Peng Lu
|
00030e3f7d
|
[Fix] refactor _prepare_pos_embed in ViT to fix bug in loading old checkpoint (#1679)
|
2023-07-03 11:36:44 +08:00 |
Wangbo Zhao(黑色枷锁)
|
68758db7a8
|
[Fix] freeze pre norm in vision transformer. (#1672)
|
2023-06-28 17:00:27 +08:00 |
Yixiao Fang
|
70ff2abbf7
|
[Refactor] Refactor _prepare_pos_embed in ViT (#1656)
* deal with cls_token
* Update implement
---------
Co-authored-by: mzr1996 <mzr1996@163.com>
|
2023-06-20 17:37:08 +08:00 |
|
Yixiao Fang
|
d9e561a09d
|
[Feature] Support dinov2 backbone (#1522)
* support dinov2 backbone
* update metafile and readme
* compatible to use_layer_scale
* update SwiGLUFFN
* add deprecation warning
* update
|
2023-05-05 16:59:37 +08:00 |
|
Yixiao Fang
|
75dceaa78f
|
[Refactor] Add ln to vit avg_featmap output (#1447)
|
2023-04-06 11:59:39 +08:00 |
Ma Zerun
|
dbf3df21a3
|
[Refactor] Use `out_type` to specify ViT-like backbone output. (#1408)
* [Refactor] Use to specify ViT-like backbone output.
* Fix ClsBatchNormNeck
* Update mmpretrain/models/necks/mae_neck.py
---------
Co-authored-by: Yixiao Fang <36138628+fangyixiao18@users.noreply.github.com>
|
2023-03-09 11:02:58 +08:00 |
|
Ma Zerun
|
274a67223e
|
[Feature] Implement layer-wise learning rate decay optimizer constructor. (#1399)
* [Feature] Implement layer-wise learning rate decay optimizer constructor.
* Use num_layers instead of max_depth to avoid misleading
* Add UT
* Update docstring
* Update log info
* update LearningRateDecay configs
---------
Co-authored-by: fangyixiao18 <fangyx18@hotmail.com>
|
2023-03-07 17:30:39 +08:00 |
|
mzr1996
|
0979e78573
|
Rename the package name to `mmpretrain`.
|
2023-02-17 15:20:55 +08:00 |