Skip to content

Add 2 3 1 release note #5068

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 16 commits into from
Jul 12, 2022
Merged

Add 2 3 1 release note #5068

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
16 commits
Select commit Hold shift + click to select a range
c5784d0
Update release_note_cn.md
TCChenlong Apr 30, 2022
8f81716
Update release_note_en.md
TCChenlong Apr 30, 2022
552a353
Update release_note_en.md
TCChenlong Apr 30, 2022
9eb73bc
Update release_note_cn.md
TCChenlong Apr 30, 2022
7f66707
Update release_note_cn.md
TCChenlong Apr 30, 2022
87125f9
Update release_note_en.md
TCChenlong Apr 30, 2022
7db26eb
Merge branch 'PaddlePaddle:develop' into develop
TCChenlong Apr 30, 2022
5a6deeb
Merge branch 'PaddlePaddle:develop' into develop
TCChenlong May 11, 2022
1ea84c8
Merge branch 'PaddlePaddle:develop' into develop
TCChenlong May 19, 2022
4e3f6ba
Merge branch 'PaddlePaddle:develop' into develop
TCChenlong Jun 17, 2022
527e647
Merge branch 'PaddlePaddle:develop' into develop
TCChenlong Jun 30, 2022
8e21e09
Merge branch 'develop' of https://github.com/PaddlePaddle/docs into d…
TCChenlong Jul 1, 2022
a949707
Merge branch 'develop' of https://github.com/PaddlePaddle/docs into d…
TCChenlong Jul 7, 2022
93f0925
Merge branch 'develop' of https://github.com/PaddlePaddle/docs into d…
TCChenlong Jul 12, 2022
e1c6b73
add 2.3.1 release note
TCChenlong Jul 12, 2022
33038bc
Update release_note_cn.md
jzhang533 Jul 12, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
116 changes: 116 additions & 0 deletions docs/release_note_cn.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,120 @@

# 2.3.1 Release Note

## 1. 重要更新

- 2.3.1 版本是在 2.3 版本的基础上修复了已知问题，并且发布了支持 CUDA 11.6 的安装包。

## 2. 训练框架（含分布式）

### （1）功能优化

#### API

- 修改`paddle.nn.initializer.KaimingUniform`和`paddle.nn.initializer.KaimingNormal` 两种初始化方式，使其支持多种类型的激活函数。([#43721](https://github.com/PaddlePaddle/Paddle/pull/43721), [#43827](https://github.com/PaddlePaddle/Paddle/pull/43827))
- 优化 `paddle.io.DataLoader` 的数据预读取功能，使其支持设置了 `prefetch_factor` 设定的预读取数据的缓存数量，避免在读取大块数据时出现 IO 阻塞。([#43674](https://github.com/PaddlePaddle/Paddle/pull/43674) )

#### 新动态图执行机制

- 修改新动态图 API 逻辑中 optional 类型 Tensor 的初始化方法，防止被提前析构导致数据异常。([#42561](https://github.com/PaddlePaddle/Paddle/pull/42561))

#### 全新静态图执行器

- 延迟初始化执行器中的线程池，避免只执行一轮的 `program`（如 `save、load、startup_program`等）创建线程池。([#43768](https://github.com/PaddlePaddle/Paddle/pull/43768))

#### 混合精度训练

- 设置 `paddle.nn.Layer` 中 `set_state_dict`中禁用 `state_dict` hook。([#43407](https://github.com/PaddlePaddle/Paddle/pull/43407))

#### 分布式训练

- 使 `paddle.incubate.nn.functional.fused_attention`和 `paddle.incubate.nn.functional.fused_feedforward`支持张量模型并行。([#43505](https://github.com/PaddlePaddle/Paddle/pull/43505))

#### 其他

- 调整框架算子内核打印字符串的格式，便于进行自动化拆分解析。([#42931](https://github.com/PaddlePaddle/Paddle/pull/42931))
- 更新模型量化 API，支持`rounding to nearest ties to even`的四舍五入方式，支持量化取值范围 [-128, 127]。([#43829](https://github.com/PaddlePaddle/Paddle/pull/43829))
- 量化感知训练适配支持 AMP 混合精度训练。([#43689](https://github.com/PaddlePaddle/Paddle/pull/43689))
- 量化感知训练在启动时新增 `progress bar`，便于查看量化初始化进度，统计 out_threshold 时跳过 scale op，加速初始化过程。([#43454](https://github.com/PaddlePaddle/Paddle/pull/43454))
- 动态图量化训练支持 `conv` 和 `bn` 融合，静态图离线量化支持设置 `skip_tensor_list` 来跳过某些层不做量化。([#43301](https://github.com/PaddlePaddle/Paddle/pull/43301))

