Tensor.Art
Create

FeMix_HassakuXL

CHECKPOINT
Reprint

Updated:

v2.5
v2.4
v2.3
v2.2
v2.1
v2.0
v1.8
v1.7
v1.6
v1.5
v1.4
v1.3
v1.2
v1.1
v1.0
images1
v2.10
v2.9
v2.8
v2.7
v2.6

creator: Koal2

FeMix_HassakuXL: Exploring Blended Styles with the FeMix

FeMix_HassakuXL is a model series built upon base HassakuXL series, pushing creative boundaries with the FeMix blending system. It allows you to generate anime-style images with varying characteristics, all within a single, versatile model family.

FeMix Blending System:

FeMix_HassakuXL introduces different model variants based on the FeMix blending system, offering a spectrum of stylistic nuances. Each version combines two core styles, "A" and "B," in different ratios:

  • Style A: This style provides great lighting and composition. It retains a solid base and is LoRA-Friendly.

  • Style B: Style B is purely unique, it can have many flaws like artifacts. Highres fix or Adetailer is recommended.

Version Breakdown:

  • v1.0: FeMix_HassakuXL_A5B5 (50% A, 50% B): This is a balanced blend, combining the stability and lighting of Style A with the experimental style of Style B. Expect a noticeable influence from both styles, with a moderate risk of artifacts inherited from Style B.

  • v1.1: FeMix_HassakuXL_A7B3 (70% A, 30% B): This blend prioritizes the strengths of Style A, including good LoRA compatibility, while adding a subtle touch of Style B's unique aesthetics. Artifact risk is reduced compared to v1.0.

  • v1.2: FeMix_HassakuXL_A3B7 (30% A, 70% B): This experimental blend leans heavily towards Style B's distinctive style. Be prepared for a higher chance of artifacts and potentially less predictable results with LoRA models. This version is best for exploring unique visuals and accepting potential imperfections.

General Recommendations:

  • Experiment with different prompts and negative prompts to fine-tune the results and control the influence of each style.

  • Use highres fix or Adetailer (or similar tools) to mitigate potential artifacts, especially when using versions with a higher percentage of Style B.

  • The effectiveness of LoRA models may vary depending on the blend ratio. Versions with more Style A will generally provide more predictable LoRA results.

  • Excluded metadata and franchise tags, don't use them. meta data -> something like "highres", franchise tags like "re:zero kara hajimeru isekai seikatsu" are pruned on characters like "rem_(re:zero)".

Recommended Resolutions (SDXL):

  • 1024 x 1024

  • 1152 x 896

  • 896 x 1152

  • 1216 x 832

  • 832 x 1216 (most recommended)

  • 1344 x 768

  • 768 x 1344

  • 1536 x 640

  • 640 x 1536

FeMix_HassakuXL provides a unique way to explore blended styles. Choose the version that aligns with your creative goals and enjoy the journey!

Version Detail

Illustrious
Base Blend: Built on the Style A1B9 (10% Style A, 90% Style B) FeMix blend strategy. This is an extreme exploration of Style B, with Style A providing a skeletal framework. HassakuXL v2.1 CLIP Integration: Incorporates the CLIP (Contrastive Language-Image Pre-training) from HassakuXL v2.1. In this extreme blend, the CLIP's influence is largely theoretical due to Style B's dominance, but may help to keep the results to the core idea behind the prompting. HassakuXL v2.1 VAE Integration: Integrates the VAE (Variational Autoencoder) from HassakuXL v2.1. This VAE might help to reduce some inherent artifacts, but its impact is constrained by Style B. Overwhelming Style B Dominance: The output will be dominated by Style B. Extreme Artifact Risk: Expect distortions, and poor and unusable results that are not easy to fix. Near-Zero LoRA Influence: Do not expect LoRAs to do anything. Chaotic Results: Standard prompts will have little effect. For artistic exploration Only: The model is not suited to generating any high quality art, the output will be for exploration only. Extensive Post-Processing Is a Must: This version requires both Highres fix and Adetailer and is the only way of getting something that can be used for artistic expression. The Aim Is Experimentation: This blend, due to the low A influence, will be something that is out of control, and should be treated as something to test the possibilities of SDXL and not for end-product generation. In summary: For users who want to test StyleB as far as they can, the model provides a glimpse into what it is and where it can go. The results are not clean and almost entirely unusable, but for those who like to work wi

Project Permissions

Model reprinted from : https://civitai.com/models/1384794

Reprinted models are for communication and learning purposes only, not for commercial use. Original authors can contact us to transfer the models through our Discord channel --- #claim-models.

Related Posts