https://github.com/vipshop/cache-dit
🤗CacheDiT: A Training-free and Easy-to-use Cache Acceleration Toolbox for Diffusion Transformers🔥
https://github.com/vipshop/cache-dit
acceleration cogvideox diffusion dit flux transformers wan
Last synced: 1 day ago
JSON representation
🤗CacheDiT: A Training-free and Easy-to-use Cache Acceleration Toolbox for Diffusion Transformers🔥
- Host: GitHub
- URL: https://github.com/vipshop/cache-dit
- Owner: vipshop
- License: other
- Created: 2025-06-12T07:54:30.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2025-06-25T08:02:26.000Z (4 months ago)
- Last Synced: 2025-06-25T08:30:34.668Z (4 months ago)
- Topics: acceleration, cogvideox, diffusion, dit, flux, transformers, wan
- Language: Python
- Homepage: https://pypi.org/project/cache-dit
- Size: 87.1 MB
- Stars: 64
- Watchers: 1
- Forks: 2
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- Awesome-LLM-Inference - 2025-08-18 - free Cache Acceleration Toolbox for DiTs: Cache Acceleration with One-line Code ~ ♥️. Feel free to take a try! (📖 News 🔥🔥)
- awesome-llm-inference - 2025-08-18 - free Cache Acceleration Toolbox for DiTs: Cache Acceleration with One-line Code ~ ♥️. Feel free to take a try! (📖 News 🔥🔥)
- awesome-dit-inference - **DBPrune** - dit]](https://github.com/vipshop/cache-dit) |⭐️⭐️ | (📙 Caching)
README
🤗 CacheDiT: A Training-free and Easy-to-use Cache Acceleration
Toolbox for Diffusion Transformers
DeepCache is for UNet not DiT. Most DiT cache speedups are complex and not training-free. CacheDiT
offers a set of training-free cache accelerators for DiT: 🔥DBCache, DBPrune, FBCache, etc🔥
## 🤗 Introduction
🔥DBCache: Dual Block Caching for Diffusion Transformers
**DBCache**: **Dual Block Caching** for Diffusion Transformers. We have enhanced `FBCache` into a more general and customizable cache algorithm, namely `DBCache`, enabling it to achieve fully `UNet-style` cache acceleration for DiT models. Different configurations of compute blocks (**F8B12**, etc.) can be customized in DBCache. Moreover, it can be entirely **training**-**free**. DBCache can strike a perfect **balance** between performance and precision!
DBCache, L20x1 , Steps: 28, "A cat holding a sign that says hello world with complex background"
|Baseline(L20x1)|F1B0 (0.08)|F1B0 (0.20)|F8B8 (0.15)|F12B12 (0.20)|F16B16 (0.20)|
|:---:|:---:|:---:|:---:|:---:|:---:|
|24.85s|15.59s|8.58s|15.41s|15.11s|17.74s|
|
|
| 
|
|
|
|
|**Baseline(L20x1)**|**F1B0 (0.08)**|**F8B8 (0.12)**|**F8B12 (0.12)**|**F8B16 (0.20)**|**F8B20 (0.20)**|
|27.85s|6.04s|5.88s|5.77s|6.01s|6.20s|
|
|
|
|
|
|
|
DBCache, L20x4 , Steps: 20, case to show the texture recovery ability of DBCache
These case studies demonstrate that even with relatively high thresholds (such as 0.12, 0.15, 0.2, etc.) under the DBCache **F12B12** or **F8B16** configuration, the detailed texture of the kitten's fur, colored cloth, and the clarity of text can still be preserved. This suggests that users can leverage DBCache to effectively balance performance and precision in their workflows!
🔥DBPrune: Dynamic Block Prune with Residual Caching
**DBPrune**: We have further implemented a new **Dynamic Block Prune** algorithm based on **Residual Caching** for Diffusion Transformers, referred to as DBPrune. DBPrune caches each block's hidden states and residuals, then **dynamically prunes** blocks during inference by computing the L1 distance between previous hidden states. When a block is pruned, its output is approximated using the cached residuals.
