gte-en-mlm-base

We introduce GTE-v1.5 series, new generalized text encoder, embedding and reranking models that the context length of up to 8192. The models are built upon the transformer++ encoder backbone (BERT + RoPE + GLU, code refer to Alibaba-NLP/new-impl) as well as the vocabulary of bert-base-uncased.

This text encoder is the GTEv1.5-en-MLM-base-8192 in table 13 of our paper.

Model list

Models Language Model Size Max Seq. Length GLUE XTREME-R
gte-multilingual-mlm-base Multiple 306M 8192 83.47 64.44
gte-en-mlm-base English - 8192 85.61 -
gte-en-mlm-large English - 8192 87.58 -

Training Details

Training Data

  • Masked language modeling (MLM): c4-en

Training Procedure

To enable the backbone model to support a context length of 8192, we adopted a multi-stage training strategy. The model first undergoes preliminary MLM pre-training on shorter lengths. And then, we resample the data, reducing the proportion of short texts, and continue the MLM pre-training.

The entire training process is as follows:

  • MLM-2048: lr 5e-4, mlm_probability 0.3, batch_size 4096, num_steps 70000, rope_base 10000
  • MLM-8192: lr 5e-5, mlm_probability 0.3, batch_size 1024, num_steps 20000, rope_base 500000

Evaluation

Models Language Model Size Max Seq. Length GLUE XTREME-R
gte-multilingual-mlm-base Multiple 306M 8192 83.47 64.44
gte-en-mlm-base English 137M 8192 85.61 -
gte-en-mlm-large English 435M 8192 87.58 -
MosaicBERT-base English 137M 128 85.4 -
MosaicBERT-base-2048 English 137M 2048 85 -
JinaBERT-base English 137M 512 85 -
nomic-bert-2048 English 137M 2048 84 -
MosaicBERT-large English 434M 128 86.1 -
JinaBERT-large English 434M 512 83.7 -
XLM-R-base Multiple 279M 512 80.44 62.02
RoBERTa-base English 125M 512 86.4 -
RoBERTa-large English 355M 512 88.9 -

Citation

If you find our paper or models helpful, please consider citing them as follows:

@misc{zhang2024mgte,
  title={mGTE: Generalized Long-Context Text Representation and Reranking Models for Multilingual Text Retrieval}, 
  author={Xin Zhang and Yanzhao Zhang and Dingkun Long and Wen Xie and Ziqi Dai and Jialong Tang and Huan Lin and Baosong Yang and Pengjun Xie and Fei Huang and Meishan Zhang and Wenjie Li and Min Zhang},
  year={2024},
  eprint={2407.19669},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  url={https://arxiv.org/abs/2407.19669}, 
}
Downloads last month
1,357
Safetensors
Model size
137M params
Tensor type
F32
·
Inference Examples
Unable to determine this model's library. Check the docs .

Model tree for Alibaba-NLP/gte-en-mlm-base

Finetunes
1 model

Dataset used to train Alibaba-NLP/gte-en-mlm-base

Collection including Alibaba-NLP/gte-en-mlm-base