You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
imyzx2017 26b49a43b3 add readme 2 years ago
src 2021-pangu_alpha_KD-V0 2 years ago
FFN_tiny_pangu_generalKD.py 2021-pangu_alpha_KD-V0 2 years ago
MiniLM_pangu.py 2021-pangu_alpha_KD-V0 2 years ago
dataset_restore_data0.py 2021-pangu_alpha_KD-V0 2 years ago
jieba-0.42.1.tar.gz 2021-pangu_alpha_KD-V0 2 years ago
ma-pre-start.sh 2021-pangu_alpha_KD-V0 2 years ago
readme.md add readme 2 years ago
sentencepiece-0.1.94-cp37-cp37m-linux_aarch64.whl 2021-pangu_alpha_KD-V0 2 years ago
tinyPangu_stage1_pretraingStu.py 2021-pangu_alpha_KD-V0 2 years ago
tiny_pangu.py 2021-pangu_alpha_KD-V0 2 years ago
tiny_pangu_KD_finetune.py 2021-pangu_alpha_KD-V0 2 years ago
tiny_pangu_KD_withCrossEntropy.py 2021-pangu_alpha_KD-V0 2 years ago
tokenization_jieba.py 2021-pangu_alpha_KD-V0 2 years ago
utils_fix.py 2021-pangu_alpha_KD-V0 2 years ago

支持在鹏城云脑2或者其他NPU服务器上,进行盘古大模型的知识蒸馏,目前实验代码为2.6B盘古模型蒸馏出368M

Python Shell

Contributors (1)