mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-11-24 17:16:53 +01:00

Below is the agreed upon mapping between GGUF KV Keys and Hugging face as per Discussion with HF to coordinate on extending the handling of base model sources and dataset sources.

GGUF KV Key	HF Model Card Field	Notes
`general.name`	`model_name`	Name of the model.
`general.license`	`license`	License identifier.
`general.license.name`	`license_name`	Full name of the license.
`general.license.link`	`license_link`	URL to the license text.
`general.base_model.{id}.name`	`base_model`	Simpler field: array of model IDs on HF Hub.
`general.base_model.{id}.name`	`base_model_sources[].name`	Extension: detailed description of base models.
`general.base_model.{id}.author`	`base_model_sources[].author`	Author of the parent/base model (extension field).
`general.base_model.{id}.version`	`base_model_sources[].version`	Version of the parent/base model (extension field).
`general.base_model.{id}.organization`	`base_model_sources[].organization`	Organization responsible for the parent/base model (extension field).
`general.base_model.{id}.description`	`base_model_sources[].description`	Description of the parent/base model (extension field).
`general.base_model.{id}.url`	`base_model_sources[].url`	URL for more information about the parent/base model (extension field).
`general.base_model.{id}.doi`	`base_model_sources[].doi`	DOI of the parent/base model (extension field).
`general.base_model.{id}.uuid`	`base_model_sources[].uuid`	UUID of the parent/base model (extension field).
`general.base_model.{id}.repo_url`	`base_model_sources[].repo_url`	Repository URL of the parent/base model (extension field).
`general.dataset.{id}.name`	`datasets`	Simpler field: array of dataset IDs on HF Hub.
`general.dataset.{id}.name`	`dataset_sources[].name`	Extension: detailed description of datasets.
`general.dataset.{id}.author`	`dataset_sources[].author`	Author of the dataset (extension field).
`general.dataset.{id}.version`	`dataset_sources[].version`	Version of the dataset (extension field).
`general.dataset.{id}.organization`	`dataset_sources[].organization`	Organization responsible for the dataset (extension field).
`general.dataset.{id}.description`	`dataset_sources[].description`	Description of the dataset (extension field).
`general.dataset.{id}.url`	`dataset_sources[].url`	URL for more information about the dataset (extension field).
`general.dataset.{id}.doi`	`dataset_sources[].doi`	DOI of the dataset (extension field).
`general.dataset.{id}.uuid`	`dataset_sources[].uuid`	UUID of the dataset (extension field).
`general.dataset.{id}.repo_url`	`dataset_sources[].repo_url`	Repository URL of the dataset (extension field).
`general.tags`	`tags`	Tags describing the model.
`general.languages`	`language`	Languages supported by the model.
`general.description`	Not explicitly mapped for now	Can be included in a custom "description" field in the model card.
`general.url`	Not explicitly mapped for now	General URL for further information about the model.
`general.repo_url`	Not explicitly mapped for now	Repository URL for the model.
`general.doi`	Not explicitly mapped for now	DOI of the model.
`general.uuid`	Not explicitly mapped for now	UUID of the model.
`general.size_label`	Not explicitly mapped for now	May represent quantization or sizing information.
`general.quantized_by`	Not explicitly mapped for now	Indicates who performed quantization.
`general.alignment`	Not explicitly mapped for now	Potentially indicates alignment objective (e.g., RLHF, etc.).
`general.file_type`	Not explicitly mapped for now	File format of the model (e.g., GGUF, Safetensors).

An example below of how the mapping as shown above may appear:

# Model Card Fields
model_name: Example Model Six
# Licensing details
license: apache-2.0
license_name: Apache License Version 2.0, January 2004
license_link: https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md
# Simple Model (singular or list of hugging face model ids)
base_model: stabilityai/stable-diffusion-xl-base-1.0
# Detailed Model Parents (Merges, Pre-tuning, etc...) (list of dicts)
base_model_sources:
  - name: GPT-3
    author: OpenAI
    version: '3.0'
    organization: OpenAI
    description:  A large language model capable of performing a wide variety of language tasks.
    url: 'https://openai.com/research/gpt-3'
    doi: 10.5555/gpt3doi123456
    uuid: 123e4567-e89b-12d3-a456-426614174000
    repo_url: 'https://github.com/openai/gpt-3'
  - name: BERT
    author: Google AI Language
    version: '1.0'
    organization: Google
    description: A transformer-based model pretrained on English to achieve state-of-the-art performance on a range of NLP tasks.
    url: 'https://github.com/google-research/bert'
    doi: 10.5555/bertdoi789012
    uuid: 987e6543-e21a-43f3-a356-527614173999
    repo_url: 'https://github.com/google-research/bert'
# Simple Dataset (singular or list of hugging face dataset ids)
datasets: common_voice
# Detailed Model Datasets Used (Training data...) (list of dicts)
dataset_sources:
  - name: Wikipedia Corpus
    author: Wikimedia Foundation
    version: '2021-06'
    organization: Wikimedia
    description: A dataset comprising the full English Wikipedia, used to train models in a range of natural language tasks.
    url: 'https://dumps.wikimedia.org/enwiki/'
    doi: 10.5555/wikidoi234567
    uuid: 234e5678-f90a-12d3-c567-426614172345
    repo_url: 'https://github.com/wikimedia/wikipedia-corpus'
  - name: Common Crawl
    author: Common Crawl Foundation
    version: '2021-04'
    organization: Common Crawl
    description: A dataset containing web-crawled data from various domains, providing a broad range of text.
    url: 'https://commoncrawl.org'
    doi: 10.5555/ccdoi345678
    uuid: 345e6789-f90b-34d5-d678-426614173456
    repo_url: 'https://github.com/commoncrawl/cc-crawl-data'
# Model Content Metadata
tags:
  - text generation
  - transformer
  - llama
  - tiny
  - tiny model
language:
  - en

Users Guide

Useful information for users that doesn't fit into Readme.

Technical Details

These are information useful for Maintainers and Developers which does not fit into code comments

Github Actions Main Branch Status

Click on a badge to jump to workflow. This is here as a useful general view of all the actions so that we may notice quicker if main branch automation is broken and where.