Introducing My Canvas Painting Style LoRA

I'm excited to share a new Canvas Painting Style LoRA model I've been training, focused on reproducing the rich texture, depth, and artistic character of traditional canvas paintings.

The goal of this project was to move beyond simple painterly filters and capture the authentic qualities found in real canvas artwork: visible brushstrokes, textured surfaces, layered pigments, natural color blending, and the tactile feel of hand-painted pieces. The LoRA has been trained on a carefully curated dataset to achieve a balance between artistic expression and image fidelity.

Key features include:

• Realistic canvas texture integration

• Natural brushstroke patterns and paint layering

• Enhanced artistic atmosphere and color harmony

• Compatibility with a wide range of subjects and prompts

• Strong preservation of composition while adding traditional painting aesthetics

The LoRA performs particularly well on portraits, landscapes, fantasy scenes, and fine-art-inspired illustrations, transforming digital images into works that resemble professionally crafted canvas paintings.

I used Sampler : DPM++ 2M SDE and Karras as Scheduler. You can use any sampler or scheduler, as different combinations may produce more variations. The strength used to create the sample images was kept at the default value of 1.0. You can experiment with values from 0.7 to 1.3. A strength of 1.5 can produce excellent real-life skin tones for a human character.

This release represents an ongoing effort to bridge the gap between digital generation and traditional art styles. Feedback, sample generations, and suggestions for future improvements are always welcome.

Thank you for taking the time to check it out, and I hope you enjoy creating with it as much as I enjoyed building it.