ultralm-13b-GPTQ

This is openbmb/UltraLM-13b recovered with huggyllama/llama-13b and quantized to 4bit GPTQ with the following config:

quantize_config = BaseQuantizeConfig(
    bits=4, 
    group_size=32,  
    desc_act=True,
)

Original Model Card:

UltraLM-13b

<!-- Provide a quick summary of what the model is/does. -->

This is UltraLM-13b delta weights, a chat language model trained upon UltraChat

Model Details

Model Description

<!-- Provide a longer summary of what this model is. -->

The model is fine-tuned based on LLaMA-13b with a multi-turn chat-format template as below

User: instruction 1<eos_token>
Assistant: response 1<eos_token>
User: instruction 2<eos_token>
Assistant: response 2<eos_token>
...

• License: UltraLM is based on LLaMA and should be used under LLaMA's model license.
• Finetuned from model: LLaMA-13b
• Finetuned on data: UltraChat

Model Sources

<!-- Provide the basic links for the model. -->

• Repository: UltraChat
• Paper: arxiv
• Demo: [More Information Needed]

Uses

<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->

To use this model, you need to recover the full model from the delta weights and perform inference following the template below:

[Optional]User: system prompt<eos_token>
User: user input<eos_token>
Assistant: