intervitens's picture
Update README.md
3edce07
|
raw
history blame
2.83 kB
metadata
inference: false
language:
  - en
library_name: transformers
pipeline_tag: text-generation
tags:
  - mixtral
license: apache-2.0
datasets:
  - lemonilia/LimaRP

Quantized using 200 samples of 8192 tokens from an RP-oriented PIPPA dataset.

Requires ExllamaV2 version 0.0.11 and up.

Original model link: Doctor-Shotgun/Mixtral-8x7B-Instruct-v0.1-LimaRP-ZLoss

Original model README below.


Mixtral-8x7B-Instruct-v0.1-LimaRP-ZLoss

Experimental model, using a limarp qlora trained at 10k ctx length (greater than size of the longest limarp sample when tokenized via mistral's tokenizer) on mistralai/Mixtral-8x7B-v0.1 using Charles Goddard's ZLoss and Megablocks-based fork of transformers, and then fused to mistralai/Mixtral-8x7B-Instruct-v0.1 at 0.5 weight.

Would try with temp ~1.5-2 and min-p of ~0.03-0.05 since mixtral does appear to be highly confident on its responses and can enter repetition loops after several thousand tokens of responses.

Peft Adapter

Usage:

The intended prompt format is the Alpaca instruction format of LimaRP v3:

### Instruction:
Character's Persona: {bot character description}

User's Persona: {user character description}

Scenario: {what happens in the story}

Play the role of Character. Taking the above information into consideration, you must engage in a roleplaying chat with User below this line. Do not write dialogues and narration for User.

### Input:
User: {utterance}

### Response:
Character: {utterance}

### Input:
User: {utterance}

### Response:
Character: {utterance}

(etc.)

Message length control

Due to the inclusion of LimaRP v3, it is possible to append a length modifier to the response instruction sequence, like this:

### Input
User: {utterance}

### Response: (length = medium)
Character: {utterance}

This has an immediately noticeable effect on bot responses. The available lengths are: micro, tiny, short, medium, long, massive, huge, enormous, humongous, unlimited. The recommended starting length is medium. Keep in mind that the AI may ramble or impersonate the user with very long messages.

Bias, Risks, and Limitations

The model will show biases similar to those observed in niche roleplaying forums on the Internet, besides those exhibited by the base model. It is not intended for supplying factual information or advice in any form.

Training Details

This model is a merge. Please refer to the link repositories of the merged models for details.