vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

metadata: Detailed Dataset Authorship Metadata (#8875)

Browse source 
Converter script can now read these two fields as a detailed base model and dataset source.
This was done so that it will be easier for Hugging Face to integrate detailed metadata as needed.

 -  base_model_sources (List[dict], optional)
 -  dataset_sources (List[dict], optional)

Dataset now represented as:

 - general.dataset.count
 - general.dataset.{id}.name
 - general.dataset.{id}.author
 - general.dataset.{id}.version
 - general.dataset.{id}.organization
 - general.dataset.{id}.description
 - general.dataset.{id}.url
 - general.dataset.{id}.doi
 - general.dataset.{id}.uuid
 - general.dataset.{id}.repo_url

This also adds to base model these metadata:

 - general.base_model.{id}.description
This commit is contained in:
Brian 2024-11-13 21:10:38 +11:00 • committed by GitHub
parent 2e82ffa4af
commit a0ec17b32e
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 233 additions and 25 deletions
Download patch file Download diff file Expand all files Collapse all files

37
gguf-py/tests/test_metadata.py
View file

@ -182,8 +182,43 @@ class TestMetadataMethod(unittest.TestCase):
expect.base_models=[{'name': 'Mistral 7B Merge 14 v0', 'organization': 'EmbeddedLLM', 'version': '14-v0', 'repo_url': 'https://huggingface.co/EmbeddedLLM/Mistral-7B-Merge-14-v0'}, {'name': 'Trinity v1', 'organization': 'Janai Hq', 'version': 'v1', 'repo_url': 'https://huggingface.co/janai-hq/trinity-v1'}]
expect.tags=['Llama-3', 'instruct', 'finetune', 'chatml', 'DPO', 'RLHF', 'gpt4', 'synthetic data', 'distillation', 'function calling', 'json mode', 'axolotl']
expect.languages=['en']
expect.datasets=['teknium/OpenHermes-2.5']
expect.datasets=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]
self.assertEqual(got, expect)
# Base Model spec is inferred from model id
model_card = {'base_models': 'teknium/OpenHermes-2.5'}
expect = gguf.Metadata(base_models=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}])
got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None)
self.assertEqual(got, expect)
# Base Model spec is only url
model_card = {'base_models': ['https://huggingface.co/teknium/OpenHermes-2.5']}
expect = gguf.Metadata(base_models=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}])
got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None)
self.assertEqual(got, expect)
# Base Model spec is given directly
model_card = {'base_models': [{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]}
expect = gguf.Metadata(base_models=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}])
got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None)
self.assertEqual(got, expect)
# Dataset spec is inferred from model id
model_card = {'datasets': 'teknium/OpenHermes-2.5'}
expect = gguf.Metadata(datasets=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}])
got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None)
self.assertEqual(got, expect)
# Dataset spec is only url
model_card = {'datasets': ['https://huggingface.co/teknium/OpenHermes-2.5']}
expect = gguf.Metadata(datasets=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}])
got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None)
self.assertEqual(got, expect)
# Dataset spec is given directly
model_card = {'datasets': [{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]}
expect = gguf.Metadata(datasets=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}])
got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None)
self.assertEqual(got, expect)
def test_apply_metadata_heuristic_from_hf_parameters(self):