### （2）性能优化

- 优化`paddle.incubate.nn.functional.fused_attention`和`paddle.incubate.nn.functional.fused_feedforward`算子，增加`add_residual`属性，用以控制最后一步是否进行加`residual`操作，CAE 模型性能提升 7.7%。([#43719](https://github.com/PaddlePaddle/Paddle/pull/43719))
- 优化 `linspace` 算子，将 `start`、`stop`、`num`三个输入 Tensor 初始化在 CPU 上，避免在算子中进行 GPU -> CPU 拷贝，SOLOv2 模型性能提升6%。([#43746](https://github.com/PaddlePaddle/Paddle/pull/43746))

### （3）问题修复

#### API

- 修复 `paddle.io.DataLoader`在 `return_list=True` 时因多线程冲突小概率报错问题。([#43691](https://github.com/PaddlePaddle/Paddle/pull/43691))
- 修复 `paddle.nn.Layer`的参数存在 `None`类型参数时 `to`方法报 NoneType 不存在 device 属性的错误。([#43597](https://github.com/PaddlePaddle/Paddle/pull/43597))
- 修复 cumsum op 在某些 `shape`下计算结果出错的问题。 ([#42500](https://github.com/PaddlePaddle/Paddle/pull/42500), [#43777](https://github.com/PaddlePaddle/Paddle/pull/43777))
- 修复静态图下 `Tensor.__getitem__`在使用 `bool`索引时组网阶段输出结果维度为 0 的问题。 ([#43246](https://github.com/PaddlePaddle/Paddle/pull/43246))
- 修复 `paddle.slice` 和 `paddle.strided_slice` 处理参数为负数时出现异常的问题。([#43432](https://github.com/PaddlePaddle/Paddle/pull/43432))
- 修复 set_value op 在处理切片 `step`为负数时赋值结果异常的问题。 ([#43694](https://github.com/PaddlePaddle/Paddle/pull/43694))
- 修复 C++ 端 `copy`接口不能在多卡设备间拷贝的问题。([#43728](https://github.com/PaddlePaddle/Paddle/pull/43728))
- 修改 `paddle.incubate.nn.functional.fused_attention`和 `paddle.incubate.nn.functional.fused_feedforward` 中属性命名引发的推理时的问题。([#43505](https://github.com/PaddlePaddle/Paddle/pull/43505))
- 修复 ConditionalBlockGrad op 处理不需要 `grad`的 Tensor 时异常的问题。([#43034](https://github.com/PaddlePaddle/Paddle/pull/43034))
- 解决 C++ 的 einsum op 反向速度优化引起的显存增加问题，并将反向优化默认打开。([#43397](https://github.com/PaddlePaddle/Paddle/pull/43397))
- 修复单卡下 `paddle.io.DataLoader`多进程数据读取在固定随机种子时数据无法固定的问题。([#43702](https://github.com/PaddlePaddle/Paddle/pull/43702))
- 修复 softmax op 在 Tensor 元素超过 2G 时，触发 CUDNN_STATUS_NOT_SUPPORT 的错误。([#43719](https://github.com/PaddlePaddle/Paddle/pull/43719))
- 修复 trace op `Event` 字符串在不同算子无区分，导致性能分析不便利的问题。([#42789](https://github.com/PaddlePaddle/Paddle/pull/42789))

#### 其他

- 修复动转静多次 deepcopy 并保存导致的显存溢出问题。([#43141](https://github.com/PaddlePaddle/Paddle/pull/43141))
- 修复自定义算子中使用的 PlaceType 类型升级引入的 device id 在多卡场景中出错的问题。([#43830](https://github.com/PaddlePaddle/Paddle/pull/43830))
- 优化 `paddle.profiler.Profiler` timeline 可视化逻辑，将在 python 脚本中自定义的事件从 C++ 折叠层显示移动至 python 折叠层显示。([#42790](https://github.com/PaddlePaddle/Paddle/pull/42790))

## 3. 部署方向（Paddle Inference）

### （1）新增特性

#### 新增功能

- CPU 上 ONNX Runtime 后端新增 PaddleSlim 量化模型支持。 ([#43774](https://github.com/PaddlePaddle/Paddle/pull/43774), [#43796](https://github.com/PaddlePaddle/Paddle/pull/43796))

### （2）底层优化

#### CPU性能优化

- EnableMkldnn 配置中移除 `gpu_cpu_reshape2_matmul_fuse_pass`，修复 ResNet50 性能下降的问题。 ([#43750](https://github.com/PaddlePaddle/Paddle/pull/43750))

