issues
search
yuanzhoulvpi2017
/
zero_nlp
中文nlp解决方案(大模型、数据、模型、训练、推理)
MIT License
2.81k
stars
351
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
bloom 也是 casulmModel 体系下的，是否可以用cpu加速推理
#133
xx-zhang
closed
1 year ago
1
请问满足模型并行化的条件是什么
#132
taofennanhai
opened
1 year ago
10
train model all error
#131
yxk9810
opened
1 year ago
3
运行Chatglm6b_ModelParallel代码时候，模型是下载 huggingface上的THUDM/chatglm-6bcommit为d2bbc82a2出错
#130
Ardang666
closed
1 year ago
2
torch
#129
yangliuIOC
closed
1 year ago
0
chinese_bloom 支持上下文对话吗
#128
gebilaoman
opened
1 year ago
1
chinese bloom的默认padding side为什么改成了right
#127
DZ9
opened
1 year ago
1
裁剪词表
#126
yangliuIOC
opened
1 year ago
1
Lora训练一段时间后出现OOM报错
#125
976311200
opened
1 year ago
2
Chatglm6b_ModelParalle子项尝试失败，遇到模型加载问题
#124
shaoqing404
opened
1 year ago
4
chinese_bloom通过ds训练报错：RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0! (when checking argument for argument weight in method wrapper_CUDA__native_layer_norm)
#123
shaoqing404
opened
1 year ago
4
大佬
#122
yangliuIOC
closed
1 year ago
1
模型并行启动方式
#121
kevinuserdd
opened
1 year ago
0
CUDA Error
#120
yuntong613
opened
1 year ago
2
ptuning训练完了，为什么还会显示模型没有训练？
#119
happyjiaojiao
opened
1 year ago
0
我想使用deepspeed训练bloom，但发现以下错误
#118
fredericklee602
opened
1 year ago
1
请问在跑Chatglm6b_ModelParallel模型的时候报这个错该怎么解决啊
#117
string-new
opened
1 year ago
2
总是报这个错，怎么才是本地文件夹，我已经下载到本地了啊。chatglm-6b is not a local folder and is not a valid model identifier listed on 'https://huggingface.co/models'
#116
cat88hzh
opened
1 year ago
2
运行train_parallel.sh报错
#115
MathamPollard
opened
1 year ago
2
想请教一下做微调的时候用 embedding 输入模型而不是 input_ids，应该在哪部分做修改呀？
#114
shtdbb
closed
1 year ago
2
code02_训练模型全部流程.ipynb 代码是否有误？
#113
ghost
closed
1 year ago
0
lora训练 多卡报错
#112
aihaidong
closed
1 year ago
3
AttributeError: 'ChatGLMTokenizer' object has no attribute 'eop_token_id'
#111
withyou971
closed
1 year ago
6
多卡运行报错(Chatglm6b_ModelParallel_ptuning, 训练)
#110
hunwenpinghao
closed
1 year ago
4
遇到报错 RuntimeError: self and mat2 must have the same dtype
#109
diaojunxian
closed
1 year ago
1
训练集测试集划分导致数据泄漏
#108
yongqiangning
closed
1 year ago
1
模型并行全量微调后，推理很慢，调用web_demo.py出不来结果
#107
laiqinghan
closed
1 year ago
0
我用了两张GPU，训练起来只占用了第二张GPU的内存，没有使用算力，这个是为什么
#106
zx19941234
closed
1 year ago
4
基于ChatGlm ptuning的也用了LoRA么？
#105
qingyuan18
closed
1 year ago
1
chinese_dolly_v2_3b无法支持fp16进行训练问题
#104
Bob199511
closed
1 year ago
3
deepspeed
#103
kevinuserdd
closed
1 year ago
0
lora微调很难把握结果质量啊 多轮结果展示 貌似900轮后过拟合 但什么时候停止微调如何把握
#102
alexhmyang
closed
1 year ago
4
quantize(8) 量化后的原始模型和lora checkpoints 不匹配，报错 RuntimeError: self and mat2 must have the same dtype，怎么统一起来？
#101
alexhmyang
opened
1 year ago
1
code02_训练模型全部流程.ipynb运行问题
#100
situjunhao
closed
1 year ago
0
单机多卡是基于哪个版本的，P-Tuning 还是lora，为什么模型又换成THUDM/chatglm-6b，而不是yuanzhoulvpi/chatglm6b-dddd
#99
YSLLYW
closed
1 year ago
1
大佬，微调完成后，要想基于微调后的版本，继续微调，你们一般是怎么弄的？
#98
cywjava
closed
1 year ago
4
大佬 ，帮忙看看这个错误 。
#97
cywjava
opened
1 year ago
1
多卡并行训练并没有提升速度反而变慢了些，这是为什么
#96
heheyzc
closed
1 year ago
9
get_masks_and_position_ids 问题请教
#95
chenjw505
opened
1 year ago
0
infer【bug】RuntimeError
#94
putdoor
opened
1 year ago
1
gpt预训练
#93
Clearloveplus7
closed
1 year ago
2
模型并行训练的问题
#92
aihaidong
opened
1 year ago
2
进行多卡训练的时候，device_map_dict应该如何进行编号设置
#91
168liuliu168
opened
1 year ago
4
请问4.1版Chatglm6b_ModelParallel训练loss下降学不到东西bug修复了嘛
#90
guoswang
opened
1 year ago
2
我的模型文件从dddd版本里下的，Chatglm6b_ModelParallel这个文件夹下，只修改了cuda的配置，训练还是遇到了问题。
#89
Rorschaaaach
opened
1 year ago
8
训练epoch的设置和效果问题
#88
GreatWildFire
opened
1 year ago
4
模型并行，数据没有并行？
#87
GUORUIWANG
closed
1 year ago
2
单机多卡训练chat_glm 有误
#86
cxj01
opened
1 year ago
6
[BUG]训练脚本报错
#85
aihaidong
closed
1 year ago
6
合并Lora权重后的模型不生成回答了
#84
heccxixi
closed
1 year ago
0
Previous
Next