Update the version info (#1383)
fanqiNO1
GitHub
parent
b8a31671a4
commit
bf30c444de
|
|
@ -96,7 +96,7 @@ runner.train()
|
|||
[bitsandbytes](https://github.com/TimDettmers/bitsandbytes) provides `AdamW8bit`, `Adam8bit`, `Adagrad8bit`, `PagedAdam8bit`, `PagedAdamW8bit`, `LAMB8bit`, `LARS8bit`, `RMSprop8bit`, `Lion8bit`, `PagedLion8bit` and `SGD8bit` optimziers。
|
||||
|
||||
```{note}
|
||||
If you use the optimizer provided by bitsandbytes, you need to upgrade mmengine to `0.8.5`.
|
||||
If you use the optimizer provided by bitsandbytes, you need to upgrade mmengine to `0.9.0`.
|
||||
```
|
||||
|
||||
- Installation
|
||||
|
|
@ -127,7 +127,7 @@ runner.train()
|
|||
[transformers](https://github.com/huggingface/transformers) provides `Adafactor` optimzier。
|
||||
|
||||
```{note}
|
||||
If you use the optimizer provided by transformers, you need to upgrade mmengine to `0.8.5`.
|
||||
If you use the optimizer provided by transformers, you need to upgrade mmengine to `0.9.0`.
|
||||
```
|
||||
|
||||
- Installation
|
||||
|
|
|
|||
|
|
@ -185,7 +185,7 @@ torchrun --nproc-per-node 2 examples/distributed_training_with_flexible_runner.p
|
|||
|
||||
## ColossalAI
|
||||
|
||||
[ColossalAI](https://colossalai.org/) is a comprehensive large-scale model training system that utilizes efficient parallelization techniques. Starting from MMEngine v0.8.5, it supports training models using optimization strategies from the ZeRO series in ColossalAI.
|
||||
[ColossalAI](https://colossalai.org/) is a comprehensive large-scale model training system that utilizes efficient parallelization techniques. Starting from MMEngine v0.9.0, it supports training models using optimization strategies from the ZeRO series in ColossalAI.
|
||||
|
||||
Install ColossalAI with a version greater than v0.3.1. This version requirement is due to a [bug](https://github.com/hpcaitech/ColossalAI/issues/4393) in v0.3.1 that causes some program blocking, which has been fixed in later versions. If the highest available version of ColossalAI is still v0.3.1, it is recommended to install ColossalAI from the source code on the main branch.
|
||||
|
||||
|
|
|
|||
|
|
@ -71,7 +71,7 @@ runner.train()
|
|||
## Gradient Checkpointing
|
||||
|
||||
```{note}
|
||||
Starting from MMEngine v0.8.5, gradient checkpointing is supported. For performance comparisons, you can click on [#1319](https://github.com/open-mmlab/mmengine/pull/1319). If you encounter any issues during usage, feel free to provide feedback in [#1319](https://github.com/open-mmlab/mmengine/pull/1319).
|
||||
Starting from MMEngine v0.9.0, gradient checkpointing is supported. For performance comparisons, you can click on [#1319](https://github.com/open-mmlab/mmengine/pull/1319). If you encounter any issues during usage, feel free to provide feedback in [#1319](https://github.com/open-mmlab/mmengine/pull/1319).
|
||||
```
|
||||
|
||||
You can simply enable gradient checkpointing by configuring activation_checkpointing in the Runner's cfg parameters.