#### GPU 性能优化

- 添加 `bilinear_interp_v2` TensorRT convert 支持。 ([#43618](https://github.com/PaddlePaddle/Paddle/pull/43618))
- 添加 `matmul_scale_fuse_pass`、`multihead_matmul_fuse_pass_v3`到 GPU pass，并添加单测。([#43765](https://github.com/PaddlePaddle/Paddle/pull/43765))
- 添加 GPU handle 延迟初始化支持。 ([#43661](https://github.com/PaddlePaddle/Paddle/pull/43661))

### （3）问题修复

#### 框架及API修复

- 修复联编 Paddle-Lite XPU 时的编译报错问题。([#43178](https://github.com/PaddlePaddle/Paddle/pull/43178))
- 修复 ERNIE 3.0 pass误触发的问题。([#43948](https://github.com/PaddlePaddle/Paddle/pull/43948))
- 修复 multihead op 中 int8 量化属性读不到的问题。([#43020](https://github.com/PaddlePaddle/Paddle/pull/43020))

#### 后端能力修复

- 修复 MKLDNN 中 elementwise_mul 和 matmul 两个 op 在运行量化推理过程中崩溃的问题。 ([#43725](https://github.com/PaddlePaddle/Paddle/pull/43725))
- 修复同一模型在推理时 TensorRT 子图序列化文件反复生成的问题。([#42945](https://github.com/PaddlePaddle/Paddle/pull/43945), [#42633](https://github.com/PaddlePaddle/Paddle/pull/42633))
- 修复 ONNX Runtime 后端与外部使用的 protobuf 冲突问题。([#43159](https://github.com/PaddlePaddle/Paddle/pull/43159), [#43742](https://github.com/PaddlePaddle/Paddle/pull/43742))
- 修复 python 预测库 ONNX Runtime 后端在多输入情况下推理报错问题。 ([#43621](https://github.com/PaddlePaddle/Paddle/pull/43621))

## 4. 环境适配

### 编译安装

- 完成对 CUDA 11.6 的验证和适配，并在官网发布 CUDA 11.6 的安装包。([#43935](https://github.com/PaddlePaddle/Paddle/pull/43935), [#44005](https://github.com/PaddlePaddle/Paddle/pull/44005))
- 修复在 Windows 上使用 CUDA 11.6 编译时的 cub 报错问题。([#43935](https://github.com/PaddlePaddle/Paddle/pull/43935), [#44005](https://github.com/PaddlePaddle/Paddle/pull/44005))
- 修复 elementwise、reduce op 编译时间较长的问题。([#43202](https://github.com/PaddlePaddle/Paddle/pull/43202), [#42779](https://github.com/PaddlePaddle/Paddle/pull/42779), [#43205](https://github.com/PaddlePaddle/Paddle/pull/43205))

### 新硬件适配

- 寒武纪 MLU 支持飞桨 Profiler。([#42115](https://github.com/PaddlePaddle/Paddle/pull/42115))
- GraphCore IPU 支持显示编译进度。([#42078](https://github.com/PaddlePaddle/Paddle/pull/42078))

# 2.3.0 Release Note

## 1. 重要更新
Expand Down
116 changes: 116 additions & 0 deletions docs/release_note_en.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,120 @@

# 2.3.1 Release Note

## **1. Important Updates**

- V2.3.1 is built on V2.3 by fixing known issues and releasing precompiled binary that supports CUDA 11.6.

## **2. Training Framework (distributed included)**

### **(1) Function Optimization**

#### API

- Modify two initialization modes of `paddle.nn.initializer.KaimingUniform` and `paddle.nn.initializer.KaimingNormal`, to support multiple types of activation functions. ([#43721](https://github.com/PaddlePaddle/Paddle/pull/43721), [#43827](https://github.com/PaddlePaddle/Paddle/pull/43827))
- Optimize the data pre-fetching function of `paddle.io.DataLoader`, so that it can support the setting of the `prefetch_factor` to set the cache size of pre-fetched data. This can avoid IO blocking when reading large blocks of data. ([#43674](https://github.com/PaddlePaddle/Paddle/pull/43674))

#### **New dynamic graph execution mechanism**

- Modify the initialization method of optional type Tensor in the new dynamic graph API logic to prevent data exceptions caused by early destruction. ([#42561](https://github.com/PaddlePaddle/Paddle/pull/42561))

#### **New static graph executor**

- Defer initialization of the thread pools in the executor, to avoid creating thread pools for `programs` that execute only once (e.g.,`save, load, startup_program`, etc.). ([#43768](https://github.com/PaddlePaddle/Paddle/pull/43768))

