issues
search
wind91725
/
gpt2-ml-finetune-
根据gpt2-ml中文模型finetune自己的数据集
Apache License 2.0
43
stars
15
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Bump requests from 2.22.0 to 2.31.0
#47
dependabot[bot]
opened
1 year ago
0
Bump tensorflow from 1.13.1 to 2.11.1
#46
dependabot[bot]
opened
1 year ago
0
Bump tensorflow-gpu from 1.13.1 to 2.12.0
#45
dependabot[bot]
opened
1 year ago
0
Bump tensorflow from 1.13.1 to 2.9.3
#44
dependabot[bot]
closed
1 year ago
1
Bump tensorflow-gpu from 1.13.1 to 2.9.3
#43
dependabot[bot]
closed
1 year ago
1
why pre_data output no text (pre_data為什麼會把文本處理成空)
#42
snowflowersnowflake
opened
1 year ago
0
Bump numpy from 1.16.2 to 1.22.0
#41
dependabot[bot]
opened
2 years ago
0
Bump tensorflow from 1.13.1 to 2.7.2
#40
dependabot[bot]
closed
1 year ago
1
Bump tensorflow-gpu from 1.13.1 to 2.7.2
#39
dependabot[bot]
closed
1 year ago
1
Bump tensorflow from 1.13.1 to 2.6.4
#38
dependabot[bot]
closed
2 years ago
1
Bump tensorflow-gpu from 1.13.1 to 2.6.4
#37
dependabot[bot]
closed
2 years ago
1
Bump tensorflow from 1.13.1 to 2.5.3
#36
dependabot[bot]
closed
2 years ago
1
Bump tensorflow-gpu from 1.13.1 to 2.5.3
#35
dependabot[bot]
closed
2 years ago
1
Bump tensorflow from 1.13.1 to 2.5.1
#34
dependabot[bot]
closed
2 years ago
1
Bump tensorflow-gpu from 1.13.1 to 2.5.1
#33
dependabot[bot]
closed
2 years ago
1
执行第二步出错
#32
ghost
opened
3 years ago
2
【讨论】gpt2-ml，30G，22w步模型微调报错解决方案
#31
NLPIG
opened
3 years ago
1
word_embed:0 ((21130, 1536))、word_embed ([8021, 1536]) 数组维度不一致有解决办法吗？
#30
NLPIG
closed
3 years ago
0
Bump tensorflow from 1.13.1 to 2.5.0
#29
dependabot[bot]
closed
3 years ago
1
请问一下 batch_size=1的微调情况下，learning rate的选择？
#28
huangdacheng
opened
3 years ago
0
/data/home/share1/gpt2-ml-Finetune/data-mayun_xiugai
#27
xukai98
opened
3 years ago
0
Bump tensorflow from 1.13.1 to 2.3.1
#26
dependabot[bot]
closed
3 years ago
1
Bump tensorflow-gpu from 1.13.1 to 2.3.1
#25
dependabot[bot]
closed
3 years ago
1
finetuning的batchsize是1 就可以了？
#24
huangdacheng
opened
3 years ago
0
为啥会出现：is not in all_model_checkpoint_paths
#23
cncbec
opened
3 years ago
1
22.json可以生成，train.tfrecord生成不了
#22
cncbec
opened
3 years ago
0
loss下降到多少是差不多了
#21
joytianya
opened
3 years ago
0
train/train_wc.py的输出路径是不是改为原模型地址
#20
cncbec
closed
3 years ago
3
Bump tensorflow from 1.13.1 to 1.15.4
#19
dependabot[bot]
closed
3 years ago
1
Bump tensorflow-gpu from 1.13.1 to 1.15.4
#18
dependabot[bot]
closed
3 years ago
1
加载原gpt2-ml的大模型，9G的显存也会满了,报OOM，请问作者是使用多大显存？
#17
huangdacheng
opened
3 years ago
1
bs=1，训练有效果么
#16
joytianya
opened
3 years ago
1
和原来gpt2-ml的demo.py 进行inference的代码对比，使用的不是同一个vocabulary？
#15
huangdacheng
opened
4 years ago
3
生成tfrecord这一步时出错
#14
dakkor
closed
4 years ago
1
AttributeError: module 'tensorflow.compat.v1' has no attribute 'contrib'
#13
qiujz
closed
4 years ago
3
gpu最低要求？
#12
kifish
closed
4 years ago
2
gpt2_ml的模型更新了，可以更新一下代码吗？还有，可以出个colab版的吗，运行起来也方便一些
#11
suxuan123
opened
4 years ago
5
可以给一个输入数据的例子吗，就是 pre_data.py的输入文件格式
#10
EiraZhang
opened
4 years ago
9
finetune时，总是找不到预训练模型文件是怎么回事呢？
#9
lrz512699597
opened
4 years ago
0
CUDA_VISIBLE_DEVICES=0 是做什么的呢？
#8
lrz512699597
opened
4 years ago
1
自己的数据集需要多大效果会比较好
#7
fred-github
opened
4 years ago
9
pre_data.py
#6
lrz512699597
opened
4 years ago
3
"iterations_per_loop", 1600,
#5
lrz512699597
opened
4 years ago
0
能不能使用多机多卡训练，小作坊用的是4*4 1080ti,
#4
SeekPoint
opened
4 years ago
0
filepath /data/home/share1/gpt2-ml-Finetune/data-mayun_xiugai
#3
SeekPoint
closed
4 years ago
1
Bump tensorflow-gpu from 1.13.1 to 1.15.2
#2
dependabot[bot]
closed
3 years ago
1
Bump tensorflow from 1.13.1 to 1.15.2
#1
dependabot[bot]
closed
3 years ago
1