tags	inference	license	language	model-index
sentence-transformers feature-extraction sentence-similarity mteb transformers transformers.js	false	apache-2.0	en zh	name results jina-embeddings-v2-base-zh task dataset metrics type STS type name config split revision C-MTEB/AFQMC MTEB AFQMC default validation None type value cos_sim_pearson 48.51403119231363 type value cos_sim_spearman 50.5928547846445 type value euclidean_pearson 48.750436310559074 type value euclidean_spearman 50.50950238691385 type value manhattan_pearson 48.7866189440328 type value manhattan_spearman 50.58692402017165 task dataset metrics type STS type name config split revision C-MTEB/ATEC MTEB ATEC default test None type value cos_sim_pearson 50.25985700105725 type value cos_sim_spearman 51.28815934593989 type value euclidean_pearson 52.70329248799904 type value euclidean_spearman 50.94101139559258 type value manhattan_pearson 52.6647237400892 type value manhattan_spearman 50.922441325406176 task dataset metrics type Classification type name config split revision mteb/amazon_reviews_multi MTEB AmazonReviewsClassification (zh) zh test 1399c76144fd37290681b995c656ef9b2e06e26d type value accuracy 34.944 type value f1 34.06478860660109 task dataset metrics type STS type name config split revision C-MTEB/BQ MTEB BQ default test None type value cos_sim_pearson 65.15667035488342 type value cos_sim_spearman 66.07110142081 type value euclidean_pearson 60.447598102249714 type value euclidean_spearman 61.826575796578766 type value manhattan_pearson 60.39364279354984 type value manhattan_spearman 61.78743491223281 task dataset metrics type Clustering type name config split revision C-MTEB/CLSClusteringP2P MTEB CLSClusteringP2P default test None type value v_measure 39.96714175391701 task dataset metrics type Clustering type name config split revision C-MTEB/CLSClusteringS2S MTEB CLSClusteringS2S default test None type value v_measure 38.39863566717934 task dataset metrics type Reranking type name config split revision C-MTEB/CMedQAv1-reranking MTEB CMedQAv1 default test None type value map 83.63680381780644 type value mrr 86.16476190476192 task dataset metrics type Reranking type name config split revision C-MTEB/CMedQAv2-reranking MTEB CMedQAv2 default test None type value map 83.74350667859487 type value mrr 86.10388888888889 task dataset metrics type Retrieval type name config split revision C-MTEB/CmedqaRetrieval MTEB CmedqaRetrieval default dev None type value map_at_1 22.072 type value map_at_10 32.942 type value map_at_100 34.768 type value map_at_1000 34.902 type value map_at_3 29.357 type value map_at_5 31.236000000000004 type value mrr_at_1 34.259 type value mrr_at_10 41.957 type value mrr_at_100 42.982 type value mrr_at_1000 43.042 type value mrr_at_3 39.722 type value mrr_at_5 40.898 type value ndcg_at_1 34.259 type value ndcg_at_10 39.153 type value ndcg_at_100 46.493 type value ndcg_at_1000 49.01 type value ndcg_at_3 34.636 type value ndcg_at_5 36.278 type value precision_at_1 34.259 type value precision_at_10 8.815000000000001 type value precision_at_100 1.474 type value precision_at_1000 0.179 type value precision_at_3 19.73 type value precision_at_5 14.174000000000001 type value recall_at_1 22.072 type value recall_at_10 48.484 type value recall_at_100 79.035 type value recall_at_1000 96.15 type value recall_at_3 34.607 type value recall_at_5 40.064 task dataset metrics type PairClassification type name config split revision C-MTEB/CMNLI MTEB Cmnli default validation None type value cos_sim_accuracy 76.7047504509922 type value cos_sim_ap 85.26649874800871 type value cos_sim_f1 78.13528724646915 type value cos_sim_precision 71.57587548638132 type value cos_sim_recall 86.01823708206688 type value dot_accuracy 70.13830426939266 type value dot_ap 77.01510412382171 type value dot_f1 73.56710042713817 type value dot_precision 63.955094991364426 type value dot_recall 86.57937806873977 type value euclidean_accuracy 75.53818400481059 type value euclidean_ap 84.34668448241264 type value euclidean_f1 77.51741608613047 type value euclidean_precision 70.65614777756399 type value euclidean_recall 85.85457096095394 type value manhattan_accuracy 75.49007817197835 type value manhattan_ap 84.40297506704299 type value manhattan_f1 77.63185324160932 type value manhattan_precision 70.03949595636637 type value manhattan_recall 87.07037643207856 type value max_accuracy 76.7047504509922 type value max_ap 85.26649874800871 type value