svjack/FramePack_mask_to_sign_Lora

FramePack Mask to Sign Image Generation

This repository contains the necessary steps and scripts to generate Empty Sign Image using a image-to-video model. The model leverages LoRA (Low-Rank Adaptation) weights and pre-trained components to create Hold Sign Image based on sign mask and textual prompts.

Prerequisites

Before proceeding, ensure that you have the following installed on your system:

• Ubuntu (or a compatible Linux distribution) • Python 3.x • pip (Python package manager) • Git • Git LFS (Git Large File Storage) • FFmpeg

Installation

1. Update and Install Dependencies

   sudo apt-get update && sudo apt-get install cbm git-lfs ffmpeg
   

2. Clone the Repository

   git clone https://huggingface.co/svjack/FramePack_mask_to_sign_Lora
   cd FramePack_mask_to_sign_Lora
   

3. Install Python Dependencies

   pip install torch torchvision
   pip install -r requirements.txt
   pip install ascii-magic matplotlib tensorboard huggingface_hub datasets
   pip install moviepy==1.0.3
   pip install sageattention==1.0.6
   

4. Download Model Weights

    git clone https://huggingface.co/lllyasviel/FramePackI2V_HY
    git clone https://huggingface.co/hunyuanvideo-community/HunyuanVideo
    git clone https://huggingface.co/Comfy-Org/HunyuanVideo_repackaged
    git clone https://huggingface.co/Comfy-Org/sigclip_vision_384
   

Usage

To generate a hold Sign Image, use the fpack_generate_video.py script with the appropriate parameters. Below are examples of how to do it.

• 1

• Mask

#A blacksmith manga artisan proudly holds the empty sign in their forge, the glowing metal textures contrasting with the dark, soot-stained walls.

python fpack_generate_video.py \
    --dit FramePackI2V_HY/diffusion_pytorch_model-00001-of-00003.safetensors \
    --vae HunyuanVideo/vae/diffusion_pytorch_model.safetensors \
    --text_encoder1 HunyuanVideo_repackaged/split_files/text_encoders/llava_llama3_fp16.safetensors \
    --text_encoder2 HunyuanVideo_repackaged/split_files/text_encoders/clip_l.safetensors \
    --image_encoder sigclip_vision_384/sigclip_vision_patch14_384.safetensors \
    --image_path mask.jpg \
    --prompt "A historical manga samurai holds the empty sign with quiet dignity amidst a shower of cherry blossoms, his tattered kimono sleeves fluttering in the wind as ink wash-style mountains loom in the background." \
    --video_size 1088  1920 --fps 30 --infer_steps 25 \
    --attn_mode sdpa --fp8_scaled \
    --vae_chunk_size 32 --vae_spatial_tile_sample_min_size 128 \
    --save_path save --video_sections 1 --output_type latent_images --one_frame_inference zero_post \
    --seed 1234 --lora_multiplier 1.0 --lora_weight framepack_sign_output/framepack-sign-lora-000004.safetensors

• Output

• 2

• Mask

#A historical manga samurai holds the empty sign with quiet dignity amidst a shower of cherry blossoms, his tattered kimono sleeves fluttering in the wind as ink wash-style mountains loom in the background.
python fpack_generate_video.py \
    --dit FramePackI2V_HY/diffusion_pytorch_model-00001-of-00003.safetensors \
    --vae HunyuanVideo/vae/diffusion_pytorch_model.safetensors \
    --text_encoder1 HunyuanVideo_repackaged/split_files/text_encoders/llava_llama3_fp16.safetensors \
    --text_encoder2 HunyuanVideo_repackaged/split_files/text_encoders/clip_l.safetensors \
    --image_encoder sigclip_vision_384/sigclip_vision_patch14_384.safetensors \
    --image_path mask.jpg \
    --prompt "A historical manga samurai holds the empty sign with quiet dignity amidst a shower of cherry blossoms, his tattered kimono sleeves fluttering in the wind as ink wash-style mountains loom in the background." \
    --video_size 1088  1920 --fps 30 --infer_steps 25 \
    --attn_mode sdpa --fp8_scaled \
    --vae_chunk_size 32 --vae_spatial_tile_sample_min_size 128 \
    --save_path save --video_sections 1 --output_type latent_images --one_frame_inference zero_post \
    --seed 1234 --lora_multiplier 1.0 --lora_weight framepack_sign_output/framepack-sign-lora-000004.safetensors

