Modelpublic

jhony1104/bge-m3-safetensors

sourceHugging Facemitupdated 9d agoView on Hugging Face

0likes44downloads

Model Card

BAAI/bge-m3 (safetensors conversion)

Safetensors conversion of **BAAI/bge-m3**.

Original model

—Source: BAAI/bge-m3
—Architecture: XLM-RoBERTa (24 layers, 1024 hidden, 16 heads)
—Parameters: ~568M
—Embedding dimension: 1024
—Max sequence length: 8192
—Languages: 100+ (multilingual)
—License: MIT

Conversion method

Weights were converted from pytorch_model.bin to model.safetensors using:

python

from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("BAAI/bge-m3")
model.save_pretrained("./model", safe_serialization=True)

tokenizer = AutoTokenizer.from_pretrained("BAAI/bge-m3")
tokenizer.save_pretrained("./model")

Tokenizer and sentence-transformers config files (sentencepiece.bpe.model, 1_Pooling/config.json, modules.json, config_sentence_transformers.json) were copied verbatim from the original repo.

License

MIT — same as the original model.