max_f1 78.13528724646915 task dataset metrics type Retrieval type name config split revision C-MTEB/CovidRetrieval MTEB CovidRetrieval default dev None type value map_at_1 69.178 type value map_at_10 77.523 type value map_at_100 77.793 type value map_at_1000 77.79899999999999 type value map_at_3 75.878 type value map_at_5 76.849 type value mrr_at_1 69.44200000000001 type value mrr_at_10 77.55 type value mrr_at_100 77.819 type value mrr_at_1000 77.826 type value mrr_at_3 75.957 type value mrr_at_5 76.916 type value ndcg_at_1 69.44200000000001 type value ndcg_at_10 81.217 type value ndcg_at_100 82.45 type value ndcg_at_1000 82.636 type value ndcg_at_3 77.931 type value ndcg_at_5 79.655 type value precision_at_1 69.44200000000001 type value precision_at_10 9.357 type value precision_at_100 0.993 type value precision_at_1000 0.101 type value precision_at_3 28.1 type value precision_at_5 17.724 type value recall_at_1 69.178 type value recall_at_10 92.624 type value recall_at_100 98.209 type value recall_at_1000 99.684 type value recall_at_3 83.772 type value recall_at_5 87.882 task dataset metrics type Retrieval type name config split revision C-MTEB/DuRetrieval MTEB DuRetrieval default dev None type value map_at_1 25.163999999999998 type value map_at_10 76.386 type value map_at_100 79.339 type value map_at_1000 79.39500000000001 type value map_at_3 52.959 type value map_at_5 66.59 type value mrr_at_1 87.9 type value mrr_at_10 91.682 type value mrr_at_100 91.747 type value mrr_at_1000 91.751 type value mrr_at_3 91.267 type value mrr_at_5 91.527 type value ndcg_at_1 87.9 type value ndcg_at_10 84.569 type value ndcg_at_100 87.83800000000001 type value ndcg_at_1000 88.322 type value ndcg_at_3 83.473 type value ndcg_at_5 82.178 type value precision_at_1 87.9 type value precision_at_10 40.605000000000004 type value precision_at_100 4.752 type value precision_at_1000 0.488 type value precision_at_3 74.9 type value precision_at_5 62.96000000000001 type value recall_at_1 25.163999999999998 type value recall_at_10 85.97399999999999 type value recall_at_100 96.63000000000001 type value recall_at_1000 99.016 type value recall_at_3 55.611999999999995 type value recall_at_5 71.936 task dataset metrics type Retrieval type name config split revision C-MTEB/EcomRetrieval MTEB EcomRetrieval default dev None type value map_at_1 48.6 type value map_at_10 58.831 type value map_at_100 59.427 type value map_at_1000 59.44199999999999 type value map_at_3 56.383 type value map_at_5 57.753 type value mrr_at_1 48.6 type value mrr_at_10 58.831 type value mrr_at_100 59.427 type value mrr_at_1000 59.44199999999999 type value mrr_at_3 56.383 type value mrr_at_5 57.753 type value ndcg_at_1 48.6 type value ndcg_at_10 63.951 type value ndcg_at_100 66.72200000000001 type value ndcg_at_1000 67.13900000000001 type value ndcg_at_3 58.882 type value ndcg_at_5 61.373 type value precision_at_1 48.6 type value precision_at_10 8.01 type value precision_at_100 0.928 type value precision_at_1000 0.096 type value precision_at_3 22.033 type value precision_at_5 14.44 type value recall_at_1 48.6 type value recall_at_10 80.10000000000001 type value recall_at_100 92.80000000000001 type value recall_at_1000 96.1 type value recall_at_3 66.10000000000001 type value recall_at_5 72.2 task dataset metrics type Classification type name config split revision C-MTEB/IFlyTek-classification MTEB IFlyTek default validation None type value accuracy 47.36437091188918 type value f1 36.60946954228577 task dataset metrics type Classification type name config split revision C-MTEB/JDReview-classification MTEB JDReview default test None type value accuracy 79.5684803001876 type value ap 42.671935929201524 type value f1 73.31912729103752 task dataset metrics type STS type name config split revision C-MTEB/LCQMC MTEB LCQMC default test None type value cos_sim_pearson 68.62670112113864 type value cos_sim_spearman 75.74009123170768 type value euclidean_pearson 73.93002595958237 type value euclidean_spearman 75.35222935003587 type value manhattan_pearson 73.89870445158144 type value manhattan_spearman 75.31714936339398 task dataset metrics type Reranking type name config split revision C-MTEB/Mmarco-reranking MTEB MMarcoReranking default dev None type value map 31.5372713650176 type value mrr 30.163095238095238 task dataset metrics type Retrieval type name config split revision C-MTEB/MMarcoRetrieval MTEB MMarcoRetrieval default dev None type value map_at_1 