FlagEmbedding

Map any text to a low-dimensional dense vector which can be used for tasks like retrieval, classification, clustering, or semantic search. It also can be used in vector database for LLMs.

Updates

08/02/2023: Release baai-general-embedding-* Models, ranking 1st in MTEB and C-MTEB respectively !
08/01/2023: We release the Chinese Massive Text Embedding Benchmark (C-MTEB), consisting of 31 test dataset.

Model List

Model	Language	Description	query instruction for retrieval
BAAI/baai-general-embedding-large-en-instruction	English	rank 1st in MTEB leaderboard	`Represent this sentence for searching relevant passages:`
BAAI/baai-general-embedding-large-zh-instruction	Chinese	rank 1st in C-MTEB bechmark	`为这个句子生成表示以用于检索相关文章：`
BAAI/baai-general-embedding-large-zh	Chinese	rank 2nd in C-MTEB bechmark	--

Usage

Sentence-Transformers

Using this model becomes easy when you have sentence-transformers installed:

pip install -U sentence-transformers

Then you can use the model like this:

from sentence_transformers import SentenceTransformer
sentences = ["样例数据-1", "样例数据-2"]
model = SentenceTransformer('BAAI/baai-general-embedding-large-zh-instruction')
embeddings = model.encode(sentences, normalize_embeddings=True)
print(embeddings)

#For retrieval task, when you use the model whose name ends with `-instruction`
#each query should start with a instruction. 
queries = ["手机开不了机怎么办？"]
passages = ["样例段落-1", "样例段落-2"]
instruction = "为这个句子生成表示以用于检索相关文章："
model = SentenceTransformer('BAAI/baai-general-embedding-large-zh-instruction')
q_embeddings = model.encode([instruction+q for q in queries], normalize_embeddings=True)
p_embeddings = model.encode(passages, normalize_embeddings=True)
scores = q_embeddings @ p_embeddings.T

HuggingFace Transformers

With transformers package, you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.

from transformers import AutoTokenizer, AutoModel
import torch
# Sentences we want sentence embeddings for
sentences = ["样例数据-1", "样例数据-2"]
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('BAAI/baai-general-embedding-large-zh-instruction')
model = AutoModel.from_pretrained('BAAI/baai-general-embedding-large-zh-instruction')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)
    # Perform pooling. In this case, cls pooling.
    sentence_embeddings = model_output[0][:, 0]
# normalize embeddings
sentence_embeddings = torch.nn.functional.normalize(sentence_embeddings, p=2, dim=1)
print("Sentence embeddings:")
print(sentence_embeddings)

Evaluation Results

MTEB:

Model Name	Model Size (GB)	Dimension	Sequence Length	Average (56)	Retrieval (15)	Clustering (11)	Pair Classification (3)	Reranking (4)	STS (10)	Summarization (1)	Classification (12)
baai-general-embedding-large-en-instruction	0.67	1024	512	63.34	53.23	48.47	86.34	59.87	81.89	30.55	72.28
gte-large	0.67	1024	512	63.13	52.22	46.84	85.00	59.13	83.35	31.66	73.33
gte-base	0.22	768	512	62.39	51.14	46.2	84.57	58.61	82.3	31.17	73.01
e5-large-v2	1.34	1024	512	62.25	50.56	44.49	86.03	56.61	82.05	30.19	75.24
instructor-xl	4.96	768	512	61.79	49.26	44.74	86.62	57.29	83.06	32.32	61.79
e5-base-v2	0.44	768	512	61.5	50.29	43.80	85.73	55.91	81.05	30.28	73.84
gte-small	0.07	384	512	61.36	49.46	44.89	83.54	57.7	82.07	30.42	72.31
text-embedding-ada-002	-	1536	8192	60.99	49.25	45.9	84.89	56.32	80.97	30.8	70.93
e5-small-v2	0.13	384	512	59.93	49.04	39.92	84.67	54.32	80.39	31.16	72.94
sentence-t5-xxl	9.73	768	512	59.51	42.24	43.72	85.06	56.42	82.63	30.08	73.42
all-mpnet-base-v2	0.44	768	514	57.78	43.81	43.69	83.04	59.36	80.28	27.49	65.07
sgpt-bloom-7b1-msmarco	28.27	4096	2048	57.59	48.22	38.93	81.9	55.65	77.74	33.6	66.19
all-MiniLM-L12-v2	0.13	384	512	56.53	42.69	41.81	82.41	58.44	79.8	27.9	63.21
all-MiniLM-L6-v2	0.09	384	512	56.26	41.95	42.35	82.37	58.04	78.9	30.81	63.05
contriever-base-msmarco	0.44	768	512	56.00	41.88	41.1	82.54	53.14	76.51	30.36	66.68
sentence-t5-base	0.22	768	512	55.27	33.63	40.21	85.18	53.09	81.14	31.39	69.81

C-MTEB:
We create a benchmark C-MTEB for chinese text embedding which consists of 31 datasets from 6 tasks. More details and evaluation scripts see evaluation.

Model	Embedding dimension	Avg	Retrieval	STS	PairClassification	Classification	Reranking	Clustering
baai-general-embedding-large-zh-instruction	1024	64.20	71.53	53.23	78.94	72.26	65.11	48.39
baai-general-embedding-large-zh	1024	63.53	70.55	50.98	76.77	72.49	64.91	50.01
m3e-base	768	57.10	56.91	48.15	63.99	70.28	59.34	47.68
m3e-large	1024	57.05	54.75	48.64	64.3	71.22	59.66	48.88
text-embedding-ada-002(OpenAI)	1536	53.02	52.0	40.61	69.56	67.38	54.28	45.68
luotuo	1024	49.37	44.4	39.41	66.62	65.29	49.25	44.39
text2vec	768	47.63	38.79	41.71	67.41	65.18	49.45	37.66
text2vec-large	1024	47.36	41.94	41.98	70.86	63.42	49.16	30.02

Train

This section will introduce the way we used to train the general embedding. The training scripts are in universal_embedding, and we provide some examples to do pre-train and fine-tune.

1. RetroMAE Pre-train
We pre-train the model following the method retromae, which shows promising improvement in retrieval task (see https://aclanthology.org/2022.emnlp-main.35.pdf). The pre-training was conducted on 24 A100(40G) GPUs with a batch size of 720. In retromae, the mask ratio of encoder and decoder are 0.3, 0.5 respectively. We used the AdamW optimizer and the learning rate is 2e-5.

Pre-training data:

English:
- Pile
- wikipedia
- msmarco
Chinese:
- Subset of wudao
- baidu-baike

2. Finetune
We fine-tune the model using a contrastive objective. The format of input data is a triple(query, positive, negative). Besides the negative in the triple, we also adopt in-batch negatives strategy.

We trained our model on 48 A100(40G) GPUs with a large batch size of 32,768. We used the AdamW optimizer and the learning rate is 1e-5. The sequence length was limited to 128 tokens. The temperature for contrastive loss is 0.01.

For the version with *-instrcution, we add instruction to the query for retrieval task in the training. For english, the instruction is Represent this sentence for searching relevant passages: ; For chinese, the instruction is 为这个句子生成表示以用于检索相关文章：. In the evaluation, the instruction should be added for sentence to passages retrieval task, not be added for other tasks.

The finetune script is accessible in this repository: universal_embedding. You can easily finetune your model with it.

Training data:

For English, we collect 230M text pairs from wikipedia, cc-net, and so on.
For chinese, we collect 120M text pairs from wudao, zhihu, news websites and so on.

The data collection is to be released in the future.

10 KiB Raw Blame History