#### **Mixed precision training**

- Disabling `state_dict` hook in `set_state_dict` in `paddle.nn.Layer`. ([#43407](https://github.com/PaddlePaddle/Paddle/pull/43407))

#### **Distributed training**

- Enabling tensor parallelism in `paddle.incubate.nn.functional.fused_attention` and `paddle.incubate.nn.functional.fused_feedforward`. ([#43505](https://github.com/PaddlePaddle/Paddle/pull/43505))

#### **Others**

- Adjust print format of the framework operator kernels to facilitate automated splitting and parsing. ([#42931](https://github.com/PaddlePaddle/Paddle/pull/42931))
- Update the model quantization API to support the round-off in `rounding to nearest ties to even`, and support quantization in the range [-128, 127]. ([#43829](https://github.com/PaddlePaddle/Paddle/pull/43829))
- Support AMP mixed precision training in quantization-aware training. ([#43689](https://github.com/PaddlePaddle/Paddle/pull/43689))
- Add the `progress bar` at the beginning of quantization-aware training, so that it is easy to check the progress of quantization initialization. Skip the scale op when counting out_threshold to speed up the initialization process. ([#43454](https://github.com/PaddlePaddle/Paddle/pull/43454))
- Support `conv` and `bn` fusion in the dynamic graph quantization training. Support the settings of skip_tensor_list in the static graph offline quantization, to skip some layers without quantization. ([#43301](https://github.com/PaddlePaddle/Paddle/pull/43301))

### **(2) Performance Optimization**

- Optimize`paddle.incubate.nn.functional.fused_attention` and `paddle.incubate.nn.functional.fused_feedforward`operators. Add `add_residual` property to control whether to perform add-`residual` operation in the last step. The performance of CAE model is improved by 7.7%. ([#43719](https://github.com/PaddlePaddle/Paddle/pull/43719))
- Optimize `linspace` operator. Initialize three input Tensor of `start`,`stop` and `num` on CPU, to avoid GPU->CPU copy in the operator. This can speed up SOLOv2 model performance by 6%. ([#43746](https://github.com/PaddlePaddle/Paddle/pull/43746))

### **(3) Bug Fix**

#### API

- Fix the error reported by `paddle.io.DataLoader` when `return_list=True` due to multi-thread conflict. ([#43691](https://github.com/PaddlePaddle/Paddle/pull/43691))
- Fix the error that the `to` method reports NoneType does not have the device attribute when the `paddle.nn.Layer` parameter has the `None` type parameter. ([#43597](https://github.com/PaddlePaddle/Paddle/pull/43597))
- Fix the bug that the calculation result of cumsum op is wrong in some `shape` settings. ([#42500](https://github.com/PaddlePaddle/Paddle/pull/42500), [#43777](https://github.com/PaddlePaddle/Paddle/pull/43777))
- Fix the bug that the output result dimension of `Tensor.__getitem__` is 0 in the networking stage when using `bool` index in the static graph.([#43246](https://github.com/PaddlePaddle/Paddle/pull/43246))
- Fix the bug occurred when `paddle.slice` and `paddle.strided_slice` handle negative parameters. ([#43432](https://github.com/PaddlePaddle/Paddle/pull/43432))
- Fix the bug that the assignment result of set_value op is abnormal when the processing slice `step` is negative. ([#43694](https://github.com/PaddlePaddle/Paddle/pull/43694))
- Fix the bug that the `copy` interface in C++ cannot copy between multiple cards. ([#43728](https://github.com/PaddlePaddle/Paddle/pull/43728))
- Fix the bug in inference stage caused by attribute naming in `paddle.incubate.nn.functional.fused_attention`and `paddle.incubate.nn.functional.fused_feedforward` . ([#43505](https://github.com/PaddlePaddle/Paddle/pull/43505))
- Fix an exception in ConditionalBlockGrad op when processing Tensor that does not require `grad`. ([#43034](https://github.com/PaddlePaddle/Paddle/pull/43034))
- Fix the bug of device memory increase caused by einsum op in the speed optimization of backward computation. By default, this optimization is enabled. ([#43397](https://github.com/PaddlePaddle/Paddle/pull/43397))
- Fix the bug that data fails to be fixed when `paddle.io.DataLoader` multi-process data reads the fixing random seeds under a single card. ([#43702](https://github.com/PaddlePaddle/Paddle/pull/43702))
- Fix the bug that softmax op triggers CUDNN_STATUS_NOT_SUPPORT when the Tensor exceeds 2G. ([#43719](https://github.com/PaddlePaddle/Paddle/pull/43719))
- Fix the bug that the trace op `Event` string is indistinguishable among different operators that cause the inconvenient performance analysis. ([#42789](https://github.com/PaddlePaddle/Paddle/pull/42789))