65.054 type value map_at_10 74.156 type value map_at_100 74.523 type value map_at_1000 74.535 type value map_at_3 72.269 type value map_at_5 73.41 type value mrr_at_1 67.24900000000001 type value mrr_at_10 74.78399999999999 type value mrr_at_100 75.107 type value mrr_at_1000 75.117 type value mrr_at_3 73.13499999999999 type value mrr_at_5 74.13499999999999 type value ndcg_at_1 67.24900000000001 type value ndcg_at_10 77.96300000000001 type value ndcg_at_100 79.584 type value ndcg_at_1000 79.884 type value ndcg_at_3 74.342 type value ndcg_at_5 76.278 type value precision_at_1 67.24900000000001 type value precision_at_10 9.466 type value precision_at_100 1.027 type value precision_at_1000 0.105 type value precision_at_3 27.955999999999996 type value precision_at_5 17.817 type value recall_at_1 65.054 type value recall_at_10 89.113 type value recall_at_100 96.369 type value recall_at_1000 98.714 type value recall_at_3 79.45400000000001 type value recall_at_5 84.06 task dataset metrics type Classification type name config split revision mteb/amazon_massive_intent MTEB MassiveIntentClassification (zh-CN) zh-CN test 31efe3c427b0bae9c22cbb560b8f15491cc6bed7 type value accuracy 68.1977135171486 type value f1 67.23114308718404 task dataset metrics type Classification type name config split revision mteb/amazon_massive_scenario MTEB MassiveScenarioClassification (zh-CN) zh-CN test 7d571f92784cd94a019292a1f45445077d0ef634 type value accuracy 71.92669804976462 type value f1 72.90628475628779 task dataset metrics type Retrieval type name config split revision C-MTEB/MedicalRetrieval MTEB MedicalRetrieval default dev None type value map_at_1 49.2 type value map_at_10 54.539 type value map_at_100 55.135 type value map_at_1000 55.19199999999999 type value map_at_3 53.383 type value map_at_5 54.142999999999994 type value mrr_at_1 49.2 type value mrr_at_10 54.539 type value mrr_at_100 55.135999999999996 type value mrr_at_1000 55.19199999999999 type value mrr_at_3 53.383 type value mrr_at_5 54.142999999999994 type value ndcg_at_1 49.2 type value ndcg_at_10 57.123000000000005 type value ndcg_at_100 60.21300000000001 type value ndcg_at_1000 61.915 type value ndcg_at_3 54.772 type value ndcg_at_5 56.157999999999994 type value precision_at_1 49.2 type value precision_at_10 6.52 type value precision_at_100 0.8009999999999999 type value precision_at_1000 0.094 type value precision_at_3 19.6 type value precision_at_5 12.44 type value recall_at_1 49.2 type value recall_at_10 65.2 type value recall_at_100 80.10000000000001 type value recall_at_1000 93.89999999999999 type value recall_at_3 58.8 type value recall_at_5 62.2 task dataset metrics type Classification type name config split revision C-MTEB/MultilingualSentiment-classification MTEB MultilingualSentiment default validation None type value accuracy 63.29333333333334 type value f1 63.03293854259612 task dataset metrics type PairClassification type name config split revision C-MTEB/OCNLI MTEB Ocnli default validation None type value cos_sim_accuracy 75.69030860855442 type value cos_sim_ap 80.6157833772759 type value cos_sim_f1 77.87524366471735 type value cos_sim_precision 72.3076923076923 type value cos_sim_recall 84.37170010559663 type value dot_accuracy 67.78559826746074 type value dot_ap 72.00871467527499 type value dot_f1 72.58722247394654 type value dot_precision 63.57142857142857 type value dot_recall 84.58289334741288 type value euclidean_accuracy 75.20303194369248 type value euclidean_ap 80.98587256415605 type value euclidean_f1 77.26396917148362 type value euclidean_precision 71.03631532329496 type value euclidean_recall 84.68848996832101 type value manhattan_accuracy 75.20303194369248 type value manhattan_ap 80.93460699513219 type value manhattan_f1 77.124773960217 type value manhattan_precision 67.43083003952569 type value manhattan_recall 90.07391763463569 type value max_accuracy 75.69030860855442 type value max_ap 80.98587256415605 type value max_f1 77.87524366471735 task dataset metrics type Classification type name config split revision C-MTEB/OnlineShopping-classification MTEB OnlineShopping default test None type value accuracy 87.00000000000001 type value ap 83.24372135949511 type value f1 86.95554191530607 task dataset metrics type STS type name config split revision C-MTEB/PAWSX MTEB PAWSX default test None type value cos_sim_pearson 37.57616811591219 type value cos_sim_spearman 41.490259084930045 type value euclidean_pearson 