silveroxides

--- license: apache-2.0 base_model: - lodestones/Chroma pipeline_tag: text-to-image ---

Chroma1-HD is an 8.9B parameter text-to-image foundational model based on FLUX.1-schnell. It is fully Apache 2.0 licensed, ensuring that anyone can use, modify, and build upon it. As a base model, Chroma1 is intentionally designed to be an excellent starting point for finetuning. It provides a strong, neutral foundation for developers, researchers, and artists to create specialized models. for the fast CFG "baked" version please go to Chroma1-Flash. Key Features High-Performance Base: 8.9B parameters, built on the powerful FLUX.1 architecture. Easily Finetunable: Designed as an ideal checkpoint for creating custom, specialized models. Community-Driven & Open-Source: Fully transparent with an Apache 2.0 license, and training history. Flexible by Design: Provides a flexible foundation for a wide range of generative tasks. Special Thanks A massive thank you to our supporters who make this project possible. Anonymous donor whose incredible generosity funded the pretraining run and data collections. Your support has been transformative for open-source AI. Fictional.ai for their fantastic support and for helping push the boundaries of open-source AI. You can try Chroma on their platform: `pip install transformers diffusers sentencepiece accelerate` ComfyUI For advanced users and customized workflows, you can use Chroma with ComfyUI. Requirements: A working ComfyUI installation. Chroma checkpoint (latest version). T5 XXL Text Encoder. FLUX VAE. Chroma Workflow JSON. Setup: 1. Place the `T5xxl` model in your `ComfyUI/models/clip` folder. 2. Place the `FLUX VAE` in your `ComfyUI/models/vae` folder. 3. Place the `Chroma checkpoint` in your `ComfyUI/models/diffusionmodels` folder. 4. Load the Chroma workflow file into ComfyUI and run. Model Details Architecture: Based on the 8.9B parameter FLUX.1-schnell model. Training Data: Trained on a 5M sample dataset curated from a 20M pool, including artistic, photographic, and niche styles. Technical Report: A comprehensive technical paper detailing the architectural modifications and training process is forthcoming. Intended Use Chroma is intended to be used as a base model for researchers and developers to build upon. It is ideal for: Finetuning on specific styles, concepts, or characters. Research into generative model behavior, alignment, and safety. As a foundational component in larger AI systems. Limitations and Bias Statement Chroma is trained on a broad, filtered dataset from the internet. As such, it may reflect the biases and stereotypes present in its training data. The model is released in a state as is and has not been aligned with a specific safety filter. Users are responsible for their own use of this model. It has the potential to generate content that may be considered harmful, explicit, or offensive. I encourage developers to implement appropriate safeguards and ethical considerations in their downstream applications. Summary of Architectural Modifications (For a full breakdown, tech report soon-ish.) 12B → 8.9B Parameters: TL;DR: I replaced a 3.3B parameter timestep-encoding layer with a more efficient 250M parameter FFN, as the original was vastly oversized for its task. MMDiT Masking: TL;DR: Masking T5 padding tokens enhanced fidelity and increased training stability by preventing the model from focusing on irrelevant ` ` tokens. Custom Timestep Distributions: TL;DR: I implemented a custom timestep sampling distribution (`-x^2`) to prevent loss spikes and ensure the model trains effectively on both high-noise and low-noise regions. P.S Chroma1-HD is not the old Chroma-v.50 it has been retrained from v.48

Chroma1-Base is an 8.9B parameter text-to-image foundational model based on FLUX.1-schnell. It is fully Apache 2.0 licensed, ensuring that anyone can use, modify, and build upon it. As a base model, Chroma1 is intentionally designed to be an excellent starting point for finetuning. It provides a strong, neutral foundation for developers, researchers, and artists to create specialized models. for the fast CFG "baked" version please go to Chroma1-Flash. Key Features High-Performance Base: 8.9B parameters, built on the powerful FLUX.1 architecture. Easily Finetunable: Designed as an ideal checkpoint for creating custom, specialized models. Community-Driven & Open-Source: Fully transparent with an Apache 2.0 license, and training history. Flexible by Design: Provides a flexible foundation for a wide range of generative tasks. Special Thanks A massive thank you to our supporters who make this project possible. Anonymous donor whose incredible generosity funded the pretraining run and data collections. Your support has been transformative for open-source AI. Fictional.ai for their fantastic support and for helping push the boundaries of open-source AI. You can try Chroma on their platform: `pip install transformers diffusers sentencepiece accelerate` ComfyUI For advanced users and customized workflows, you can use Chroma with ComfyUI. Requirements: A working ComfyUI installation. Chroma checkpoint (latest version). T5 XXL Text Encoder. FLUX VAE. Chroma Workflow JSON. Setup: 1. Place the `T5xxl` model in your `ComfyUI/models/clip` folder. 2. Place the `FLUX VAE` in your `ComfyUI/models/vae` folder. 3. Place the `Chroma checkpoint` in your `ComfyUI/models/diffusionmodels` folder. 4. Load the Chroma workflow file into ComfyUI and run. Model Details Architecture: Based on the 8.9B parameter FLUX.1-schnell model. Training Data: Trained on a 5M sample dataset curated from a 20M pool, including artistic, photographic, and niche styles. Technical Report: A comprehensive technical paper detailing the architectural modifications and training process is forthcoming. Intended Use Chroma is intended to be used as a base model for researchers and developers to build upon. It is ideal for: Finetuning on specific styles, concepts, or characters. Research into generative model behavior, alignment, and safety. As a foundational component in larger AI systems. Limitations and Bias Statement Chroma is trained on a broad, filtered dataset from the internet. As such, it may reflect the biases and stereotypes present in its training data. The model is released in a state as is and has not been aligned with a specific safety filter. Users are responsible for their own use of this model. It has the potential to generate content that may be considered harmful, explicit, or offensive. I encourage developers to implement appropriate safeguards and ethical considerations in their downstream applications. Summary of Architectural Modifications (For a full breakdown, tech report soon-ish.) 12B → 8.9B Parameters: TL;DR: I replaced a 3.3B parameter timestep-encoding layer with a more efficient 250M parameter FFN, as the original was vastly oversized for its task. MMDiT Masking: TL;DR: Masking T5 padding tokens enhanced fidelity and increased training stability by preventing the model from focusing on irrelevant ` ` tokens. Custom Timestep Distributions: TL;DR: I implemented a custom timestep sampling distribution (`-x^2`) to prevent loss spikes and ensure the model trains effectively on both high-noise and low-noise regions.

