Downloads · 30 days
19
1% of all-time downloads
YeungNLP/bloom-2b6-zh
bloom-2b6-zh is a text generation model from YeungNLP. Use it when you need the model to write or continue text. It is set up for transformers.
Downloads · 30 days
19
1% of all-time downloads
All-time downloads
1.4K
Public
Repo size
10.4 GB
Likes
5
Public
Click a slice to open those files.
.bin5.2 GB · 100%
From the Hugging Face model README
项目地址:LLMPruner:大语言模型裁剪工具
LLMPruner是一个大语言模型裁剪工具,通过对大语言模型的冗余词表进行裁剪,减少模型参数量,降低显存占用,提升训练速度,并且能够保留预训练中学习到的知识。
本项目对Bloom进行词表裁剪,保留中文token和常用的英文token,词表由250880将至46145,缩减为原来的18.39%。裁剪得到的Bloom模型如下表:
使用方法:
from transformers import BloomTokenizerFast, BloomForCausalLM
tokenizer = BloomTokenizerFast.from_pretrained('YeungNLP/bloom-1b4-zh')
model = BloomForCausalLM.from_pretrained('YeungNLP/bloom-1b4-zh')
print(tokenizer.batch_decode(model.generate(tokenizer.encode('长风破浪会有时', return_tensors='pt'))))