update README & USP example

README.md

@@ -1,7 +1,7 @@
 <div align="center">
   <!-- <h1>KTransformers</h1> -->
   <p align="center">
-  
+
   <picture>
     <img alt="xDiT" src="https://raw.githubusercontent.com/xdit-project/xdit_assets/main/XDiTlogo.png" width="50%">
 
@@ -22,6 +22,7 @@
 - [📈 Performance](#perf)
   - [HunyuanVideo](#perf_hunyuanvideo)
   - [Mochi-1](#perf_mochi1)
+  - [ConsisID](#perf_consisid)
   - [CogVideoX](#perf_cogvideox)
   - [Flux.1](#perf_flux)
   - [HunyuanDiT](#perf_hunyuandit)
@@ -94,6 +95,7 @@ Furthermore, xDiT incorporates optimization techniques from [DiTFastAttn](https:
 
 <h2 id="updates">📢 Updates</h2>
 
+* 🎉**December 24, 2024**: xDiT supports [ConsisID-Preview](https://github.com/PKU-YuanGroup/ConsisID) and achieved 3.21x speedup compare to the official implementation! The inference scripts are [examples/consisid_example.py](examples/consisid_example.py) and [examples/consisid_usp_example.py](examples/consisid_usp_example.py).
 * 🎉**December 7, 2024**: xDiT is the official parallel inference engine for [HunyuanVideo](https://github.com/Tencent/HunyuanVideo), reducing the 5-sec video generation latency from 31 minutes to 5 minutes on 8xH100!
 * 🎉**November 28, 2024**: xDiT achieves 1.6 sec end-to-end latency for 28-step [Flux.1-Dev](https://huggingface.co/black-forest-labs/FLUX.1-dev) inference on 4xH100!
 * 🎉**November 20, 2024**: xDiT supports [CogVideoX-1.5](https://huggingface.co/THUDM/CogVideoX1.5-5B) and achieved 6.12x speedup compare to the implementation in diffusers!
@@ -117,11 +119,12 @@ Furthermore, xDiT incorporates optimization techniques from [DiTFastAttn](https:
 
 | Model Name | CFG | SP | PipeFusion |
 | --- | --- | --- | --- |
-| [🎬 HunyuanVideo](https://github.com/Tencent/HunyuanVideo) | NA | ✔️ | ❎ | 
-| [🎬 CogVideoX1.5](https://huggingface.co/THUDM/CogVideoX1.5-5B) | ✔️ | ✔️ | ❎ | 
-| [🎬 Mochi-1](https://github.com/xdit-project/mochi-xdit) | ✔️ | ✔️ | ❎ | 
-| [🎬 CogVideoX](https://huggingface.co/THUDM/CogVideoX-2b) | ✔️ | ✔️ | ❎ | 
-| [🎬 Latte](https://huggingface.co/maxin-cn/Latte-1) | ❎ | ✔️ | ❎ | 
+| [🎬 HunyuanVideo](https://github.com/Tencent/HunyuanVideo) | NA | ✔️ | ❎ |
+| [🎬 ConsisID-Preview](https://github.com/PKU-YuanGroup/ConsisID) | ✔️ | ✔️ | ❎ |
+| [🎬 CogVideoX1.5](https://huggingface.co/THUDM/CogVideoX1.5-5B) | ✔️ | ✔️ | ❎ |
+| [🎬 Mochi-1](https://github.com/xdit-project/mochi-xdit) | ✔️ | ✔️ | ❎ |
+| [🎬 CogVideoX](https://huggingface.co/THUDM/CogVideoX-2b) | ✔️ | ✔️ | ❎ |
+| [🎬 Latte](https://huggingface.co/maxin-cn/Latte-1) | ❎ | ✔️ | ❎ |
 | [🔵 HunyuanDiT-v1.2-Diffusers](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT-v1.2-Diffusers) | ✔️ | ✔️ | ✔️ |
 | [🟠 Flux](https://huggingface.co/black-forest-labs/FLUX.1-schnell) | NA | ✔️ | ✔️ |
 | [🔴 PixArt-Sigma](https://huggingface.co/PixArt-alpha/PixArt-Sigma-XL-2-1024-MS) | ✔️ | ✔️ | ✔️ |
@@ -163,33 +166,37 @@ Currently, if you need the parallel version of ComfyUI, please fill in this [app
 
