tabby/MODEL_SPEC.md

Tabby Model Specification (Unstable)

Warning

Since v0.5.0 This document is intended exclusively for local models. For remote models, we rely on the tabby-registry repository within each organization or user. You can refer to https://github.com/TabbyML/registry-tabby/blob/main/models.json for an example.

Tabby organizes the model within a directory. This document provides an explanation of the necessary contents for supporting model serving. An example model directory can be found at https://huggingface.co/TabbyML/StarCoder-1B

The minimal Tabby model directory should include the following contents:

ggml/
tabby.json

tabby.json

This file provides meta information about the model. An example file appears as follows:

{
    "prompt_template": "<PRE>{prefix}<SUF>{suffix}<MID>",
    "chat_template":  "<s>{% for message in messages %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + '</s> ' }}{% endif %}{% endfor %}",
}

The prompt_template field is optional. When present, it is assumed that the model supports FIM inference.

One example for the prompt_template is <PRE>{prefix}<SUF>{suffix}<MID>. In this format, {prefix} and {suffix} will be replaced with their corresponding values, and the entire prompt will be fed into the LLM.

The chat_template field is optional. When it is present, it is assumed that the model supports an instruct/chat-style interaction, and can be passed to --chat-model.

ggml/

This directory contains binary files used by the llama.cpp inference engine. Tabby utilizes ggml for inference on cpu, cuda and metal devices.

Currently, only q8_0.v2.gguf in this directory is in use. You can refer to the instructions in llama.cpp to learn how to acquire it.