关于GCU、沐曦GPGPU、MLU、0卡V100资源4月7日恢复上架的公告>>> 关于共建具身智能开源数据集的倡议>>> 关于云脑任务中统一路径访问方式的公告>>> 关于将启智集群GPU资源迁移至智算集群的公告>>>
OpenBMB
/
BMTrain
Not watched
11
10
Fork 17
Code
Releases 0 Wiki Activity
Issues 1 Pull Requests 1 Datasets Model Cloudbrain
You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
296 Commits
1 Branch
2.4 MiB
1679 Download
 
 
 
 
 
 
Branch: main
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'main'
${ noResults }
BMTrain/bmtrain
Achazwl dc8f915188 add bf16 support 1 year ago
..
benchmark add batch_related interface 1 year ago
distributed FIX: ZERO2 incorrect indent 1 year ago
inspect FIX: nccl reduce use no avg op for compatibility 1 year ago
loss Merge branch 'main' into opendelta_interface 2 years ago
lr_scheduler fix warmup 2 years ago
nccl add bf16 support 1 year ago
optim FIX: typo 1 year ago
__init__.py unify loss_scale and normal -> OptimManager 1 year ago
block_layer.py add bf16 support 1 year ago
checkpointing.py FX: type hint 2 years ago
debug.py rename 2 years ago
global_var.py modified for merge 1 year ago
init.py Merge branch 'main' of https://github.com/MayDomine/BMTrain into MayDomine-main 1 year ago
layer.py FX: check parameter shape 2 years ago
param_init.py add test for init parameters and save/load parameters 1 year ago
parameter.py UPD: add default rank/world_size if bmtrain is not initialized. Raise RuntimeError when calling distributed functions if bmtrain is not initialized. 1 year ago
pipe_layer.py reduce_scatter only once for all micros 1 year ago
store.py FIX pipeline inspect_model stuck 1 year ago
synchronize.py FIX: nccl reduce use no avg op for compatibility 1 year ago
utils.py DOCS: update docs 1 year ago
wrapper.py FX: buffer is None 2 years ago
Delete task

Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.

cancel
confirm

高效的大模型训练（预训练、微调）工具包！

https://www.openbmb.org/

Python C++ C Cuda Markdown other

Please read the following content carefully:

Dear OpenI User

Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.

For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》

Agree and continue
Disagree, exit