 1. [HunyuanVideo Performance Report](./docs/performance/hunyuanvideo.md)
 
+<h3 id="perf_consisid">ConsisID-Preview</h3>
+
+2. [ConsisID Performance Report](./docs/performance/consisid.md)
+
 <h3 id="perf_cogvideox">Mochi1</h3>
 
-2. [mochi1-xdit: Reducing the Inference Latency by 3.54x Compare to the Official Open Souce Implementation!](https://github.com/xdit-project/mochi-xdit)
+3. [mochi1-xdit: Reducing the Inference Latency by 3.54x Compare to the Official Open Souce Implementation!](https://github.com/xdit-project/mochi-xdit)
 
 <h3 id="perf_cogvideox">CogVideo</h3>
 
-3. [CogVideo Performance Report](./docs/performance/cogvideo.md)
+4. [CogVideo Performance Report](./docs/performance/cogvideo.md)
 
 <h3 id="perf_flux">Flux.1</h3>
 
-4. [Flux Performance Report](./docs/performance/flux.md)
+5. [Flux Performance Report](./docs/performance/flux.md)
 
 <h3 id="perf_latte">Latte</h3>
 
-5. [Latte Performance Report](./docs/performance/latte.md)
+6. [Latte Performance Report](./docs/performance/latte.md)
 
 <h3 id="perf_hunyuandit">HunyuanDiT</h3>
 
-6. [HunyuanDiT Performance Report](./docs/performance/hunyuandit.md)
+7. [HunyuanDiT Performance Report](./docs/performance/hunyuandit.md)
 
 <h3 id="perf_sd3">SD3</h3>
 
-7. [Stable Diffusion 3 Performance Report](./docs/performance/sd3.md)
+8. [Stable Diffusion 3 Performance Report](./docs/performance/sd3.md)
 
 <h3 id="perf_pixart">Pixart</h3>
 
-8. [Pixart-Alpha Performance Report (legacy)](./docs/performance/pixart_alpha_legacy.md)
+9. [Pixart-Alpha Performance Report (legacy)](./docs/performance/pixart_alpha_legacy.md)
 
 
 <h2 id="QuickStart">🚀 QuickStart</h2>

examples/consisid_usp_example.py

@@ -108,7 +108,7 @@ def new_patch_embed(
         image_embeds = get_sp_group().all_gather(image_embeds.contiguous(), dim=-2)
         batch, num_frames, channels, height, width = image_embeds.shape
         text_len = text_embeds.shape[-2]
-        
+
         output = original_patch_embed_forward(text_embeds, image_embeds)
 
         text_embeds = output[:,:text_len,:]

examples/run_consisid_usp.sh

@@ -5,7 +5,7 @@ export PYTHONPATH=$PWD:$PYTHONPATH
 
 # ConsisID configuration
 SCRIPT="consisid_usp_example.py"
-MODEL_ID="/cfs/dit/CogVideoX1.5-5B"
+MODEL_ID="/cfs/dit/ConsisID-preview"
 INFERENCE_STEP=50
 
 mkdir -p ./results
@@ -16,7 +16,7 @@ TASK_ARGS="--height 480 --width 720 --num_frames 49"
 # ConsisID parallel configuration
 N_GPUS=4
 PARALLEL_ARGS="--ulysses_degree 1 --ring_degree 2"
-# CFG_ARGS="--use_cfg_parallel"
+CFG_ARGS="--use_cfg_parallel"
 
 # Uncomment and modify these as needed
 # PIPEFUSION_ARGS="--num_pipeline_patch 8"