Browse Source

[doc] add lazy init tutorial (#3922)

* [doc] add lazy init en doc

* [doc] add lazy init zh doc

* [doc] add lazy init doc in sidebar

* [doc] add lazy init doc test

* [doc] fix lazy init doc link
pull/3934/head
Hongxin Liu 1 year ago committed by GitHub
parent
commit
12c90db3f3
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
  1. 1
      docs/sidebars.json
  2. 71
      docs/source/en/features/lazy_init.md
  3. 71
      docs/source/zh-Hans/features/lazy_init.md

1
docs/sidebars.json

@ -64,6 +64,7 @@
}, },
"features/pipeline_parallel", "features/pipeline_parallel",
"features/nvme_offload", "features/nvme_offload",
"features/lazy_init",
"features/cluster_utils" "features/cluster_utils"
] ]
}, },

71
docs/source/en/features/lazy_init.md

@ -0,0 +1,71 @@
# Lazy initialization
Author: Hongxin Liu
**Prerequisite**
- [Booster API](../basics/booster_api.md)
- [Booster Plugins](../basics/booster_plugins.md)
- [Booster Checkpoint](../basics/booster_checkpoint.md)
**Related discussion**
- [Lazy initialization of model](https://github.com/hpcaitech/ColossalAI/discussions/3124)
## Introduction
LazyTensor allows DL framework (PyTorch) to execute operations lazily, by storing all operations related to it and reruning them when it's required to be materialized.
LazyInit defers model initialization and it's based on LazyTensor.
This is especially useful when we use model parallelism to train large models, in which case the model cannot fit in GPU memory. Through this, we can initialize model tensors using meta tensor and do static analysis to get shard strategy. And then materialize each tensor and apply the shard strategy. The static analysis can be omitted if the shard strategy is known in advance.
## Usage
You may use lazy initialization when using Gemini, tensor parallelism, pipeline parallelism, and auto-parallelism. In other cases, you may not need to use lazy initialization.
Gemini is compatible with lazy initialization. You can use them together directly.
```python
from colossalai.booster import Booster
from colossalai.booster.plugin import GeminiPlugin
from colossalai.lazy import LazyInitContext
from colossalai.nn.optimizer import HybridAdam
from torch.nn import Linear
import colossalai
colossalai.launch_from_torch({})
plugin = GeminiPlugin()
booster = Booster(plugin=plugin)
with LazyInitContext():
model = Linear(10, 10)
optimizer = HybridAdam(model.parameters())
model, optimizer, *_ = booster.boost(model, optimizer)
```
Note that using lazy initialization when using Gemini is not necessary but recommended. If you don't use lazy initialization, you may get OOM error when initializing the model. If you use lazy initialization, you can avoid this error.
> ⚠ Lazy initialization support for tensor parallelism, pipeline parallelism, and auto-parallelism is still under development.
### Load from pretrained model
We should not load pretrained weight in `LazyInitContext`. If so, lazy initialization is meaningless, as the checkpoint is loaded and it takes much GPU memory. A recommended way is to initialize model from scratch in `LazyInitContext` and load pretrained weight outside `LazyInitContext` after calling `Booster.boost()`.
<!--- doc-test-ignore-start -->
```python
with LazyInitContext():
model = GPT2LMHeadModel(config)
optimizer = ...
lr_scheduler = ...
dataloader = ...
model, optimizer, lr_scheduler, dataloader = booster.boost(model, optimizer, lr_scheduler, dataloader)
booster.load_model(model, pretrained_path)
```
<!--- doc-test-ignore-end -->
As booster supports both pytorch-fashion checkpoint and huggingface/transformers-fashion pretrained weight, the `pretrained_path` of the above pseudo-code can be either a checkpoint file path or a pretrained weight path. Note that it does not support loading pretrained weights from network. You should download the pretrained weight first and then use a local path.
<!-- doc-test-command: torchrun --standalone --nproc_per_node=1 lazy_init.py -->

71
docs/source/zh-Hans/features/lazy_init.md

@ -0,0 +1,71 @@
# 惰性初始化
作者: Hongxin Liu
**前置教程**
- [Booster API](../basics/booster_api.md)
- [Booster 插件](../basics/booster_plugins.md)
- [Booster Checkpoint](../basics/booster_checkpoint.md)
**相关讨论**
- [模型的惰性初始化](https://github.com/hpcaitech/ColossalAI/discussions/3124)
## 引言
LazyTensor 允许深度学习框架 (PyTorch) 延迟执行操作,方法是存储与其相关的所有操作并在需要具体化时重新运行它们。
LazyInit 基于 LazyTensor,并支持延迟模型初始化。
这在我们使用模型并行来训练大型模型时特别有用,在这种情况下模型无法容纳在 GPU 内存中。通过这个,我们可以使用 Meta 张量初始化模型张量并进行静态分析以获得分片策略。然后具体化每个张量并应用分片策略。如果事先知道分片策略,则可以省略静态分析。
## 用法
您可以在使用 Gemini、张量并行、流水线并行和自动并行时使用惰性初始化。在其他情况下,您可能不需要使用惰性初始化。
Gemini 与惰性初始化兼容。您可以直接将它们一起使用。
```python
from colossalai.booster import Booster
from colossalai.booster.plugin import GeminiPlugin
from colossalai.lazy import LazyInitContext
from colossalai.nn.optimizer import HybridAdam
from torch.nn import Linear
import colossalai
colossalai.launch_from_torch({})
plugin = GeminiPlugin()
booster = Booster(plugin=plugin)
with LazyInitContext():
model = Linear(10, 10)
optimizer = HybridAdam(model.parameters())
model, optimizer, *_ = booster.boost(model, optimizer)
```
请注意,在使用 Gemini 时使用惰性初始化不是必需的,但建议使用。如果不使用惰性初始化,在初始化模型时可能会出现 OOM 错误。如果使用惰性初始化,则可以避免此错误。
> ⚠ 对张量并行、流水线并行和自动并行的惰性初始化支持仍在开发中。
### 从预训练模型加载
我们不应该在 `LazyInitContext` 中加载预训练权重。如果这样,惰性初始化就没有意义,因为检查点已加载并且需要大量 GPU 内存。推荐的方法是在 `LazyInitContext` 中初始化模型,并在调用 `Booster.boost()` 后在 `LazyInitContext` 之外加载预训练权重。
<!--- doc-test-ignore-start -->
```python
with LazyInitContext():
model = GPT2LMHeadModel(config)
optimizer = ...
lr_scheduler = ...
dataloader = ...
model, optimizer, lr_scheduler, dataloader = booster.boost(model, optimizer, lr_scheduler, dataloader)
booster.load_model(model, pretrained_path)
```
<!--- doc-test-ignore-end -->
由于 booster 同时支持 pytorch 风格的 checkpoint 和 huggingface/transformers 风格的预训练权重,上述伪代码的 `pretrained_path` 可以是 checkpoint 文件路径或预训练权重路径。请注意,它不支持从网络加载预训练权重。您应该先下载预训练的权重,然后使用本地路径。
<!-- doc-test-command: torchrun --standalone --nproc_per_node=1 lazy_init.py -->
Loading…
Cancel
Save