Bump version to v1.2.0

README.md

@@ -86,6 +86,11 @@ https://github.com/open-mmlab/mmpretrain/assets/26739999/e4dcd3a2-f895-4d1b-a351
 
 ## What's new
 
+🌟 v1.2.0 was released in 04/01/2023
+
+- Support LLaVA 1.5.
+- Implement of RAM with a gradio interface.
+
 🌟 v1.1.0 was released in 12/10/2023
 
 - Support Mini-GPT4 training and provide a Chinese model (based on Baichuan-7B)

README_zh-CN.md

@@ -84,6 +84,11 @@ https://github.com/open-mmlab/mmpretrain/assets/26739999/e4dcd3a2-f895-4d1b-a351
 
 ## 更新日志
 
+🌟 2024/01/04 发布了 v1.2.0 版本
+
+- 支持了 LLaVA 1.5
+- 实现了一个 RAM 模型的 gradio 推理例程
+
 🌟 2023/10/12 发布了 v1.1.0 版本
 
 - 支持 Mini-GPT4 训练并提供一个基于 Baichuan-7B 的中文模型

docker/serve/Dockerfile

@@ -3,7 +3,7 @@ ARG CUDA="11.7"
 ARG CUDNN="8"
 FROM pytorch/torchserve:latest-gpu
 
-ARG MMPRE="1.1.1"
+ARG MMPRE="1.2.0"
 
 ENV PYTHONUNBUFFERED TRUE
 

docs/en/notes/changelog.md

@@ -1,5 +1,16 @@
 # Changelog (MMPreTrain)
 
+## v1.2.0(04/01/2024)
+
+### New Features
+
+- [Feature] Support LLaVA 1.5 ([#1853](https://github.com/open-mmlab/mmpretrain/pull/1853))
+- [Feature] Implement of RAM with a gradio interface. ([#1802](https://github.com/open-mmlab/mmpretrain/pull/1802))
+
+### Bug Fix
+
+- [Fix] Fix resize mix argument bug.
+
 ## v1.1.0(12/10/2023)
 
 ### New Features

docs/en/notes/faq.md

@@ -16,7 +16,8 @@ and make sure you fill in all required information in the template.
 
   | MMPretrain version | MMEngine version  |   MMCV version   |
   | :----------------: | :---------------: | :--------------: |
-  |    1.1.1 (main)    | mmengine >= 0.8.3 |  mmcv >= 2.0.0   |
+  |    1.2.0 (main)    | mmengine >= 0.8.3 |  mmcv >= 2.0.0   |
+  |       1.1.1        | mmengine >= 0.8.3 |  mmcv >= 2.0.0   |
   |       1.0.0        | mmengine >= 0.8.0 |  mmcv >= 2.0.0   |
   |      1.0.0rc8      | mmengine >= 0.7.1 | mmcv >= 2.0.0rc4 |
   |      1.0.0rc7      | mmengine >= 0.5.0 | mmcv >= 2.0.0rc4 |

docs/zh_CN/notes/faq.md

@@ -13,7 +13,8 @@
 
   | MMPretrain 版本 |   MMEngine 版本   |    MMCV 版本     |
   | :-------------: | :---------------: | :--------------: |
-  |  1.1.1 (main)   | mmengine >= 0.8.3 |  mmcv >= 2.0.0   |
+  |  1.2.0 (main)   | mmengine >= 0.8.3 |  mmcv >= 2.0.0   |
+  |      1.1.1      | mmengine >= 0.8.3 |  mmcv >= 2.0.0   |
   |      1.0.0      | mmengine >= 0.8.0 |  mmcv >= 2.0.0   |
   |    1.0.0rc8     | mmengine >= 0.7.1 | mmcv >= 2.0.0rc4 |
   |    1.0.0rc7     | mmengine >= 0.5.0 | mmcv >= 2.0.0rc4 |

mmpretrain/__init__.py

@@ -7,7 +7,7 @@
 from .version import __version__
 
 mmcv_minimum_version = '2.0.0'
-mmcv_maximum_version = '2.2.0'
+mmcv_maximum_version = '2.4.0'
 mmcv_version = digit_version(mmcv.__version__)
 
 mmengine_minimum_version = '0.8.3'

mmpretrain/models/utils/batch_augments/resizemix.py

@@ -87,7 +87,7 @@ def mix(self, batch_inputs: torch.Tensor,
         (y1, y2, x1, x2), lam = self.cutmix_bbox_and_lam(img_shape, lam)
         batch_inputs[:, :, y1:y2, x1:x2] = F.interpolate(
             batch_inputs[index],
-            size=(y2 - y1, x2 - x1),
+            size=(int(y2 - y1), int(x2 - x1)),
             mode=self.interpolation,
             align_corners=False)
         mixed_scores = lam * batch_scores + (1 - lam) * batch_scores[index, :]

mmpretrain/version.py

@@ -1,6 +1,6 @@
 # Copyright (c) OpenMMLab. All rights reserved
 
-__version__ = '1.1.1'
+__version__ = '1.2.0'
 
 
 def parse_version_info(version_str):

requirements/mminstall.txt

@@ -1,2 +1,2 @@
-mmcv>=2.0.0,<2.3.0
+mmcv>=2.0.0,<2.4.0
 mmengine>=0.8.3,<1.0.0

requirements/optional.txt

@@ -1,4 +1,4 @@
 albumentations>=0.3.2 --no-binary qudida,albumentations    # For Albumentations data transform
-grad-cam >= 1.3.7   # For CAM visualization
+grad-cam >= 1.3.7,<1.5.0   # For CAM visualization
 requests            # For torchserve
 scikit-learn        # For t-SNE visualization and unit tests.

tests/test_models/test_backbones/test_repmlp.py

@@ -169,4 +169,5 @@ def test_deploy_(self):
 
         assert len(feats_) == len(feats__)
         for i in range(len(feats)):
-            self.assertTrue(torch.allclose(feats__[i], feats_[i]))
+            self.assertTrue(
+                torch.allclose(feats__[i], feats_[i], rtol=0.01, atol=0.01))