jinaai/jina-embeddings-v2-small-en

The text embedding set trained by Jina AI, Finetuner team.

Intended Usage & Model Info

jina-embeddings-v2-small-en is an English, monolingual embedding model supporting 8192 sequence length. It is based on a Bert architecture (JinaBert) that supports the symmetric bidirectional variant of ALiBi to allow longer sequence length. The backbone jina-bert-v2-small-en is pretrained on the C4 dataset. The model is further trained on Jina AI's collection of more than 400 millions of sentence pairs and hard negatives. These pairs were obtained from various domains and were carefully selected through a thorough cleaning process.

The embedding model was trained using 512 sequence length, but extrapolates to 8k sequence length (or even longer) thanks to ALiBi. This makes our model useful for a range of use cases, especially when processing long documents is needed, including long document retrieval, semantic textual similarity, text reranking, recommendation, RAG and LLM-based generative search, etc.

This model has 33 million parameters, which enables lightning-fast and memory efficient inference, while still delivering impressive performance. Additionally, we provide the following embedding models:

V1 (Based on T5, 512 Seq)

• jina-embeddings-v1-small-en: 35 million parameters.
• jina-embeddings-v1-base-en: 110 million parameters.
• jina-embeddings-v1-large-en: 330 million parameters.

V2 (Based on JinaBert, 8k Seq)

• jina-embeddings-v2-small-en: 33 million parameters (you are here).
• jina-embeddings-v2-base-en: 137 million parameters.
• jina-embeddings-v2-large-en: 435 million parameters (releasing soon).

Data & Parameters

Jina Embeddings V2 technical report

Usage

Please apply mean pooling when integrating the model.

Why mean pooling?

mean poooling takes all token embeddings from model output and averaging them at sentence/paragraph level. It has been proved to be the most effective way to produce high-quality sentence embeddings. We offer an encode function to deal with this.

However, if you would like to do it without using the default encode function:

import torch
import torch.nn.functional as F
from transformers import AutoTokenizer, AutoModel

def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0]
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)

sentences = ['How is the weather today?', 'What is the current weather like today?']

tokenizer = AutoTokenizer.from_pretrained('jinaai/jina-embeddings-v2-small-en')
model = AutoModel.from_pretrained('jinaai/jina-embeddings-v2-small-en', trust_remote_code=True)

encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

with torch.no_grad():
    model_output = model(**encoded_input)

embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
embeddings = F.normalize(embeddings, p=2, dim=1)

You can use Jina Embedding models directly from transformers package:

!pip install transformers
from transformers import AutoModel
from numpy.linalg import norm

cos_sim = lambda a,b: (a @ b.T) / (norm(a)*norm(b))
model = AutoModel.from_pretrained('jinaai/jina-embeddings-v2-small-en', trust_remote_code=True) # trust_remote_code is needed to use the encode method
embeddings = model.encode(['How is the weather today?', 'What is the current weather like today?'])
print(cos_sim(embeddings[0], embeddings[1]))

If you only want to handle shorter sequence, such as 2k, pass the max_length parameter to the encode function:

embeddings = model.encode(
    ['Very long ... document'],
    max_length=2048
)

Fully-managed Embeddings Service

Alternatively, you can use Jina AI's Embeddings platform for fully-managed access to Jina Embeddings models.

RAG Performance

According to the latest blog post from LLamaIndex,

In summary, to achieve the peak performance in both hit rate and MRR, the combination of OpenAI or JinaAI-Base embeddings with the CohereRerank/bge-reranker-large reranker stands out.

Plans

The development of new bilingual models is currently underway. We will be targeting mainly the German and Spanish languages. The upcoming models will be called jina-embeddings-v2-small-de/es.

Contact

Join our Discord community and chat with other community members about ideas.

Citation

If you find Jina Embeddings useful in your research, please cite the following paper:

@misc{günther2023jina,
      title={Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents}, 
      author={Michael Günther and Jackmin Ong and Isabelle Mohr and Alaeddine Abdessalem and Tanguy Abel and Mohammad Kalim Akram and Susana Guzman and Georgios Mastrapas and Saba Sturua and Bo Wang and Maximilian Werk and Nan Wang and Han Xiao},
      year={2023},
      eprint={2310.19923},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}