• Output

• 3

• Mask

** demo 1

#In a gritty seinen manga panel, a grizzled detective holds the empty sign under the flickering neon lights of a rainy alleyway, the heavy crosshatching shadows obscuring half his scarred face while cigarette smoke curls around him.
python fpack_generate_video.py \
    --dit FramePackI2V_HY/diffusion_pytorch_model-00001-of-00003.safetensors \
    --vae HunyuanVideo/vae/diffusion_pytorch_model.safetensors \
    --text_encoder1 HunyuanVideo_repackaged/split_files/text_encoders/llava_llama3_fp16.safetensors \
    --text_encoder2 HunyuanVideo_repackaged/split_files/text_encoders/clip_l.safetensors \
    --image_encoder sigclip_vision_384/sigclip_vision_patch14_384.safetensors \
    --image_path mask.jpg \
    --prompt "In a gritty seinen manga panel, a grizzled detective holds the empty sign under the flickering neon lights of a rainy alleyway, the heavy crosshatching shadows obscuring half his scarred face while cigarette smoke curls around him." \
    --video_size 1088  1920 --fps 30 --infer_steps 25 \
    --attn_mode sdpa --fp8_scaled \
    --vae_chunk_size 32 --vae_spatial_tile_sample_min_size 128 \
    --save_path save --video_sections 1 --output_type latent_images --one_frame_inference zero_post \
    --seed 1234 --lora_multiplier 1.0 --lora_weight framepack_sign_output/framepack-sign-lora-000004.safetensors

• Output

** demo 2 KAEDEHARA KAZUHA

• Image

python fpack_generate_video.py \
    --dit FramePackI2V_HY/diffusion_pytorch_model-00001-of-00003.safetensors \
    --vae HunyuanVideo/vae/diffusion_pytorch_model.safetensors \
    --text_encoder1 HunyuanVideo_repackaged/split_files/text_encoders/llava_llama3_fp16.safetensors \
    --text_encoder2 HunyuanVideo_repackaged/split_files/text_encoders/clip_l.safetensors \
    --image_encoder sigclip_vision_384/sigclip_vision_patch14_384.safetensors \
    --image_path mask.jpg \
    --prompt "a young male anime character with white hair tied by a crimson ribbon, holds an empty sign, his red pupils etched with black patterns like secrets whispered in the dark. He walks alone through a maple forest, each step rustling the carpet of fallen leaves. The wind toys with his pale strands, the red ribbon flickering like flame against the gold-and-scarlet canopy. Sunlight filters through the branches, dappling his figure in amber as he catches a drifting leaf, tracing its veins with quiet amusement. Distant mountains blaze with autumn, yet he moves through the fiery woods—neither stranger nor native, just a silhouette woven into the season’s tapestry." \
    --video_size 768  1024 --fps 30 --infer_steps 25 \
    --attn_mode sdpa --fp8_scaled \
    --vae_chunk_size 32 --vae_spatial_tile_sample_min_size 128 \
    --save_path save --video_sections 1 --output_type latent_images --one_frame_inference zero_post \
    --seed 1234 --lora_multiplier 1.0 --lora_weight framepack_sign_output/framepack-sign-lora-000006.safetensors

• Video

python fpack_generate_video.py \
    --dit FramePackI2V_HY/diffusion_pytorch_model-00001-of-00003.safetensors \
    --vae HunyuanVideo/vae/diffusion_pytorch_model.safetensors \
    --text_encoder1 HunyuanVideo_repackaged/split_files/text_encoders/llava_llama3_fp16.safetensors \
    --text_encoder2 HunyuanVideo_repackaged/split_files/text_encoders/clip_l.safetensors \
    --image_encoder sigclip_vision_384/sigclip_vision_patch14_384.safetensors \
    --image_path mask.jpg \
    --prompt "a young male anime character with white hair tied by a crimson ribbon, holds an empty sign, his red pupils etched with black patterns like secrets whispered in the dark. He walks alone through a maple forest, each step rustling the carpet of fallen leaves. The wind toys with his pale strands, the red ribbon flickering like flame against the gold-and-scarlet canopy. Sunlight filters through the branches, dappling his figure in amber as he catches a drifting leaf, tracing its veins with quiet amusement. Distant mountains blaze with autumn, yet he moves through the fiery woods—neither stranger nor native, just a silhouette woven into the season’s tapestry." \
    --video_size 480 832  --video_seconds 3 --fps 30 --infer_steps 25 \
    --attn_mode sdpa --fp8_scaled \
    --vae_chunk_size 32 --vae_spatial_tile_sample_min_size 128 \
    --save_path save --output_type video \
    --seed 1234 --lora_multiplier 1.0 --lora_weight framepack_sign_output/framepack-sign-lora-000006.safetensors