issues
search
OpenBMB
/
MiniCPM
MiniCPM3-4B: An edge-side LLM that surpasses GPT-3.5-Turbo.
Apache License 2.0
6.95k
stars
440
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
[Feature Request]: vllm 能直接用functioncall？
#198
lonngxiang
closed
3 weeks ago
8
What is llmxmapreduce? Any reference?
#197
world2vec
closed
2 days ago
3
请问embedding模型和rerank模型怎么finetune，用的FlagEmbedding的吗？
#196
LIUKAI0815
closed
2 weeks ago
3
[Bad Case]: error
#195
lhjlhj11
closed
1 week ago
2
[Feature Request]: 可否提供正確的CUDA version, PyTorch version, 以及 deepspeed version?
#194
joyyang1215
opened
4 weeks ago
1
vllm运行你们给的demo需要多少显存
#193
lifelsl
closed
1 month ago
1
为了方便技术交流，拉了一个多模态大模型技术交流群，有需要的大家可以加入
#192
feihuamantian
closed
1 month ago
1
[Bad Case]: 为什么推理速度比9b模型都要慢很多
#191
lixiaoyuan1029
closed
1 week ago
2
请教一下你们预训练用了1.1T tokens，花了多少GPU和时间
#190
zyh3826
opened
1 month ago
0
使用vllm推理时出现错误
#189
lifelsl
closed
1 month ago
2
[Bad Case]: 多模态 MiniCPM-V 推理报错
#188
c122-ode
opened
1 month ago
1
[Bad Case]: 多模态MiniCPM-V 2.0 transformers 推理报错
#187
wangyao123456a
closed
1 month ago
5
vllm使用lora微调后的模型报错
#186
ngz-sun
closed
1 month ago
1
[Bug]: 出现报错_pickle.UnpicklingError: Weights only load failed. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
#185
weiruijinglu
closed
1 week ago
1
[Bug]: 为什么Flash Attention2里不需要repeat_kv
#184
huyiwen
opened
1 month ago
0
增加了MiniCPM的教程入口
#183
LDLINGLINGLING
closed
2 months ago
0
[Bad Case]: 无法复现模型结构缩放的最优学习率一致性实验
#182
xiaofengShi
opened
2 months ago
0
在sft的过程中，我是个新手，我该如何选取最优的checkpoint，以及如何安排训练集和验证集的比例。
#181
GromZhang
closed
1 month ago
5
增加了xtuner的开源社区链接
#180
LDLINGLINGLING
closed
2 months ago
0
[Bad Case]: 安卓部署问题
#179
No3cat
opened
2 months ago
0
[Bad Case]: android部署问题
#178
No3cat
closed
1 month ago
1
增加了主页llama_factory的导航
#177
LDLINGLINGLING
closed
2 months ago
0
增加了qlora的训练方式
#176
LDLINGLINGLING
closed
2 months ago
0
请扫微信二维码加群，如果群失效，可以添加我微信加入：yx116169
#175
feihuamantian
closed
2 months ago
1
训练loss正常，但是验证集loss是nan
#174
lifelsl
closed
1 month ago
4
请问有MiniCPM-2B-sft-fp32版本的finetune代码吗
#173
lifelsl
closed
2 months ago
0
增加了使用langchain做多文件rag的demo，能够在6g以下的显卡上运行
#172
LDLINGLINGLING
closed
2 months ago
0
有兴趣可以加一下多模态群，大家一起交流多模态技术实战中遇到的问题
#171
feihuamantian
closed
2 months ago
1
增加了快速导航以及量化、llama_factory等等内容到readme_en
#170
LDLINGLINGLING
closed
2 months ago
0
增加了bnb的量化以及快速导航
#169
LDLINGLINGLING
closed
2 months ago
0
[Feature Request]: 可以取消每次加载模型都需要联网check revision吗？Could the check revision be cancelled before loading the model?
#168
Yonggie
opened
2 months ago
0
how to start Instruct finetune by MiniCPM？
#167
scalaboy
closed
2 months ago
1
增加了minicpm-s-1b模型的powerinfer部署示例
#166
LDLINGLINGLING
closed
2 months ago
0
[Bad Case]: how can i use model.forward to inference model?
#165
ziyanxzy
closed
1 month ago
1
相同的问题，期待解决方案
#164
haohenggang
closed
2 months ago
6
[Bad Case]: 推理时间过长
#163
jichaoqun
closed
2 months ago
5
修复了mlx中的两个bug
#162
LDLINGLINGLING
closed
3 months ago
0
增加了llama_factory的示例
#161
LDLINGLINGLING
closed
3 months ago
0
Adding sft data in pre-training
#160
jordane95
closed
3 months ago
1
[Bad Case]: 为什么pip install -r requirements.txt会不停地进行安装 requirements.txt以外的包
#159
Firestar117
closed
2 months ago
1
使用vllm推理模型，出现异常。
#158
GromZhang
closed
3 months ago
3
add support autoawq for minicpm
#157
LDLINGLINGLING
closed
3 months ago
0
原始代码的usertoken是针对2b的，其他模型会有问题,现在根据不同模型都会调整
#156
LDLINGLINGLING
closed
3 months ago
0
[Bug]: The checkpoint you are trying to load has model type `cpm_dragonfly` but Transformers does not recognize this architecture.
#155
uRENu
closed
3 months ago
3
在手机端输入长文本报告维度不对错误
#154
rudaoshi
closed
1 month ago
1
minicpm-2b的gsm8k复现结果与论文报告的差异
#153
Abigail61
opened
3 months ago
0
What sequence length was used during pretraining?
#152
petroskarypis
closed
3 months ago
1
模型权重转换Llama格式
#151
SCUT-ChenBD
closed
2 months ago
1
[Feature Request]: 我的使用场景需要实时计算PPL,但是发现没有比huggingface原生计算更方便的框架
#150
ShadowTeamCN
closed
3 months ago
4
[Bad Case]: 使用tokenizer.apply_chat_template数据时，数据结尾没有添加"</s>"，但是我看模型special文件中显示"eos_token"是"</s>"。
#149
mst272
closed
3 months ago
1
Previous
Next