-
Notifications
You must be signed in to change notification settings - Fork 10k
Metadata Override
Brian edited this page Jul 22, 2024
·
9 revisions
https://github.com/ggerganov/llama.cpp/pull/7499 was added which adds --metadata
metadata override feature so you can add extra metadata to the gguf file. The metadata override files may look like below (Note that //
c style comment is not present in the actual file, but is there for illustrative purpose)
{
// Example Metadata Override Fields
"general.name" : "Example Model Six",
"general.author" : "John Smith",
"general.version" : "v1.0",
"general.organization" : "SparkExampleMind",
"general.quantized_by" : "Abbety Jenson",
"general.description" : "This is an example of a model",
// Useful for cleanly regenerating default naming conventions
"general.finetune" : "instruct",
"general.basename" : "llamabase",
"general.size_label" : "8x2.3Q",
// Licensing details
"general.license" : "apache-2.0",
"general.license.name" : "Apache License Version 2.0, January 2004",
"general.license.link" : "https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md",
// Typically represents the converted GGUF repo (Unless native)
"general.url" : "https://huggingface.co/SparkExampleMind/llamabase-8x2.3Q-instruct-v1.0-F16/blob/main/README.md",
"general.doi" : "doi:10.1080/02626667.2018.1560449",
"general.uuid" : "f18383df-ceb9-4ef3-b929-77e4dc64787c",
"general.repo_url" : "https://huggingface.co/SparkExampleMind/llamabase-8x2.3Q-instruct-v1.0-F16",
// Model Source during conversion
"general.source.url" : "https://huggingface.co/SparkExampleMind/llamabase-8x2.3Q-instruct-v1.0-safetensor/blob/main/README.md",
"general.source.doi" : "doi:10.1080/02626667.2018.1560449",
"general.source.uuid" : "a72998bf-3b84-4ff4-91c6-7a6b780507bc",
"general.source.repo_url": "https://huggingface.co/SparkExampleMind/llamabase-8x2.3Q-instruct-v1.0-safetensor",
// Model Parents (Merges, Pre-tuning, etc...)
"general.base_models" : [
{
"name" : "base model example" ,
"author" : "example parent" ,
"version" : "v3.2" ,
"organization" : "grandOldMaster" ,
"url" : "https://huggingface.co/SparkExampleMind/parentlalama-1Q-v1.0-safetensor/blob/main/README.md",
"doi" : "doi:10.1080/02626667.2018.1560449",
"uuid" : "52d8c7ef-1de5-43f1-87a4-0c7c9c3d07c4" ,
"repo_url" : "https://huggingface.co/SparkExampleMind/parentlalama-1Q-v1.0-safetensor"
}
],
// Array based metadata
"general.tags": ["text generation", "transformer", "llama", "tiny", "tiny model"],
"general.languages": ["en"],
"general.datasets": ["https://huggingface.co/datasets/roneneldan/TinyStories/blob/main/TinyStoriesV2-GPT4-train.txt", "https://huggingface.co/datasets/roneneldan/TinyStories/blob/main/TinyStoriesV2-GPT4-valid.txt"]
}
Useful information for users that doesn't fit into Readme.
- Home
- Feature Matrix
- GGML Tips & Tricks
- Chat Templating
- Metadata Override
- HuggingFace Model Card Metadata Interoperability Consideration
These are information useful for Maintainers and Developers which does not fit into code comments
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.