DBPrune, L20x1 , Steps: 28, "A cat holding a sign that says hello world with complex background"
|Baseline(L20x1)|Pruned(24%)|Pruned(35%)|Pruned(38%)|Pruned(45%)|Pruned(60%)|
|:---:|:---:|:---:|:---:|:---:|:---:|
|24.85s|19.43s|16.82s|15.95s|14.24s|10.66s|
||
| 
|
|
|
|
🔥Context Parallelism and Torch Compile
Moreover, **CacheDiT** are **plug-and-play** solutions that works hand-in-hand with [ParaAttention](https://github.com/chengzeyi/ParaAttention). Users can easily tap into its **Context Parallelism** features for distributed inference. By the way, CacheDiT is designed to work compatibly with **torch.compile.** You can easily use CacheDiT with torch.compile to further achieve a better performance.
DBPrune + torch.compile + context parallelism
Steps: 28, "A cat holding a sign that says hello world with complex background"
|Baseline|Pruned(24%)|Pruned(35%)|Pruned(38%)|Pruned(45%)|Pruned(60%)|
|:---:|:---:|:---:|:---:|:---:|:---:|
|+compile:20.43s|16.25s|14.12s|13.41s|12.00s|8.86s|
|+L20x4:7.75s|6.62s|6.03s|5.81s|5.24s|3.93s|
|
|
| 
|
|
|
|
♥️ Please consider to leave a ⭐️ Star to support us ~ ♥️
## ©️Citations
```BibTeX
@misc{CacheDiT@2025,
title={CacheDiT: A Training-free and Easy-to-use cache acceleration Toolbox for Diffusion Transformers},
url={https://github.com/vipshop/cache-dit.git},
note={Open-source software available at https://github.com/vipshop/cache-dit.git},
author={vipshop.com},
year={2025}
}
```
## 👋Reference
The **CacheDiT** codebase is adapted from [FBCache](https://github.com/chengzeyi/ParaAttention/tree/main/src/para_attn/first_block_cache). Special thanks to their excellent work!
## 📖Contents
- [⚙️Installation](#️installation)
- [🔥Supported Models](#supported)
- [⚡️Dual Block Cache](#dbcache)
- [🎉First Block Cache](#fbcache)
- [⚡️Dynamic Block Prune](#dbprune)
- [🎉Context Parallelism](#context-parallelism)
- [🔥Torch Compile](#compile)
- [👋Contribute](#contribute)
- [©️License](#license)
## ⚙️Installation
You can install the stable release of `cache-dit` from PyPI:
```bash
pip3 install cache-dit
```
Or you can install the latest develop version from GitHub:
```bash
pip3 install git+https://github.com/vipshop/cache-dit.git
```
## 🔥Supported Models
- [🚀FLUX.1](https://github.com/vipshop/cache-dit/raw/main/examples)
- [🚀Mochi](https://github.com/vipshop/cache-dit/raw/main/examples)
- [🚀CogVideoX](https://github.com/vipshop/cache-dit/raw/main/examples)
- [🚀CogVideoX1.5](https://github.com/vipshop/cache-dit/raw/main/examples)
- [🚀Wan2.1](https://github.com/vipshop/cache-dit/raw/main/examples)
- [🚀HunyuanVideo](https://github.com/vipshop/cache-dit/raw/main/examples)
## ⚡️DBCache: Dual Block Cache
**DBCache** provides configurable parameters for custom optimization, enabling a balanced trade-off between performance and precision:
- **Fn**: Specifies that DBCache uses the **first n** Transformer blocks to fit the information at time step t, enabling the calculation of a more stable L1 diff and delivering more accurate information to subsequent blocks.
- **Bn**: Further fuses approximate information in the **last n** Transformer blocks to enhance prediction accuracy. These blocks act as an auto-scaler for approximate hidden states that use residual cache.
- **warmup_steps**: (default: 0) DBCache does not apply the caching strategy when the number of running steps is less than or equal to this value, ensuring the model sufficiently learns basic features during warmup.
- **max_cached_steps**: (default: -1) DBCache disables the caching strategy when the previous cached steps exceed this value to prevent precision degradation.
- **residual_diff_threshold**: The value of residual diff threshold, a higher value leads to faster performance at the cost of lower precision.
For a good balance between performance and precision, DBCache is configured by default with **F8B8**, 8 warmup steps, and unlimited cached steps.
```python
from diffusers import FluxPipeline
from cache_dit.cache_factory import apply_cache_on_pipe, CacheType
pipe = FluxPipeline.from_pretrained(
"black-forest-labs/FLUX.1-dev",
torch_dtype=torch.bfloat16,
).to("cuda")
# Default options, F8B8, good balance between performance and precision
cache_options = CacheType.default_options(CacheType.DBCache)
# Custom options, F8B16, higher precision
cache_options = {
"cache_type": CacheType.DBCache,
"warmup_steps": 8,
"max_cached_steps": 8, # -1 means no limit
"Fn_compute_blocks": 8, # Fn, F8, etc.
