issues
search
ZhengkunTian
/
rnn-transducer
A Pytorch Implementation of Transducer Model for End-to-End Speech Recognition
231
stars
57
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Email error
#27
mflodzinski
opened
9 months ago
0
请问预训练和LM的代码在哪里啊？
#25
kartily
opened
1 year ago
0
关于encoder代码的问题
#24
ZR-HH
closed
3 years ago
0
关于CER，在验证集上最好也就28？
#23
a122760
opened
3 years ago
0
请问训练过程中loss降到负了是正常的吗？
#22
zhoumumu
opened
3 years ago
1
loss降低到 四五十后，不再下降了，在稳步上升。是啥情况。
#21
786440445
opened
3 years ago
0
about loss
#20
cst781
opened
3 years ago
1
Vocab
#19
Jiyongma06
opened
4 years ago
0
Issue about eval.py
#18
Shengqiang-Li
opened
4 years ago
0
“RuntimeError: CUDA out of the memory. ” happens during training(epoch:0(34%~36%))
#17
jinsongpan
opened
4 years ago
0
这要训练多少轮效果才好，感觉损失都好高，效果很差
#16
ZephyrChenzf
opened
4 years ago
8
一个batch内的识别结果相同是什么原因？
#15
ospeech
closed
4 years ago
7
RNNT+pretrain+LM中的pretrain指的是预训练的什么呢？
#14
oshindow
closed
4 years ago
1
LIne 140 in train.py
#13
oshindow
closed
4 years ago
1
Is there a pre-trained model available?
#12
vasurobo
closed
4 years ago
1
train.py return these errors
#11
oshindow
closed
4 years ago
5
how to preprocess the dataset?
#10
w5688414
closed
4 years ago
2
how to preprocess the dataset?
#9
w5688414
closed
4 years ago
0
请问160维MFCC特征是如何产生的？
#8
RenJinan
closed
4 years ago
3
Structure of egs/timit/data/{train,dev,test}
#7
someshjaishwal
closed
4 years ago
4
RNNTLoss
#6
cdliang11
closed
4 years ago
1
PRETRAINED MODEL
#5
tulasiram58827
closed
4 years ago
1
A question of recognize function implement
#4
huangnengCSU
closed
4 years ago
1
loss值为负数
#3
Augus-smile
closed
4 years ago
3
A question about vocab
#2
jackson1895
closed
4 years ago
1
A question about beam search
#1
Marcovaldong
closed
5 years ago
2