gte-Qwen2-7B-instruct is the latest model in the gte (General Text Embedding) model family that ranks No.1 in both English and Chinese evaluations on the Massive Text Embedding Benchmark MTEB benchmark (as of June 16, 2024).

Recently, the Qwen team released the Qwen2 series models, and we have trained the gte-Qwen2-7B-instruct model based on the Qwen2-7B LLM model. Compared to the gte-Qwen1.5-7B-instruct model, the gte-Qwen2-7B-instruct model uses the same training data and training strategies during the finetuning stage, with the only difference being the upgraded base model to Qwen2-7B. Considering the improvements in the Qwen2 series models compared to the Qwen1.5 series, we can also expect consistent performance enhancements in the embedding models.

The model incorporates several key advancements:

• Integration of bidirectional attention mechanisms, enriching its contextual understanding.
• Instruction tuning, applied solely on the query side for streamlined efficiency
• Comprehensive training across a vast, multilingual text corpus spanning diverse domains and scenarios. This training leverages both weakly supervised and supervised data, ensuring the model's applicability across numerous languages and a wide array of downstream tasks.

• Model Size: 7B
• Embedding Dimension: 3584
• Max Input Tokens: 32k

transformers>=4.39.2
flash_attn>=2.5.6

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("Alibaba-NLP/gte-Qwen2-7B-instruct", trust_remote_code=True)
# In case you want to reduce the maximum length:
model.max_seq_length = 8192

queries = [
    "how much protein should a female eat",
    "summit define",
]
documents = [
    "As a general guideline, the CDC's average requirement of protein for women ages 19 to 70 is 46 grams per day. But, as you can see from this chart, you'll need to increase that if you're expecting or training for a marathon. Check out the chart below to see how much protein you should be eating each day.",
    "Definition of summit for English Language Learners. : 1  the highest point of a mountain : the top of a mountain. : 2  the highest level. : 3  a meeting or series of meetings between the leaders of two or more governments.",
]

query_embeddings = model.encode(queries, prompt_name="query")
document_embeddings = model.encode(documents)

scores = (query_embeddings @ document_embeddings.T) * 100
print(scores.tolist())

Observe the config_sentence_transformers.json to see all pre-built prompt names. Otherwise, you can use model.encode(queries, prompt="Instruct: ...\nQuery: " to use a custom prompt of your choice.

import torch
import torch.nn.functional as F

from torch import Tensor
from transformers import AutoTokenizer, AutoModel


def last_token_pool(last_hidden_states: Tensor,
                 attention_mask: Tensor) -> Tensor:
    left_padding = (attention_mask[:, -1].sum() == attention_mask.shape[0])
    if left_padding:
        return last_hidden_states[:, -1]
    else:
        sequence_lengths = attention_mask.sum(dim=1) - 1
        batch_size = last_hidden_states.shape[0]
        return last_hidden_states[torch.arange(batch_size, device=last_hidden_states.device), sequence_lengths]


def get_detailed_instruct(task_description: str, query: str) -> str:
    return f'Instruct: {task_description}\nQuery: {query}'


# Each query must come with a one-sentence instruction that describes the task
task = 'Given a web search query, retrieve relevant passages that answer the query'
queries = [
    get_detailed_instruct(task, 'how much protein should a female eat'),
    get_detailed_instruct(task, 'summit define')
]
# No need to add instruction for retrieval documents
documents = [
    "As a general guideline, the CDC's average requirement of protein for women ages 19 to 70 is 46 grams per day. But, as you can see from this chart, you'll need to increase that if you're expecting or training for a marathon. Check out the chart below to see how much protein you should be eating each day.",
    "Definition of summit for English Language Learners. : 1  the highest point of a mountain : the top of a mountain. : 2  the highest level. : 3  a meeting or series of meetings between the leaders of two or more governments."
]
input_texts = queries + documents

tokenizer = AutoTokenizer.from_pretrained('Alibaba-NLP/gte-Qwen2-7B-instruct', trust_remote_code=True)
model = AutoModel.from_pretrained('Alibaba-NLP/gte-Qwen2-7B-instruct', trust_remote_code=True)

max_length = 8192

# Tokenize the input texts
batch_dict = tokenizer(input_texts, max_length=max_length, padding=True, truncation=True, return_tensors='pt')
outputs = model(**batch_dict)
embeddings = last_token_pool(outputs.last_hidden_state, batch_dict['attention_mask'])

# normalize embeddings
embeddings = F.normalize(embeddings, p=2, dim=1)
scores = (embeddings[:2] @ embeddings[2:].T) * 100
print(scores.tolist())

Usage via infinity, a MIT Licensed inference server.