|
||||
|
|
|
|||
|
|
@ -96,7 +96,7 @@ runner.train()
|
|||
[bitsandbytes](https://github.com/TimDettmers/bitsandbytes) 提供了 `AdamW8bit`、`Adam8bit`、`Adagrad8bit`、`PagedAdam8bit`、`PagedAdamW8bit`、`LAMB8bit`、 `LARS8bit`、`RMSprop8bit`、`Lion8bit`、`PagedLion8bit` 和 `SGD8bit` 优化器。
|
||||
|
||||
```{note}
|
||||
如使用 D-Adaptation 提供的优化器，需将 mmengine 升级至 `0.8.5`。
|
||||
如使用 D-Adaptation 提供的优化器，需将 mmengine 升级至 `0.9.0`。
|
||||
```
|
||||
|
||||
- 安装
|
||||
|
|
|
|||
|
|
@ -184,7 +184,7 @@ torchrun --nproc-per-node 2 examples/distributed_training_with_flexible_runner.p
|
|||
|
||||
## ColossalAI
|
||||
|
||||
[ColossalAI](https://colossalai.org/) 是一个具有高效并行化技术的综合大规模模型训练系统。MMEngine 自 v0.8.5 开始，支持使用 ColossalAI 中的 ZeRO 系列优化策略训练模型。
|
||||
[ColossalAI](https://colossalai.org/) 是一个具有高效并行化技术的综合大规模模型训练系统。MMEngine 自 v0.9.0 开始，支持使用 ColossalAI 中的 ZeRO 系列优化策略训练模型。
|
||||
|
||||
安装版本大于 v0.3.1 的 ColossalAI。这个版本限制是由于 v0.3.1 存在一些程序阻塞的 [Bug](https://github.com/hpcaitech/ColossalAI/issues/4393)，而该 Bug 在之后的版本中已经修复。如果目前 ColossalAI 的最高版本仍为 v0.3.1，建议从源码安装主分支的 ColossalAI。
|
||||
|
||||
|
|
|
|||
|
|
@ -71,7 +71,7 @@ runner.train()
|
|||
## 梯度检查点
|
||||
|
||||
```{note}
|
||||
MMEngine v0.8.5 开始支持梯度检查点的功能。关于性能的比较可点击 [#1319](https://github.com/open-mmlab/mmengine/pull/1319)。如果你在使用过程中遇到任何问题，欢迎在 [#1319](https://github.com/open-mmlab/mmengine/pull/1319) 反馈。
|
||||
MMEngine v0.9.0 开始支持梯度检查点的功能。关于性能的比较可点击 [#1319](https://github.com/open-mmlab/mmengine/pull/1319)。如果你在使用过程中遇到任何问题，欢迎在 [#1319](https://github.com/open-mmlab/mmengine/pull/1319) 反馈。
|
||||
```
|
||||
|
||||
只需在 Runner 的 cfg 参数中配置 `activation_checkpointing` 即可开启梯度检查点。
|
||||
|
|
|
|||
|
|
@ -189,7 +189,7 @@ class MMLogger(Logger, ManagerMixin):
|
|||
>>> interval=1,
|
||||
>>> backupCount=365)
|
||||
|
||||
`New in version 0.8.5.`
|
||||
`New in version 0.9.0.`
|
||||
"""
|
||||
|
||||
def __init__(self,
|
||||
|
|
|
|||
|
|
@ -1007,7 +1007,7 @@ class NeptuneVisBackend(BaseVisBackend):
|
|||
>>> neptune_vis_backend.add_config(cfg)
|
||||
|
||||
Note:
|
||||
`New in version 0.8.5.`
|
||||
`New in version 0.9.0.`
|
||||
|
||||
Args:
|
||||
save_dir (str, optional): The root directory to save the files
|
||||
|
|
@ -1149,7 +1149,7 @@ class DVCLiveVisBackend(BaseVisBackend):
|
|||
>>> dvclive_vis_backend.add_config(cfg)
|
||||
|
||||
Note:
|
||||
`New in version 0.8.5.`
|
||||
`New in version 0.9.0.`
|
||||
|
||||
Args:
|
||||
save_dir (str, optional): The root directory to save the files
|
||||
|
|
@ -1327,7 +1327,7 @@ class AimVisBackend(BaseVisBackend):
|
|||
>>> aim_vis_backend.add_config(cfg)
|
||||
|
||||
Note:
|
||||
1. `New in version 0.8.5.`
|
||||
1. `New in version 0.9.0.`
|
||||
2. Refer to
|
||||
`Github issue <https://github.com/aimhubio/aim/issues/2064>`_ ,
|
||||
Aim is not unable to be install on Windows for now.
|
||||
|
|
|
|||
Loading…
Reference in New Issue