38.9155043692188 type value euclidean_spearman 39.16056534305623 type value manhattan_pearson 38.76569892264335 type value manhattan_spearman 38.99891685590743 task dataset metrics type STS type name config split revision C-MTEB/QBQTC MTEB QBQTC default test None type value cos_sim_pearson 35.44858610359665 type value cos_sim_spearman 38.11128146262466 type value euclidean_pearson 31.928644189822457 type value euclidean_spearman 34.384936631696554 type value manhattan_pearson 31.90586687414376 type value manhattan_spearman 34.35770153777186 task dataset metrics type STS type name config split revision mteb/sts22-crosslingual-sts MTEB STS22 (zh) zh test 6d1ba47164174a496b7fa5d3569dae26a6813b80 type value cos_sim_pearson 66.54931957553592 type value cos_sim_spearman 69.25068863016632 type value euclidean_pearson 50.26525596106869 type value euclidean_spearman 63.83352741910006 type value manhattan_pearson 49.98798282198196 type value manhattan_spearman 63.87649521907841 task dataset metrics type STS type name config split revision C-MTEB/STSB MTEB STSB default test None type value cos_sim_pearson 82.52782476625825 type value cos_sim_spearman 82.55618986168398 type value euclidean_pearson 78.48190631687673 type value euclidean_spearman 78.39479731354655 type value manhattan_pearson 78.51176592165885 type value manhattan_spearman 78.42363787303265 task dataset metrics type Reranking type name config split revision C-MTEB/T2Reranking MTEB T2Reranking default dev None type value map 67.36693873615643 type value mrr 77.83847701797939 task dataset metrics type Retrieval type name config split revision C-MTEB/T2Retrieval MTEB T2Retrieval default dev None type value map_at_1 25.795 type value map_at_10 72.258 type value map_at_100 76.049 type value map_at_1000 76.134 type value map_at_3 50.697 type value map_at_5 62.324999999999996 type value mrr_at_1 86.634 type value mrr_at_10 89.792 type value mrr_at_100 89.91900000000001 type value mrr_at_1000 89.923 type value mrr_at_3 89.224 type value mrr_at_5 89.608 type value ndcg_at_1 86.634 type value ndcg_at_10 80.589 type value ndcg_at_100 84.812 type value ndcg_at_1000 85.662 type value ndcg_at_3 82.169 type value ndcg_at_5 80.619 type value precision_at_1 86.634 type value precision_at_10 40.389 type value precision_at_100 4.93 type value precision_at_1000 0.513 type value precision_at_3 72.104 type value precision_at_5 60.425 type value recall_at_1 25.795 type value recall_at_10 79.565 type value recall_at_100 93.24799999999999 type value recall_at_1000 97.595 type value recall_at_3 52.583999999999996 type value recall_at_5 66.175 task dataset metrics type Classification type name config split revision C-MTEB/TNews-classification MTEB TNews default validation None type value accuracy 47.648999999999994 type value f1 46.28925837008413 task dataset metrics type Clustering type name config split revision C-MTEB/ThuNewsClusteringP2P MTEB ThuNewsClusteringP2P default test None type value v_measure 54.07641891287953 task dataset metrics type Clustering type name config split revision C-MTEB/ThuNewsClusteringS2S MTEB ThuNewsClusteringS2S default test None type value v_measure 53.423702062353954 task dataset metrics type Retrieval type name config split revision C-MTEB/VideoRetrieval MTEB VideoRetrieval default dev None type value map_at_1 55.7 type value map_at_10 65.923 type value map_at_100 66.42 type value map_at_1000 66.431 type value map_at_3 63.9 type value map_at_5 65.225 type value mrr_at_1 55.60000000000001 type value mrr_at_10 65.873 type value mrr_at_100 66.36999999999999 type value mrr_at_1000 66.381 type value mrr_at_3 63.849999999999994 type value mrr_at_5 65.17500000000001 type value ndcg_at_1 55.7 type value ndcg_at_10 70.621 type value ndcg_at_100 72.944 type value ndcg_at_1000 73.25399999999999 type value ndcg_at_3 66.547 type value ndcg_at_5 68.93599999999999 type value precision_at_1 55.7 type value precision_at_10 8.52 type value precision_at_100 0.958 type value precision_at_1000 0.098 type value precision_at_3 24.733 type value precision_at_5 16 type value recall_at_1 55.7 type value recall_at_10 85.2 type value recall_at_100 95.8 type value recall_at_1000 98.3 type value recall_at_3 74.2 type value recall_at_5 80 task dataset metrics type Classification type name config split revision C-MTEB/waimai-classification MTEB Waimai default test None type value accuracy 84.54 type value ap 66.13603199670062 type value f1 82.61420654584116

