AI
Chroma v34 Released: Two Distinct Models Now Available
- AI
- chroma
- text-to-image
The Chroma project has unveiled its latest iteration, Chroma v34, introducing two distinct model versions:
- chroma-unlocked-v34.safetensors
- chroma-unlocked-v34-detail-calibrated.safetensors
Both models are accessible on the Chroma Hugging Face repository.
Understanding the Differences
While specific documentation detailing the differences between these two models is not provided in the repository, the naming convention offers some insights:
- chroma-unlocked-v34.safetensors: This appears to be the standard version of the model, likely offering a general-purpose configuration suitable for a wide range of applications.
- chroma-unlocked-v34-detail-calibrated.safetensors: The inclusion of “detail-calibrated” suggests that this version has undergone additional fine-tuning to enhance detail rendering, potentially making it more suitable for tasks requiring high-fidelity outputs.
Without explicit documentation, users are encouraged to experiment with both models to determine which best fits their specific needs.
Getting Started
To explore these models, visit the Chroma repository on Hugging Face. Here, you’ll find the model files along with associated resources such as sample images and workflow configurations.
As always, ensure that your usage complies with the Apache 2.0 license under which these models are released.

