issues
search
bojone
/
SimCSE
SimCSE在中文任务上的简单实验
591
stars
83
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
关于准确率和F1值的一些困惑
#17
hfy-only
opened
8 months ago
1
tensorflow.python.framework.errors_impl.InternalError: Blas xGEMMBatched launch failed : a.shape=[1536,64,64], b.shape=[1536,64,64], m=64, n=64, k=64, batch_size=1536
#16
yqstar
opened
1 year ago
0
词语位置对句子向量的影响
#15
XingyuZhu-Pamela
opened
2 years ago
0
评测相关系数的疑惑
#14
MingFL
opened
2 years ago
0
使用 tf.keras 训练报错
#13
Dcyx
opened
2 years ago
1
同一个batch中相同的句子一起输入模型，在这一次正向传播中 dropout是一样的吧?
#12
lihaixiang
opened
3 years ago
2
The pooling method for SIMCSE is not linear but tanh
#11
world2vec
opened
3 years ago
0
有关MLM+CL做无监督训练的问题
#10
joczu
opened
3 years ago
0
simcse loss的一些疑惑
#9
liujingqiao
opened
3 years ago
0
请问同一个batch经过dropout的不应该相同吗，为什么同一句子产生的embedding不同呢，困惑很久了
#8
zhongqiqianga
closed
3 years ago
1
simcse_loss中y_true的作用是什么，loss是根据什么定义的
#7
694344851
opened
3 years ago
0
在计算output pooling时不用mask吗？
#6
shaomai00
opened
3 years ago
1
请问损失函数该如何理解呢？
#5
FanWan
closed
3 years ago
0
怎么修改batch_size
#4
yuxuan2015
opened
3 years ago
1
batch_labels = np.zeros_like(batch_token_ids[:, :1]) ；batch_labels都为0，那batch_token_ids和dropout的batch_token_ids的相似度矩阵y_pred不应该为1吗
#3
lonngxiang
opened
3 years ago
4
期待能出个基于simcse的中文预训练模型
#2
lonngxiang
opened
3 years ago
3
运行环境同 readme, simcse_loss 中 `K.eye(K.shape(y_pred)[0])` 在 `encoder.compile` 时直接卡死
#1
ericxsun
closed
3 years ago
2