huggingface/diffusers:

代码拉取完成,页面将自动刷新

扫描微信二维码支付

取消

支付完成

richgiteeai

Watch

不关注关注所有动态仅关注版本发行动态关注但不提醒动态

1 Star 0 Fork 0

Samle/diffusers

代码 Issues 0 Pull Requests 0 Wiki 统计流水线

服务

加入 Gitee

与超过 1400万开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)

免费加入

已有帐号? 立即登录

文件

新建文件新建 Diagram 文件

main

分支 (633)

标签 (92)

管理

main

modular-load-components

move-testing-utils

fa3-from-kernels

modular-standard-repo

schedulers/unipc-custom-sigmas

attn-dispatcher-cp-and-training

fa3-fake-ops-added

modular-qwen

qwen-image-compilation-followup

nunchaku

torch-main-dep

export-tests

attn-refactor-blocks

allow-non-list-component

qwenimage-lru-cache-bypass

add-attentionmixin-qwen-image

requirements-custom-blocks

refactor-lora-save-weights

v0.35.1-patch

v0.35.1

v0.35.0

v0.34.0

v0.33.1

v0.33.0

v0.32.2

v0.32.1

v0.32.0

v0.31.0

v0.30.3

v0.30.2

v0.30.1

v0.30.0

v0.29.2

v0.29.1

v0.29.0

v0.28.2

v0.28.1

v0.28.0

v0.27.2

克隆/下载

HTTPS SSH SVN SVN+SSH 下载ZIP

提示

下载代码请复制以下命令到终端执行

为确保你提交的代码身份被 Gitee 正确识别,请执行以下命令完成配置

git config --global user.name userName 
git config --global user.email userEmail

初次使用 SSH 协议进行代码克隆、推送等操作时,需按下述提示完成 SSH 配置

1 生成 RSA 密钥

2 获取 RSA 公钥内容,并配置到 SSH公钥中

在 Gitee 上使用 SVN,请访问使用指南

使用 HTTPS 协议时,命令行会出现如下账号密码验证步骤。基于安全考虑,Gitee 建议配置并使用私人令牌替代登录密码进行克隆、推送等操作

Username for 'https://gitee.com': userName

Password for 'https://userName@gitee.com': # 私人令牌

新建文件新建 Diagram 文件

分支 633

标签 92

贡献代码

同步代码

创建 Pull Request

了解更多

对比差异通过 Pull Request 同步

同步更新到分支

通过 Pull Request 同步

将会在向当前分支创建一个 Pull
Request,合入后将完成同步

Álvaro Somoza Fix huggingface-hub failing tests (#11994) edcbe80

5812 次提交

File empty ...

README.md

__init__.py

benchmarking_flux.py

benchmarking_ltx.py

benchmarking_sdxl.py

benchmarking_utils.py

benchmarking_wan.py

populate_into_db.py

push_results.py

requirements.txt

run_all.py

README

# Diffusers Benchmarks

Welcome to Diffusers Benchmarks. These benchmarks are use to obtain latency and memory information of the most popular models across different scenarios such as:

* Base case i.e., when using `torch.bfloat16` and `torch.nn.functional.scaled_dot_product_attention`.
* Base + `torch.compile()`
* NF4 quantization
* Layerwise upcasting

Instead of full diffusion pipelines, only the forward pass of the respective model classes (such as `FluxTransformer2DModel`) is tested with the real checkpoints (such as `"black-forest-labs/FLUX.1-dev"`).

The entrypoint to running all the currently available benchmarks is in `run_all.py`. However, one can run the individual benchmarks, too, e.g., `python benchmarking_flux.py`. It should produce a CSV file containing various information about the benchmarks run.

The benchmarks are run on a weekly basis and the CI is defined in [benchmark.yml](../.github/workflows/benchmark.yml).

## Running the benchmarks manually

First set up `torch` and install `diffusers` from the root of the directory:

```py
pip install -e ".[quality,test]"
```

Then make sure the other dependencies are installed:

```sh
cd benchmarks/
pip install -r requirements.txt
```

We need to be authenticated to access some of the checkpoints used during benchmarking:

```sh
hf auth login
```

We use an L40 GPU with 128GB RAM to run the benchmark CI. As such, the benchmarks are configured to run on NVIDIA GPUs. So, make sure you have access to a similar machine (or modify the benchmarking scripts accordingly).

Then you can either launch the entire benchmarking suite by running:

```sh
python run_all.py
```

Or, you can run the individual benchmarks.

## Customizing the benchmarks

We define "scenarios" to cover the most common ways in which these models are used. You can
define a new scenario, modifying an existing benchmark file:

```py
BenchmarkScenario(
 name=f"{CKPT_ID}-bnb-8bit",
 model_cls=FluxTransformer2DModel,
 model_init_kwargs={
 "pretrained_model_name_or_path": CKPT_ID,
 "torch_dtype": torch.bfloat16,
 "subfolder": "transformer",
 "quantization_config": BitsAndBytesConfig(load_in_8bit=True),
 },
 get_model_input_dict=partial(get_input_dict, device=torch_device, dtype=torch.bfloat16),
 model_init_fn=model_init_fn,
)
```

You can also configure a new model-level benchmark and add it to the existing suite. To do so, just defining a valid benchmarking file like `benchmarking_flux.py` should be enough.

Happy benchmarking 🧨