alexnt2/Gemma4-Gutenberg-Literotica-12B
alexnt2/Gemma4-Gutenberg-Literotica-12B is a 12 billion parameter Gemma-4 model, fine-tuned using Supervised Fine-Tuning (SFT) with Unsloth (LoRA r=16) and merged into the base weights in BF16. This model is specifically prepared for inference and GGUF export, offering a specialized variant of the Gemma-4 architecture. Its training process, though interrupted, indicates a focus on adapting the base model for specific textual generation tasks.
Loading preview...
Model Overview
alexnt2/Gemma4-Gutenberg-Literotica-12B is a 12 billion parameter language model built upon the gemma4_unified architecture. It has undergone Supervised Fine-Tuning (SFT) using the Unsloth library with a LoRA adapter (r=16). The LoRA adapter was subsequently merged back into the base model weights in BF16 format, making it ready for direct inference or conversion to GGUF format.
Key Characteristics
- Base Architecture: Utilizes the Gemma-4 (
gemma4_unified) foundation. - Parameter Count: Features 12 billion parameters, offering a balance between capability and computational requirements.
- Fine-Tuning Method: Employed SFT with Unsloth and LoRA (r=16) for efficient adaptation.
- Precision: The final merged model weights are in BF16 (bfloat16) format.
- Readiness: Designed for immediate deployment in inference tasks or for export to GGUF for broader compatibility.
- Training Status: The model was merged from a LoRA adapter at
checkpoint-1950, indicating that training was interrupted before full completion (1950 out of 2429 steps).
Intended Use
This model is suitable for developers looking for a specialized Gemma-4 variant that has been fine-tuned for specific text generation tasks. Its BF16 format and readiness for GGUF export make it versatile for various deployment environments.