The text embedding set trained by Jina AI.

Quick Start

The easiest way to starting using jina-embeddings-v2-base-zh is to use Jina AI's Embedding API.

Intended Usage & Model Info

jina-embeddings-v2-base-zh is a Chinese/English bilingual text 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. We have designed it for high performance in mono-lingual & cross-lingual applications and trained it specifically to support mixed Chinese-English input without bias. Additionally, we provide the following embedding models:

jina-embeddings-v2-base-zh 是支持中英双语的文本向量模型，它支持长达8192字符的文本编码。 该模型的研发基于BERT架构(JinaBERT)，JinaBERT是在BERT架构基础上的改进，首次将ALiBi应用到编码器架构中以支持更长的序列。 不同于以往的单语言/多语言向量模型，我们设计双语模型来更好的支持单语言（中搜中）以及跨语言（中搜英）文档检索。 除此之外，我们也提供其它向量模型:

• jina-embeddings-v2-small-en: 33 million parameters.
• jina-embeddings-v2-base-en: 137 million parameters.
• jina-embeddings-v2-base-zh: 161 million parameters Chinese-English Bilingual embeddings (you are here).
• jina-embeddings-v2-base-de: 161 million parameters German-English Bilingual embeddings.
• jina-embeddings-v2-base-es: Spanish-English Bilingual embeddings (soon).
• jina-embeddings-v2-base-code: 161 million parameters code embeddings.

