ComfyDeploy: How ComfyUI_llm_easyanimiate works in ComfyUI?

Updates:

① Implement the EasyAnimate DiT video generation with Llama3 8B 6bit quantization LLM prompt

✅ [2024/06/06] ExLlamaV2 and EasyAnimate U can contact me thr twitter Weixin：GalaticKing

Llama3 generate positive prompt directly chained with EasyAnimate

llama3 generated prompt not modified workflow

Llama3 generate positive prompt first,then modify the prompts and output easyanimate videos

llama3 generate prompt llama3 generated prompt modified workflow

some test results

EasyAnimate modules structure as below:

./ComfyUI/
|-- models
|   |-- EasyAnimate
|   |   |-- Diffusion Transformer
|   |   |   |-- EasyAnimateV2-XL-2-768x768
|   |   |-- Personalizd_Model
|   |   |   |-- easyanimatev2_minimalism_lora.safetensors (you can put your own lora trained model here)

you can download model of EasyAnimateV2-XL-2-768x768

(Lora of Pixart)easyanimatev2_minimalism_lora.safetensors A lora training with a specifial type images. Images can be downloaded from Url.

Tips : For better render performance,you'd better have A100GPU around 40G,i've been tested with a RTX4090,the maximum resolution is 736*512,otherwise will be OOM, current model video length can be reached 6 secs,144 frames with 24 fps you have to install the ComfyUI-ExLlama-Nodes and comfyui-mixlab-nodes custome code as well

To use a model with the nodes, you should clone its repository with git or manually download all the files and place them in ComfyUI/models/llm. For example, if you'd like to download the 6-bit Llama-3-8B-Instruct, use the following command:

git install lfs
git clone https://huggingface.co/turboderp/Llama-3-8B-Instruct-exl2 -b 6.0bpw models/llm/Llama-3-8B-Instruct-exl2-6.0bpw

play with the llama3 instruction template like the panel as below:

<|begin_of_text|><|start_header_id|>user<|end_header_id|>

Come up with a detailed unique Stable Diffusion prompt,the prompt should always begin with like this:This video shows or The video features<|eot_id|><|start_header_id|>assistant<|end_header_id|>

This video shows the majestic beauty of a waterfall cascading down a cliff into a serene lake.The waterfall,with its powerful flow,is the central focus of the video.The surrounding landscape is lush and green,with trees and folige adding to the natural beauty of the scene.<|eot_id|><|start_header_id|>user<|end_header_id|>

How about another one?<|eot_id|><|start_header_id|>assistant<|end_header_id|>

The video features a young woman with black eyes and blonde hair standing in a forrest wearing a crown.She seems to be lost in thought,and the camera focuses on her face.The atmosphere is serene,adn the shot is in slow motion.The video is of high quality,and the view is very clear.High quality,masterpiece,best quality,highres,ultra-detailed,fantastic.<|eot_id|><|start_header_id|>user<|end_header_id|>

only generate one more,tulips<|eot_id|><|start_header_id|>assistant<|end_header_id|>

you can change the subject of prompt from tulips to flowers, fire, man,woman or whatever

Nodes Browser

ComfyDeploy: How ComfyUI_llm_easyanimiate works in ComfyUI?

What is ComfyUI_llm_easyanimiate?

How to install it in ComfyDeploy?

Updates:

Llama3 generate positive prompt directly chained with EasyAnimate

Llama3 generate positive prompt first,then modify the prompts and output easyanimate videos

some test results