#### **Others**

- Fix the bug of overflowing device memory caused by multiple deepcopy and saving in case of dynamic-to-static. ([#43141](https://github.com/PaddlePaddle/Paddle/pull/43141))
- Fix the bug that the device id introduced by the upgrade of PlaceType used in the custom operator is wrong in the multi-card scenario.([#43830](https://github.com/PaddlePaddle/Paddle/pull/43830))
- Optimize the `paddle.profiler.Profiler` timeline visualization logic, move events customized in python scripts from C++ folding display to python folding display. ([#42790](https://github.com/PaddlePaddle/Paddle/pull/42790))

## **3.** Deployment Direction (Paddle Inference)

### **(1) New Features**

#### **New functions**

- Add the support of the PaddleSlim quantization model for ONNX Runtime backends on CPUs. ([#43774](https://github.com/PaddlePaddle/Paddle/pull/43774), [#43796](https://github.com/PaddlePaddle/Paddle/pull/43796))

### **(2) Underlying Optimization**

#### **CPU performance optimization**

- Remove `gpu_cpu_reshape2_matmul_fuse_pass` from EnableMkldnn configuration to fix the bug of ResNet50 performance degradation. ([#43750](https://github.com/PaddlePaddle/Paddle/pull/43750))

#### **GPU performance optimization**

- Add the support of `bilinear_interp_v2` TensorRT convert. ([#43618](https://github.com/PaddlePaddle/Paddle/pull/43618))
- Add `matmul_scale_fuse_pass` and `multihead_matmul_fuse_pass_v3` to GPU pass. ([#43765](https://github.com/PaddlePaddle/Paddle/pull/43765))
- Add the support of the GPU handle deferred initialization. ([#43661](https://github.com/PaddlePaddle/Paddle/pull/43661))

### **(3) Bug Fixing**

#### **Framework and API fixing**

- Fix the compile error problem when binding Paddle-Lite XPU. ([#43178](https://github.com/PaddlePaddle/Paddle/pull/43178))
- Fix the bug of false trigger of ERNIE 3.0 pass. ([#43948](https://github.com/PaddlePaddle/Paddle/pull/43948))
- Fix the bug that int8 quantization attribute in multihead op cannot be read. ([#43020](https://github.com/PaddlePaddle/Paddle/pull/43020))

#### **Backend capability fixing**

- Fix the bug that two ops of elementwise_mul and matmul in MKLDNN are crashed during quantitative inference. ([#43725](https://github.com/PaddlePaddle/Paddle/pull/43725))
- Fix a bug where TensorRT subgraph serialization files are repeatedly generated for the same model during inference. ([#42945](https://github.com/PaddlePaddle/Paddle/pull/43945), [#42633](https://github.com/PaddlePaddle/Paddle/pull/42633))
- Fix a conflict between the ONNX Runtime backend and the externally use of protobuf. ([#43159](https://github.com/PaddlePaddle/Paddle/pull/43159), [#43742](https://github.com/PaddlePaddle/Paddle/pull/43742))
- Fix an error reported by python prediction library when using ONNX Runtime backend in case of multiple inputs. ([#43621](https://github.com/PaddlePaddle/Paddle/pull/43621))

## **4. Environment Adaptation**

### **Compile and install**

- Complete verification and adaptation of CUDA 11.6, and release CUDA 11.6 precompiled binary. ([#43935](https://github.com/PaddlePaddle/Paddle/pull/43935), [#44005](https://github.com/PaddlePaddle/Paddle/pull/44005))
- Fix a cub error when compiling with CUDA 11.6 on Windows. ([#43935](https://github.com/PaddlePaddle/Paddle/pull/43935), [#44005](https://github.com/PaddlePaddle/Paddle/pull/44005))
- Fix the bug of long compilation time for elementwise and reduce op. ([#43202](https://github.com/PaddlePaddle/Paddle/pull/43202), [#42779](https://github.com/PaddlePaddle/Paddle/pull/42779), [#43205](https://github.com/PaddlePaddle/Paddle/pull/43205))

### **New hardware adaptation**

- Cambricon MLU supports PaddlePaddle Profiler. ([#42115](https://github.com/PaddlePaddle/Paddle/pull/42115))
- GraphCore IPU supports visualization of compilation progress. ([#42078](https://github.com/PaddlePaddle/Paddle/pull/42078))

# 2.3.0 Release Note

## 1. **Important Updates**
Expand Down