roiaware_pool3d with mlu-ops. ( #2699 )
...
* [Feature] Repalce the implementation of roiaware_pool3d with mlu-ops.
* [Feature] Repalce the implementation of roiaware_pool3d with mlu-ops.
2023-04-03 23:35:07 +08:00
a55f4b7f40
[Enhancement] Replace the implementation of three_nn_forward with mlu-ops ( #2719 )
2023-04-03 23:30:40 +08:00
f946a933bb
[Enhancement] Add the dtype limit of nms_npu to maintain consistency with the GPU ( #2724 )
...
* Increase the dtype limit to maintain consistency with the gpu.
* update error message
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
---------
Co-authored-by: momo609 <963372609.qq.com>
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
2023-04-03 10:02:39 +08:00
8ceac9347e
Remove .owners.yml ( #2712 )
2023-03-28 17:06:37 +08:00
66c5515174
[Enhance] Add a default value for MMCV_MLU_ARGS ( #2688 )
...
* add mlu_args
* add mlu_args
* Modify the code
---------
Co-authored-by: budefei <budefei@cambricon.com>
2023-03-24 19:59:43 +08:00
4c4ba6cb17
[Fix] Force bbox_overlaps calculation with FP32 for ascend device ( #2697 )
...
* modify bbox_overlaps op adapter
* update
2023-03-24 19:08:57 +08:00
5f1d55885e
[Enhance] Ignore mlu-ops files ( #2691 )
...
Co-authored-by: budefei <budefei@cambricon.com>
2023-03-24 17:00:16 +08:00
9fcf48c4c8
[Enhancement] Replace the implementation of deform_roi_pool with mlu-ops ( #2598 )
...
* [Feature] Replace the implementation of deform_roi_pool with mlu-ops
* [Feature] Modify code
---------
Co-authored-by: budefei <budefei@cambricon.com>
2023-03-24 10:27:45 +08:00
0d1b224fb1
[Feature] Support NmsRotated with cambricon MLU backend ( #2643 )
...
* [Feature] Support NmsRotated with cambricon MLU backend
* [Feature] remove foolproofs in nms_rotated_mlu.cpp
* [Feature] fix lint in test_nms_rotated.py
* [Feature] fix kMLU not found in nms_rotated.cpp
* [Feature] modify mlu support in nms.py
* [Feature] modify nms_rotated support in ops.md
* [Feature] modify ops/nms.py
2023-03-23 11:36:32 +08:00
e6f434c440
[Enhance] Add a comment for the use of masked_conv on Ascend device ( #2689 )
...
* Update masked_conv.py
* Update mmcv/ops/masked_conv.py
---------
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
2023-03-22 20:37:15 +08:00
1de3aeffd7
Fix bbox_overlaps of parrots ( #2684 )
...
* Fix parrots bbox_overlaps
* Fix
2023-03-22 11:39:55 +08:00
4e8972fbf9
Move voxel_size to the device where points are ( #2681 )
...
* ...
* ...
* ...
2023-03-20 22:08:30 +08:00
06fa32853b
[Feature] Add MLU support for Sparse Convolution op ( #2589 )
...
* [Feature] Add sparse convolution MLU API
* [Feature] update cpp code style
* end-of-file
* delete libext.a
* code style
* update ops.md
---------
Co-authored-by: budefei <budefei@cambricon.com>
2023-03-20 19:28:13 +08:00
01a0f53ea4
[Feature] Add Ascend support for RoIAlign op ( #2638 )
...
* add roi_algin op
* add unit test
* update ops.md
* fix typo
---------
Co-authored-by: momo609 <963372609.qq.com>
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
2023-03-20 19:10:42 +08:00
a3e97360a9
[Fix] nms_rotated_npu accepts angle parameter ( #2670 )
2023-03-17 16:40:01 +08:00
75290685b7
[Enhancement] Refine the implementation of NMS op in Ascend device ( #2658 )
2023-03-15 23:09:33 +08:00
9a671e48fb
[Fix] Prevent divide-by-zero error on Ascend device for bbox_overlaps ( #2644 )
2023-03-15 23:07:41 +08:00
126077df5e
[Feature] Make voxelization operator support the mlu290 platform ( #2652 )
...
* [Feature]: Make voxelization operator support the mlu290 platform.
* [Feature]: Make voxelization operator support the mlu290 platform.
* [Feature]: Make voxelization operator support the mlu290 platform.
---------
Co-authored-by: maxiangjun <maxiangjun@cambricon.com>
2023-03-10 11:21:43 +08:00
745985937a
[Feature] Add the support of voxelization op for ascend device ( #2613 )
...
