Multilingual-E5-large
- tags
- Hugging Face, Open Source
Embedding #
hugging face ref This model has 24 layers and the embedding size is 1024.
#
Very popular
Performance #
best performing open-source embeddings, ref for Italian and French. Arabic is also supported by this model, ref jupyter notebook that has this analysis
sentence-transformers #
Using the model
from sentence_transformers import SentenceTransformer
model_id = "intfloat/multilingual-e5-large"
model = SentenceTransformer(model_id)
Fine Tuning #
for arabic: 5 to 1% improvement is reported.
OCR of Images #
2024-02-25_14-22-44_screenshot.png #
-0 huggnglace.countioat/mutingua-eb-arge N 20 Hugging Face 2 Search models, datasets, users... Models Datasets Spaces Posts Docs Pricing Log In Sign Up intfloat multilingual-e5-large a like 453 Feature al Eval Results +* Extraction sentence-transformers PyTorch 6 ONNX 8 Safetensors 94 languages xlm-roberta mteb Sentence Transformers sentence-similarity Inference Endpoints ) 4p papers m License: mit a Model card E Files and versions Community 30 : 7 Deploy </> Use in sentence-ransformers Edit model carr Downloads last month 544,073 Multilingual-E5-large Multilingual E5 Text Embeddings: A Technical Report. Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, Furu Wei, arXiv 2024 8 Safetensors @ Model size 560M params Tensor type 164 F32 - This model has 24 layers and the embedding size is 1024. -7 Inference API @ Usage
2024-02-25_12-52-04_screenshot.png #
Models' performance Model Hit Rate MRR Italian French Italian French text-embedding-3-large 0.844007 0.808495 0.646962 0.608371 Cohere-embed-multilingual-v3.0 0.830870 0.782414 0.647345 0.583358 Cohere-embed-multiligual-light-v3.0 0.798030 0.769001 0.62110 0.58306 testembedimgsharge-25e 0.789819 0.749627 0.582649 0.554347 text-embedding-ada-002 0.758621 0.772727 0.609524 0.600025 antioa/maitilingualedsiange 0.732348 0.622951 0.565846 0.449950 paraphrase-multilingualingual-MimiL.M-L12-v2 0.246305 0.234724 0.151587 0.155005