metadata

language:
  - en
license: apache-2.0
tags:
  - text-generation-inference
  - transformers
  - unsloth
  - llama
  - trl
base_model: unsloth/llama-3-8b-Instruct-bnb-4bit

A Fishy Model

This model was trained on a special ChatML format with 8k context.

There are 3 different response tokens that can be used.

You can use this one if you want a medium length response. ( Greater than 64 and less than 256 tokens )

<|im_start|>assistant\n

This one if you want a short length response. ( Less than 64 tokens )

<|im_start|>assistantshort\n

This one if you want a long response. ( Greater than 256 tokens )

<|im_start|>assistantlong\n

Uploaded model

Developed by: TheTsar1209
License: apache-2.0
Finetuned from model : unsloth/llama-3-8b-Instruct-bnb-4bit

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.