Data & Parameters

The data and training details are described in this 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?', '今天天气怎么样?']

tokenizer = AutoTokenizer.from_pretrained('jinaai/jina-embeddings-v2-base-zh')
model = AutoModel.from_pretrained('jinaai/jina-embeddings-v2-base-zh', trust_remote_code=True, torch_dtype=torch.bfloat16)

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
import torch
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-base-zh', trust_remote_code=True, torch_dtype=torch.bfloat16)
embeddings = model.encode(['How is the weather today?', '今天天气怎么样?'])
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
)

If you want to use the model together with the sentence-transformers package, make sure that you have installed the latest release and set trust_remote_code=True as well:

!pip install -U sentence-transformers
from sentence_transformers import SentenceTransformer
from numpy.linalg import norm

cos_sim = lambda a,b: (a @ b.T) / (norm(a)*norm(b))
model = SentenceTransformer('jinaai/jina-embeddings-v2-base-zh', trust_remote_code=True)
embeddings = model.encode(['How is the weather today?', '今天天气怎么样?'])
print(cos_sim(embeddings[0], embeddings[1]))

Using the its latest release (v2.3.0) sentence-transformers also supports Jina embeddings (Please make sure that you are logged into huggingface as well):

!pip install -U sentence-transformers
from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim

model = SentenceTransformer(
    "jinaai/jina-embeddings-v2-base-zh", # switch to en/zh for English or Chinese
    trust_remote_code=True
)

# control your input sequence length up to 8192
model.max_seq_length = 1024

embeddings = model.encode([
    'How is the weather today?',
    '今天天气怎么样?'
])
print(cos_sim(embeddings[0], embeddings[1]))

Alternatives to Using Transformers Package

1. Managed SaaS: Get started with a free key on Jina AI's Embedding API.
2. Private and high-performance deployment: Get started by picking from our suite of models and deploy them on AWS Sagemaker.

Use Jina Embeddings for RAG

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.

Trouble Shooting

Loading of Model Code failed

If you forgot to pass the trust_remote_code=True flag when calling AutoModel.from_pretrained or initializing the model via the SentenceTransformer class, you will receive an error that the model weights could not be initialized. This is caused by tranformers falling back to creating a default BERT model, instead of a jina-embedding model:

Some weights of the model checkpoint at jinaai/jina-embeddings-v2-base-zh were not used when initializing BertModel: ['encoder.layer.2.mlp.layernorm.weight', 'encoder.layer.3.mlp.layernorm.weight', 'encoder.layer.10.mlp.wo.bias', 'encoder.layer.5.mlp.wo.bias', 'encoder.layer.2.mlp.layernorm.bias', 'encoder.layer.1.mlp.gated_layers.weight', 'encoder.layer.5.mlp.gated_layers.weight', 'encoder.layer.8.mlp.layernorm.bias', ...

User is not logged into Huggingface

The model is only availabe under gated access. This means you need to be logged into huggingface load load it. If you receive the following error, you need to provide an access token, either by using the huggingface-cli or providing the token via an environment variable as described above:

OSError: jinaai/jina-embeddings-v2-base-zh is not a local folder and is not a valid model identifier listed on 'https://huggingface.co/models'
If this is a private repository, make sure to pass a token having permission to this repo with `use_auth_token` or log in with `huggingface-cli login` and pass `use_auth_token=True`.

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:

@article{mohr2024multi,
  title={Multi-Task Contrastive Learning for 8192-Token Bilingual Text Embeddings},
  author={Mohr, Isabelle and Krimmel, Markus and Sturua, Saba and Akram, Mohammad Kalim and Koukounas, Andreas and G{\"u}nther, Michael and Mastrapas, Georgios and Ravishankar, Vinit and Mart{\'\i}nez, Joan Fontanals and Wang, Feng and others},
  journal={arXiv preprint arXiv:2402.17016},
  year={2024}
}