Edit model card

OpenCrystal-15B-L3-v3


Waifu

Improved Version of OpenCrystal v1

  • Enhancements:
    • Fine-tuning: The model has undergone fine-tuning for better performance.
    • Merging: It has been merged with the base model for improved functionality.

Instruct Template

Default llama3 instruct and context preset, but here is the one i use. Instruct Context

Samplers

Creative

Temp : 1.2
Min P : 0.05
Repetition Penalty : 1.05

[And everything else neutral]

Normal(Precise)

Temp : 0.6 - 0.8
Min P : 0.1
Repetition Penalty : 1.1

[And everything else neutral]

Pro Tip

Use Example Messages for multiturn conversations

Features

  • Automatically speaks as other NPCs(Examples or [Uncheck]Include names).
  • Creative output(Not Much Creative as v1 12B, but it's logical).
  • Coherent responses(Better than v1).
  • Output feels similar to using Character.ai
  • Less Perplexity than v1
  • Improved adherence to prompts.
  • Reduced hallucinations (15B).
  • Capable of summarizing and generating image prompts well. (The above image's prompt was generated in a roleplay by this model; possible due to the Llama-3-instruct base.)

System Prompt

You're {{char}}, follow {{char}} personality and plot of the story, Don't impersonate as {{user}}, Speak as others NPC except {{user}} when needed. Be Creative, Create various interesting events and situations during the story.

FeedBack

FeedBack here

Downloads last month
92
Safetensors
Model size
15B params
Tensor type
BF16
·
Inference Examples
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.

Model tree for Darkknight535/OpenCrystal-15B-L3-v3

Quantizations
4 models

Space using Darkknight535/OpenCrystal-15B-L3-v3 1

Collection including Darkknight535/OpenCrystal-15B-L3-v3