# requires ~16-32GB VRAM NVIDIA Compute Capability >= 8.0
docker run \
-v $PWD/data:/app/.cache --gpus "0" -p "7997":"7997" \
michaelf34/infinity:0.0.68-trt-onnx \
v2 --model-id Alibaba-NLP/gte-Qwen2-7B-instruct --revision "refs/pr/38" --dtype bfloat16 --batch-size 8 --device cuda --engine torch --port 7997 --no-bettertransformer

You can use the scripts/eval_mteb.py to reproduce the following result of gte-Qwen2-7B-instruct on MTEB(English)/C-MTEB(Chinese):

Model Name	MTEB(56)	C-MTEB(35)	MTEB-fr(26)	MTEB-pl(26)
bge-base-en-1.5	64.23	-	-	-
bge-large-en-1.5	63.55	-	-	-
gte-large-en-v1.5	65.39	-	-	-
gte-base-en-v1.5	64.11	-	-	-
mxbai-embed-large-v1	64.68	-	-	-
acge_text_embedding	-	69.07	-	-
stella-mrl-large-zh-v3.5-1792d	-	68.55	-	-
gte-large-zh	-	66.72	-	-
multilingual-e5-base	59.45	56.21	-	-
multilingual-e5-large	61.50	58.81	-	-
e5-mistral-7b-instruct	66.63	60.81	-	-
gte-Qwen1.5-7B-instruct	67.34	69.52	-	-
NV-Embed-v1	69.32	-	-	-
gte-Qwen2-7B-instruct	70.24	72.05	68.25	67.86
gte-Qwen2-1.5B-instruc(https://huggingface.co/Alibaba-NLP/gte-Qwen2-1.5B-instruct)	67.16	67.65	66.60	64.04

The gte series models have consistently released two types of models: encoder-only models (based on the BERT architecture) and decode-only models (based on the LLM architecture).

Models	Language	Max Sequence Length	Dimension	Model Size (Memory Usage, fp32)
GTE-large-zh	Chinese	512	1024	1.25GB
GTE-base-zh	Chinese	512	512	0.41GB
GTE-small-zh	Chinese	512	512	0.12GB
GTE-large	English	512	1024	1.25GB
GTE-base	English	512	512	0.21GB
GTE-small	English	512	384	0.10GB
GTE-large-en-v1.5	English	8192	1024	1.74GB
GTE-base-en-v1.5	English	8192	768	0.51GB
GTE-Qwen1.5-7B-instruct	Multilingual	32000	4096	26.45GB
GTE-Qwen2-7B-instruct	Multilingual	32000	3584	26.45GB
GTE-Qwen2-1.5B-instruct	Multilingual	32000	1536	6.62GB

In addition to the open-source GTE series models, GTE series models are also available as commercial API services on Alibaba Cloud.

• Embedding Models: Three versions of the text embedding models are available: text-embedding-v1/v2/v3, with v3 being the latest API service.
• ReRank Models: The gte-rerank model service is available.

Note that the models behind the commercial APIs are not entirely identical to the open-source models.

GTE models can be fine-tuned with a third party framework SWIFT.

pip install ms-swift -U

# check: https://swift.readthedocs.io/en/latest/BestPractices/Embedding.html
nproc_per_node=8
NPROC_PER_NODE=$nproc_per_node \
USE_HF=1 \
swift sft \
    --model Alibaba-NLP/gte-Qwen2-7B-instruct \
    --train_type lora \
    --dataset 'sentence-transformers/stsb' \
    --torch_dtype bfloat16 \
    --num_train_epochs 10 \
    --per_device_train_batch_size 2 \
    --per_device_eval_batch_size 1 \
    --gradient_accumulation_steps $(expr 64 / $nproc_per_node) \
    --eval_steps 100 \
    --save_steps 100 \
    --eval_strategy steps \
    --use_chat_template false \
    --save_total_limit 5 \
    --logging_steps 5 \
    --output_dir output \
    --warmup_ratio 0.05 \
    --learning_rate 5e-6 \
    --deepspeed zero3 \
    --dataloader_num_workers 4 \
    --task_type embedding \
    --loss_type cosine_similarity \
    --dataloader_drop_last true

If you find our paper or models helpful, please consider cite:

@article{li2023towards,
  title={Towards general text embeddings with multi-stage contrastive learning},
  author={Li, Zehan and Zhang, Xin and Zhang, Yanzhao and Long, Dingkun and Xie, Pengjun and Zhang, Meishan},
  journal={arXiv preprint arXiv:2308.03281},
  year={2023}
}