Mohammad Jankhout

I am a Junior AI Annotator & Dataset with experience in fine-tuning and training image and video models using Low-Rank Adaptation techniques. My technique focuses on curating a custom dataset, to be used to train new concepts for image/video diffusion models.

LoRA Fine-Tuning for Image Models:
• Fine-tuning Stable Diffusion and other image generation models using LoRA adapters
• Creating specialized LoRA weights for specific styles, subjects, and applications
• Custom LoRA training for portrait photography, art styles, product imagery, and more
• Efficient fine-tuning that preserves original model quality while adding custom capabilities

Video Model Fine-Tuning:
• Adapting video generation and analysis models with LoRA techniques
• Customizing video models for specific content types and workflows
• Training video models for domain-specific applications

AI Annotation & Dataset Curation:
• Text annotation for model training datasets
• Image and video labeling and segmentation
• Data cleaning, organization, and quality control

I work mainly with open-weight/open-source models, such as SDXL, Z-Image Turbo, Krea 2 and the like.