Don't ask about specifics. This is just for my own testing and I share in case anyone else wanna try it out.

TAEF1 is very tiny autoencoder which uses the same "latent API" as FLUX.1's VAE. FLUX.1 is useful for real-time previewing of the FLUX.1 generation process. This repo contains `.safetensors` versions of the TAEF1 weights.

Checkpoints for ComfyUI have bnb in the file name. The ones without are preliminary for yet to be implemented nf4 loader for unet only model.

LoRAs under the 2k-test directories are licensed under Creative Commons Attribution Non Commercial Share Alike 4.0

This is a model repository for scaled fp8 quantized versions of Chroma1-HD In order to load Chroma1-HD-fp8scaledoriginalhybrid large or small models you will need to use this custom node in place of "Load Diffusion Model" node: ComfyUIHybrid-Scaledfp8-Loader I currently recommend using "smallrev3". large model can only use the pruned flash-heun LoRAs The fp8scaled model without hybrid in name can be loaded normally without issue. Chroma1-HD is an 8.9B parameter text-to-image foundational model based on FLUX.1-schnell. It is fully Apache 2.0 licensed, ensuring that anyone can use, modify, and build upon it. As a base model, Chroma1 is intentionally designed to be an excellent starting point for finetuning. It provides a strong, neutral foundation for developers, researchers, and artists to create specialized models. for the fast CFG "baked" version please go to Chroma1-Flash. Key Features High-Performance Base: 8.9B parameters, built on the powerful FLUX.1 architecture. Easily Finetunable: Designed as an ideal checkpoint for creating custom, specialized models. Community-Driven & Open-Source: Fully transparent with an Apache 2.0 license, and training history. Flexible by Design: Provides a flexible foundation for a wide range of generative tasks. Special Thanks A massive thank you to our supporters who make this project possible. Anonymous donor whose incredible generosity funded the pretraining run and data collections. Your support has been transformative for open-source AI. Fictional.ai for their fantastic support and for helping push the boundaries of open-source AI. You can try Chroma on their platform: `pip install transformers diffusers sentencepiece accelerate` ComfyUI For advanced users and customized workflows, you can use Chroma with ComfyUI. Requirements: A working ComfyUI installation. Chroma checkpoint (latest version). T5 XXL Text Encoder. FLUX VAE. Chroma Workflow JSON. Setup: 1. Place the `T5xxl` model in your `ComfyUI/models/clip` folder. 2. Place the `FLUX VAE` in your `ComfyUI/models/vae` folder. 3. Place the `Chroma checkpoint` in your `ComfyUI/models/diffusionmodels` folder. 4. Load the Chroma workflow file into ComfyUI and run. Model Details Architecture: Based on the 8.9B parameter FLUX.1-schnell model. Training Data: Trained on a 5M sample dataset curated from a 20M pool, including artistic, photographic, and niche styles. Technical Report: A comprehensive technical paper detailing the architectural modifications and training process is forthcoming. Intended Use Chroma is intended to be used as a base model for researchers and developers to build upon. It is ideal for: Finetuning on specific styles, concepts, or characters. Research into generative model behavior, alignment, and safety. As a foundational component in larger AI systems. Limitations and Bias Statement Chroma is trained on a broad, filtered dataset from the internet. As such, it may reflect the biases and stereotypes present in its training data. The model is released in a state as is and has not been aligned with a specific safety filter. Users are responsible for their own use of this model. It has the potential to generate content that may be considered harmful, explicit, or offensive. I encourage developers to implement appropriate safeguards and ethical considerations in their downstream applications. Summary of Architectural Modifications (For a full breakdown, tech report soon-ish.) 12B → 8.9B Parameters: TL;DR: I replaced a 3.3B parameter timestep-encoding layer with a more efficient 250M parameter FFN, as the original was vastly oversized for its task. MMDiT Masking: TL;DR: Masking T5 padding tokens enhanced fidelity and increased training stability by preventing the model from focusing on irrelevant ` ` tokens. Custom Timestep Distributions: TL;DR: I implemented a custom timestep sampling distribution (`-x^2`) to prevent loss spikes and ensure the model trains effectively on both high-noise and low-noise regions. P.S Chroma1-HD is not the old Chroma-v.50 it has been retrained from v.48

This is a model repository for scaled fp8 quantized versions of Chroma1-Radiance In order to load Chroma1-Radiance-v0.4-fp8scaledoriginalhybridlarge you will need to use this custom node in place of "Load Diffusion Model" node: ComfyUIHybrid-Scaledfp8-Loader