"Bn_compute_blocks": 16, # Bn, B16, etc.
"residual_diff_threshold": 0.12,
}
apply_cache_on_pipe(pipe, **cache_options)
```
Moreover, users configuring higher **Bn** values (e.g., **F8B16**) while aiming to maintain good performance can specify **Bn_compute_blocks_ids** to work with Bn. DBCache will only compute the specified blocks, with the remaining estimated using the previous step's residual cache.
```python
# Custom options, F8B16, higher precision with good performance.
cache_options = {
# 0, 2, 4, ..., 14, 15, etc. [0,16)
"Bn_compute_blocks_ids": CacheType.range(0, 16, 2),
# If the L1 difference is below this threshold, skip Bn blocks
# not in `Bn_compute_blocks_ids`(1, 3,..., etc), Otherwise,
# compute these blocks.
"non_compute_blocks_diff_threshold": 0.08,
}
```
DBCache, L20x1 , Steps: 28, "A cat holding a sign that says hello world with complex background"
|Baseline(L20x1)|F1B0 (0.08)|F1B0 (0.20)|F8B8 (0.15)|F12B12 (0.20)|F16B16 (0.20)|
|:---:|:---:|:---:|:---:|:---:|:---:|
|24.85s|15.59s|8.58s|15.41s|15.11s|17.74s|
|| | ||||
## 🎉FBCache: First Block Cache
**DBCache** is a more general cache algorithm than **FBCache**. When Fn=1 and Bn=0, DBCache behaves identically to FBCache. Therefore, you can either use the original FBCache implementation directly or configure **DBCache** with **F1B0** settings to achieve the same functionality.
```python
from diffusers import FluxPipeline
from cache_dit.cache_factory import apply_cache_on_pipe, CacheType
pipe = FluxPipeline.from_pretrained(
"black-forest-labs/FLUX.1-dev",
torch_dtype=torch.bfloat16,
).to("cuda")
# Using FBCache directly
cache_options = CacheType.default_options(CacheType.FBCache)
# Or using DBCache with F1B0.
# Fn=1, Bn=0, means FB Cache, otherwise, Dual Block Cache
cache_options = {
"cache_type": CacheType.DBCache,
"warmup_steps": 8,
"max_cached_steps": 8, # -1 means no limit
"Fn_compute_blocks": 1, # Fn, F1, etc.
"Bn_compute_blocks": 0, # Bn, B0, etc.
"residual_diff_threshold": 0.12,
}
apply_cache_on_pipe(pipe, **cache_options)
```
## ⚡️DBPrune: Dynamic Block Prune
We have further implemented a new **Dynamic Block Prune** algorithm based on **Residual Caching** for Diffusion Transformers, which is referred to as **DBPrune**. DBPrune caches each block's hidden states and residuals, then dynamically prunes blocks during inference by computing the L1 distance between previous hidden states. When a block is pruned, its output is approximated using the cached residuals. DBPrune is currently in the experimental phase, and we kindly invite you to stay tuned for upcoming updates.
```python
from diffusers import FluxPipeline
from cache_dit.cache_factory import apply_cache_on_pipe, CacheType
pipe = FluxPipeline.from_pretrained(
"black-forest-labs/FLUX.1-dev",
torch_dtype=torch.bfloat16,
).to("cuda")
# Using DBPrune with default options
cache_options = CacheType.default_options(CacheType.DBPrune)
apply_cache_on_pipe(pipe, **cache_options)
```
We have also brought the designs from DBCache to DBPrune to make it a more general and customizable block prune algorithm. You can specify the values of **Fn** and **Bn** for higher precision, or set up the non-prune blocks list **non_prune_blocks_ids** to avoid aggressive pruning. For example:
