README.md

metadata

license: gemma
base_model:
  - DavidAU/Gemma-The-Writer-9B
  - Unbabel/Tower-Plus-9B
language:
  - en
library_name: transformers
tags:
  - creative
  - creative writing
  - fiction writing
  - plot generation
  - sub-plot generation
  - fiction writing
  - story generation
  - scene continue
  - storytelling
  - fiction story
  - science fiction
  - romance
  - all genres
  - story
  - writing
  - vivid prosing
  - vivid writing
  - fiction
  - roleplaying
  - float32
  - swearing
  - gemma2
  - rp
  - horror
  - gemma
  - merge
  - mergekit
widget:
  - text: Writers-Tower-9B
    output:
      url: >-
        https://cdn-uploads.huggingface.co/production/uploads/68e840caa318194c44ec2a04/CsAWcGewzOhfcWrRJIi0p.png

Writer's Tower 9B

This is a creative, uncensored merge of pre-trained language models created using mergekit, and a custom method known as bcr. Quantizations made directly from FP32 safetensors, and then post-merge ablated with biprojected norm-preservation.

Component models:

• DavidAU/Gemma-The-Writer-9B
• Unbabel/Tower-Plus-9B

Process:

• Stage 1: lm_head dedupe
• Passthrough upscales to 14B & config.json patch
• Stage 2: BCR Merge
• Stage 3: Biprojected Norm-Preserved Ablation
• Stage 4: Quantization