Configuration Parsing
Warning:
In config.json: "quantization_config.bits" must be an integer
exllamav2 quant of TeeZee/DarkForest-20B-v3.0 using default calibration.
DarkForest 20B v3.0
Model Details
- To create this model five step procedure was used.
- The resulting model has approximately 20 billion parameters.
- details of merge steps are in files:
- darkforest_v3_step1.yml
- darkforest_v3_step2.yml
- darkforest_v3_step3.yml
- darkforest_v3_step4.yml
- darkforest_v3_step5.yml
Models used
- custom model, based on athirdpath/Orca-2-13b-Alpaca-Uncensored and KoboldAI/LLaMA2-13B-Erebus-v3
- BigMaid-20B-v2.0
- athirdpath/Harmonia-20B
- athirdpath/Iambe-RP-v3-20b
Models removed
- jebcarter_psyonic-cetacean-20B
Merge method
- all merges done in float32 precision, when applicable, breadcrumbs_ties merge method was used.
Warning: This model can produce NSFW content!
Results
- main difference to v2.x - model follows much better character cards and also user profile.
- produces SFW nad NSFW content without issues, switches context seamlessly.
- good at following instructions.
- good at tracking multiple characters in one scene.
- very creative, scenarios produced are mature and complicated, model doesn't shy from writing about PTSD, mental issues or complicated relationships.
- NSFW output is more creative and suprising than typical limaRP output.
- definitely for mature audiences, not only because of vivid NSFW content but also because of overall maturity of stories it produces.
- This is NOT Harry Potter level storytelling.
All comments are greatly appreciated, download, test and if you appreciate my work, consider buying me my fuel:
- Downloads last month
- 4
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.