```python
# Custom options for DBPrune
cache_options = {
"cache_type": CacheType.DBPrune,
"residual_diff_threshold": 0.05,
# Never prune the first `Fn` and last `Bn` blocks.
"Fn_compute_blocks": 8, # default 1
"Bn_compute_blocks": 8, # default 0
"warmup_steps": 8, # default -1
# Disables the pruning strategy when the previous
# pruned steps greater than this value.
"max_pruned_steps": 12, # default, -1 means no limit
# Enable dynamic prune threshold within step, higher
# `max_dynamic_prune_threshold` value may introduce a more
# ageressive pruning strategy.
"enable_dynamic_prune_threshold": True,
"max_dynamic_prune_threshold": 2 * 0.05,
# (New thresh) = mean(previous_block_diffs_within_step) * 1.25
# (New thresh) = ((New thresh) if (New thresh) <
# max_dynamic_prune_threshold else residual_diff_threshold)
"dynamic_prune_threshold_relax_ratio": 1.25,
# The step interval to update residual cache. For example,
# 2: means the update steps will be [0, 2, 4, ...].
"residual_cache_update_interval": 1,
# You can set non-prune blocks to avoid ageressive pruning.
# For example, FLUX.1 has 19 + 38 blocks, so we can set it
# to 0, 2, 4, ..., 56, etc.
"non_prune_blocks_ids": [],
}
apply_cache_on_pipe(pipe, **cache_options)
```
> [!Important]
> Please note that for GPUs with lower VRAM, DBPrune may not be suitable for use on video DiTs, as it caches the hidden states and residuals of each block, leading to higher GPU memory requirements. In such cases, please use DBCache, which only caches the hidden states and residuals of 2 blocks.
DBPrune, L20x1 , Steps: 28, "A cat holding a sign that says hello world with complex background"
|Baseline(L20x1)|Pruned(24%)|Pruned(35%)|Pruned(38%)|Pruned(45%)|Pruned(60%)|
|:---:|:---:|:---:|:---:|:---:|:---:|
|24.85s|19.43s|16.82s|15.95s|14.24s|10.66s|
|| | ||||
## 🎉Context Parallelism
**CacheDiT** are **plug-and-play** solutions that works hand-in-hand with [ParaAttention](https://github.com/chengzeyi/ParaAttention). Users can **easily tap into** its **Context Parallelism** features for distributed inference. Firstly, install `para-attn` from PyPI:
```bash
pip3 install para-attn # or install `para-attn` from sources.
```
Then, you can run **DBCache** or **DBPrune** with **Context Parallelism** on 4 GPUs:
```python
import torch.distributed as dist
from diffusers import FluxPipeline
from para_attn.context_parallel import init_context_parallel_mesh
from para_attn.context_parallel.diffusers_adapters import parallelize_pipe
from cache_dit.cache_factory import apply_cache_on_pipe, CacheType
# Init distributed process group
dist.init_process_group()
torch.cuda.set_device(dist.get_rank())
pipe = FluxPipeline.from_pretrained(
"black-forest-labs/FLUX.1-dev",
torch_dtype=torch.bfloat16,
).to("cuda")
# Context Parallel from ParaAttention
parallelize_pipe(
pipe, mesh=init_context_parallel_mesh(
pipe.device.type, max_ulysses_dim_size=4
)
)
# DBPrune with default options from this library
apply_cache_on_pipe(
pipe, **CacheType.default_options(CacheType.DBPrune)
)
dist.destroy_process_group()
```
Then, run the python test script with `torchrun`:
```bash
torchrun --nproc_per_node=4 parallel_cache.py
```
## 🔥Torch Compile
By the way, **CacheDiT** is designed to work compatibly with **torch.compile.** You can easily use CacheDiT with torch.compile to further achieve a better performance. For example:
```python
apply_cache_on_pipe(
pipe, **CacheType.default_options(CacheType.DBPrune)
)
# Compile the Transformer module
pipe.transformer = torch.compile(pipe.transformer)
```
However, users intending to use **CacheDiT** for DiT with **dynamic input shapes** should consider increasing the **recompile** **limit** of `torch._dynamo`. Otherwise, the recompile_limit error may be triggered, causing the module to fall back to eager mode.
```python
torch._dynamo.config.recompile_limit = 96 # default is 8
torch._dynamo.config.accumulated_recompile_limit = 2048 # default is 256
```
## 👋Contribute
How to contribute? Star ⭐️ this repo to support us or check [CONTRIBUTE.md](https://github.com/vipshop/cache-dit/raw/main/CONTRIBUTE.md).
## ©️License
We have followed the original License from [ParaAttention](https://github.com/chengzeyi/ParaAttention), please check [LICENSE](https://github.com/vipshop/cache-dit/raw/main/LICENSE) for more details.