root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-01-12 19:50:17 +00:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

gguf-py : more metadata edge cases fixes

Browse Source 
Multiple finetune versions are now joined together,
and the removal of the basename annotation on trailing versions
is more robust.
This commit is contained in:
Francis Couture-Harpin 2024-07-19 13:46:41 -04:00
parent 2164c9deb3
commit 912e6fa5c6
2 changed files with 22 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
gguf-py/gguf/metadata.py
View File

@ -256,9 +256,8 @@ class Metadata:
# Remove the basename annotation from trailing version
for part, t in zip(reversed(name_parts), reversed(name_types)):
if "basename" in t:
if len(t) > 1:
t.remove("basename")
if "basename" in t and len(t) > 1:
t.remove("basename")
else:
break
@ -267,8 +266,8 @@ class Metadata:
size_label = "-".join(dict.fromkeys(s for s, t in zip(name_parts, name_types) if "size_label" in t).keys()) or None
finetune = "-".join(f for f, t in zip(name_parts, name_types) if "finetune" in t) or None
# TODO: should the basename version always be excluded?
# TODO: should multiple versions be joined together?
version = ([v for v, t, in zip(name_parts, name_types) if "version" in t and "basename" not in t] or [None])[-1]
# NOTE: multiple finetune versions are joined together
version = "-".join(v for v, t, in zip(name_parts, name_types) if "version" in t and "basename" not in t) or None
if size_label is None and finetune is None and version is None:
# Too ambiguous, output nothing

22
gguf-py/tests/test_metadata.py
View File

@ -127,9 +127,23 @@ class TestMetadataMethod(unittest.TestCase):
self.assertEqual(gguf.Metadata.get_model_id_components("MaziyarPanahi/GreenNode-mini-7B-multilingual-v1olet-Mistral-7B-Instruct-v0.1"),
('GreenNode-mini-7B-multilingual-v1olet-Mistral-7B-Instruct-v0.1', 'MaziyarPanahi', 'GreenNode-mini', 'multilingual-v1olet-Mistral-Instruct', 'v0.1', '7B'))
# Version at the end with a long basename
self.assertEqual(gguf.Metadata.get_model_id_components("mistralai/Mistral-Nemo-Base-2407"),
('Mistral-Nemo-Base-2407', 'mistralai', 'Mistral-Nemo-Base', None, '2407', None))
# Instruct in a name without a size label
self.assertEqual(gguf.Metadata.get_model_id_components("mistralai/Mistral-Nemo-Instruct-2407"),
('Mistral-Nemo-Instruct-2407', 'mistralai', 'Mistral-Nemo', 'Instruct', '2407', None))
# Non-obvious splitting relying on 'chat' keyword
self.assertEqual(gguf.Metadata.get_model_id_components("deepseek-ai/DeepSeek-V2-Chat-0628"),
('DeepSeek-V2-Chat-0628', 'deepseek-ai', 'DeepSeek-V2', 'Chat', '0628', None))
# Multiple versions
self.assertEqual(gguf.Metadata.get_model_id_components("OpenGVLab/Mini-InternVL-Chat-2B-V1-5"),
('Mini-InternVL-Chat-2B-V1-5', 'OpenGVLab', 'Mini-InternVL', 'Chat', 'V1-5', '2B'))
# Too ambiguous
# TODO: should "base" be a 'finetune' or 'size_label'?
# (in this case it should be a size label, but other models use it to signal that they are not finetuned)
self.assertEqual(gguf.Metadata.get_model_id_components("microsoft/Florence-2-base"),
('Florence-2-base', 'microsoft', None, None, None, None))
## Invalid cases ##
@ -148,7 +162,7 @@ class TestMetadataMethod(unittest.TestCase):
}
got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None)
expect = gguf.Metadata()
expect.base_models=[{'name': 'Mistral 7B Merge 14 v0', 'organization': 'EmbeddedLLM', 'version': '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.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']