Downloads · 30 days
11
1% of all-time downloads
bert-from-clip-chinese-1M
YeungNLP/bert-from-clip-chinese-1M
bert-from-clip-chinese-1M is a feature extraction model from YeungNLP. Use it when you need embeddings to search or compare text. It is set up for transformers.
该预训练权重是从BertCLIP预训练模型中单独取出来的Bert的权重，BertCLIP是中文版本的CLIP预训练模型，基于LiT-tuning（Locked-image Text tuning）的策略， 使用140万中文图文对数据进行多模态对比学习预训练。
All-time downloads
742
Public
Repo size
818 MB
Likes
5
Public
Repo makeup
Click a slice to open those files.
.bin409 MB · 100%
At a glance
- Task
- Feature Extraction
- Library
- transformers
- Model type
- bert
- Access
- Public
- Created
- Dec 3, 2022
- Updated
- Dec 5, 2022
- SHA
- af92b3d6
Model card
From the Hugging Face model README
bert-from-clip-chinese pretrained model
该预训练权重是从BertCLIP预训练模型中单独取出来的Bert的权重，BertCLIP是中文版本的CLIP预训练模型，基于LiT-tuning（Locked-image Text tuning）的策略， 使用140万中文图文对数据进行多模态对比学习预训练。
Github: CLIP-Chinese
Bolg: CLIP-Chinese：中文多模态对比学习CLIP预训练模型
Usage
可以直接使用Huggingface的BertModel直接加载该预训练权重，进行下游的任务。
from transformers import BertTokenizer, BertModel
model_name_or_path = 'YeungNLP/bert-from-clip-chinese-1M'
tokenizer = BertTokenizer.from_pretrained(model_name_or_path)
model = BertModel.from_pretrained(model_name_or_path)
- Task
- Feature Extraction
- Library
- transformers
- Type
- bert
- Created
- Dec 3, 2022
- Updated
- Dec 5, 2022