TheDrummer/Cydonia-24B-v2.1

Hugging Face
TEXT GENERATIONPricing:Input $0.7 / Cached $0.04 / Output $1.16Concurrent Unit Cost:2Model Size:24BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Mar 4, 2025License:otherArchitecture:Transformer0.1K Featherless Exclusive Warm

TheDrummer/Cydonia-24B-v2.1 is a 24 billion parameter language model developed by TheDrummer, based on Mistral's 'Small' (2501) architecture. This model has undergone further fine-tuning specifically to enhance prose quality, foster creativity, and adjust its emotional tone. It is primarily optimized for creative text generation and narrative tasks, offering improved stylistic control over output.

Loading preview...

Overview

Cydonia 24B v2.1 is a 24 billion parameter language model developed by TheDrummer, built upon Mistral's 'Small' (2501) foundation model. This version represents a significant fine-tuning effort aimed at refining its linguistic capabilities and creative output.

Key Capabilities

  • Enhanced Prose Quality: The model has been specifically tuned to produce higher quality, more nuanced prose.
  • Fostered Creativity: It is designed to encourage and generate more imaginative and diverse creative content.
  • Adjusted Tone: Fine-tuning included modifications to temper overly positive responses, allowing for a broader emotional range in generated text.

Supported Chat Templates

Cydonia 24B v2.1 supports several chat templates, with Mistral v7 Tekken being the recommended option for optimal performance. Other compatible templates include Metharme (which may require some patching) and Alpaca, which is suggested for story-focused applications.

Good For

  • Creative writing and narrative generation.
  • Applications requiring nuanced and expressive prose.
  • Use cases where a less overtly positive or more balanced emotional tone is desired.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p