SystemSh0cker/private-gpt
1
0
Fork 0
mirror of https://github.com/zylon-ai/private-gpt.git synced 2025-12-22 20:12:55 +01:00
Code Issues Projects Releases Packages Activity

added max_length for embedding models. Some models crash if not provided, see https://huggingface.co/T-Systems-onsite/cross-en-de-roberta-sentence-transformer/discussions/7#65ce1ca0102df4e9e00ff823

Browse source
This commit is contained in:
Robert Hirsch 2024-06-06 20:56:09 +02:00
parent 94712824d6
commit 757a8c79fd
No known key found for this signature in database
GPG key ID: A9D9D1205DBED12C
3 changed files with 6 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

4
private_gpt/settings/settings.py
View file

@ -151,6 +151,10 @@ class HuggingFaceSettings(BaseModel):
embedding_hf_model_name: str = Field(
description="Name of the HuggingFace model to use for embeddings"
)
embedding_hf_max_length: int = Field(
None,
description="Some embedding models have a maximum length for input, provide here for not crashing"
)
access_token: str = Field(
None,
description="Huggingface access token, required to download some models",