* update
* update
2023-03-03 20:07:36 +08:00
5f122293bf
[Enhancement] ms_deform_attn performance optimization ( #2616 )
...
* ms_opt_v2
* ms_opt_v2_1
* optimize MultiScaleDeformableAttention ops for MLU
* ms_opt_v2_1
* [Feature] ms_deform_attn performance optimization V2
* [Feature] ms_deform_attn performance optimization V2
* [Feature] ms_deform_attn performance optimization V2
* [Feature] ms_deform_attn performance optimization V2
* [Feature] ms_deform_attn performance optimization V2
* [Feature] ms_deform_attn performance optimization V2
* [Feature] ms_deform_attn performance optimization V2
---------
Co-authored-by: dongchengwei <dongchengwei@cambricon.com>
2023-03-03 11:18:19 +08:00
ec63932333
[Enhance] The argument scale of fused_bias_leakyrelu should be greater than 0 ( #2623 )
...
* Add verification for scale, when scale less than 0
* Update mmcv/ops/fused_bias_leakyrelu.py
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
* Update fused_bias_leakyrelu.py
* Update fused_bias_leakyrelu.py
---------
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
2023-02-27 19:03:46 +08:00
4adddfddf0
Fix the paper link error in the docstring ( #2618 )
2023-02-27 14:53:24 +08:00
8e8ab22686
[Enhancement] Support MultiScaleDeformableAttention with AMP ( #2541 )
...
* [Enhance] Support FP16 for MSDeformAttn
* [Fix] Data type mismatch
* Update mmcv/ops/multi_scale_deform_attn.py
* Add UT
Author: nijkah <nijkah@gmail.com>
* Add cuda available condition
---------
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
2023-02-17 19:27:14 +08:00
cb94ffb672
[Fix] Fix a typo in the docstring of MSDeformAttn ( #2599 )
2023-02-13 16:59:25 +08:00
5a61e534ea
[Feature] Add Ascend support for bbox_overlaps ( #2580 )
...
* Add support for Ascend devices with bbox_overlaps
* ...
* ...
2023-02-06 18:33:16 +08:00
615d28176a
[Feature] Add Ascend support for RoIPoolGrad op ( #2569 )
...
* add roipoolgrad adapter
* amend
2023-02-06 18:29:59 +08:00
04e8727adf
Remove unused files ( #2586 )
2023-02-06 12:50:07 +08:00
18e8f32fff
Refactor the issue templates ( #2585 )
...
* Refactor the issue template
* fix links
* fix
2023-02-06 11:51:23 +08:00
46ea829b67
[Fix] Fix tensor descriptor setting in MLU ball_query. ( #2579 )
2023-02-02 21:44:42 +08:00
497c2df495
[Fix] Add missing contiguous calls for nms_rotated ( #2547 )
...
Addresses issue https://github.com/open-mmlab/mmcv/issues/2542
For nms_rotated_cuda only adds contiguous call to dets_sorted which is accessed directly.
For nms_rotated_cpu the implementation now matches what's in the detectron2 repo.
2023-02-01 14:44:56 +08:00
87768b7f2d
[Docs] Steps to compile mmcv-full on MLU machine ( #2571 )
...
* [Docs] Steps to compile mmcv-full on MLU machine
* [Docs] Adjust paragraph order
* Update docs/zh_cn/get_started/build.md
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
* Update docs/zh_cn/get_started/build.md
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
* Update docs/en/get_started/build.md
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
* Update docs/en/get_started/build.md
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
* [Docs] Modify the format
---------
Co-authored-by: budefei <budefei@cambricon.com>
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
2023-01-31 19:40:07 +08:00
98766f4def
[Docs] Refine the docstring of EvalHook ( #2576 )
...
* 📃 docs(hooks): EvalHook comments
Refine the EvalHook comments
* Refine comments EvalHook
Update mmcv/runner/hooks/evaluation.py
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
* 📃 docs(Hooks): update Eval Hook
Slove line too long
---------
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
2023-01-31 17:31:03 +08:00
7e063e3e61
Use official isort ( #2574 )
2023-01-31 12:59:35 +08:00
Yinlei Sun
liuhw
duzekun
tudejiang79
liuduanhui
qipengh
David de la Iglesia Castro
bdf
Zaida Zhou
Zhang
luomaoling
sherie
liuyuan1-v
zcc9707
CokeDong
NoCakeAnyMore
mahxn0
BinZheng
jayggh
BainOuO
Hakjin Lee
xinliang123
Allan Zelener
Daojiang
tripleMu