master

文件	最后提交记录	最后更新时间
core_model	【INFO update!!!】url替换 Co-authored-by: EX_mitsu<yangjie409@h-partners.com> # message auto-generated for no-merge-commit merge: merge master into master 【INFO update!!!】url替换 Created-by: EX_mitsuX Commit-by: EX_mitsu Merged-by: ascend-robot Description: 清理替换gitee相关url，添加新商发版本相关信息。 See merge request: Ascend/MindSpeed!2897	1 年前
deepseek	【INFO update!!!】url替换 Co-authored-by: EX_mitsu<yangjie409@h-partners.com> # message auto-generated for no-merge-commit merge: merge master into master 【INFO update!!!】url替换 Created-by: EX_mitsuX Commit-by: EX_mitsu Merged-by: ascend-robot Description: 清理替换gitee相关url，添加新商发版本相关信息。 See merge request: Ascend/MindSpeed!2897	1 年前
feature_precision_guarding	!2688 代码安全整改 Merge pull request !2688 from yuqi/safecode1	1 年前
feature_tests	!2822 新增fsdp的ST Merge pull request !2822 from yuqi/fsdp_st	1 年前
fused_ema_adamw	!2356 refactor: Switch v2 to v1 Merge pull request !2356 from 邓佳/master_refactor	1 年前
gpt	fix for te Co-authored-by: libaokui<libaokui@huawei.com> # message auto-generated for no-merge-commit merge: merge vmaster into master fix for te Created-by: libaokui Commit-by: libaokui Merged-by: ascend-robot Description: fix for te See merge request: Ascend/MindSpeed!2862	1 年前
llama2	!2748 fix ddp overlap-param-gather issue when param order and forward mismatch Merge pull request !2748 from 郭鹏/master	1 年前
llama3	Add offline pad_data Co-authored-by: wuweiqiang24<wuweiqiang11@huawei.com> # message auto-generated for no-merge-commit merge: !2938 merge revise_preprocess_data into master Add offline pad_data Created-by: wuweiqiang24 Commit-by: wuweiqiang24 Merged-by: ascend-robot Description: 增加离线预处理pack数据集功能，可提前将数据padding到2\*CP倍，在线使用CP功能时可节约padding部分耗时 * 精度与非离线padding版本存在一定差异  * Llama2-7b，单机16k，GBS=8场景下，性能提升4.8%  See merge request: Ascend/MindSpeed!2938	11 个月前
llava	!2284 fix: Adapt to core0.10.0; v2 is aligned with v1 Merge pull request !2284 from 邓佳/master_te	1 年前
mixtral	Add offline pad_data Co-authored-by: wuweiqiang24<wuweiqiang11@huawei.com> # message auto-generated for no-merge-commit merge: !2938 merge revise_preprocess_data into master Add offline pad_data Created-by: wuweiqiang24 Commit-by: wuweiqiang24 Merged-by: ascend-robot Description: 增加离线预处理pack数据集功能，可提前将数据padding到2\*CP倍，在线使用CP功能时可节约padding部分耗时 * 精度与非离线padding版本存在一定差异  * Llama2-7b，单机16k，GBS=8场景下，性能提升4.8%  See merge request: Ascend/MindSpeed!2938	11 个月前
multi_modal	fix for te Co-authored-by: libaokui<libaokui@huawei.com> # message auto-generated for no-merge-commit merge: merge vmaster into master fix for te Created-by: libaokui Commit-by: libaokui Merged-by: ascend-robot Description: fix for te See merge request: Ascend/MindSpeed!2862	1 年前
pipeline_parallel	!2356 refactor: Switch v2 to v1 Merge pull request !2356 from 邓佳/master_refactor	1 年前
resume_training	!2733 add resume training ST for dist ckpt and torch ckpt Merge pull request !2733 from wangyuansheng8/master	1 年前
verl	修改verl.md，新增示例训练脚本 Co-authored-by: wangjinyi6<wangjinyi6@huawei.com> # message auto-generated for no-merge-commit merge: !3055 merge master into master 修改verl.md，新增示例训练脚本 Created-by: wangjinyi6 Commit-by: wangjinyi6 Merged-by: ascend-robot Description: 修改verl.md，新增示例训练脚本 See merge request: Ascend/MindSpeed!3055	10 个月前
yaml_args_example	!2688 代码安全整改 Merge pull request !2688 from yuqi/safecode1	1 年前
env_npu.sh	!537 删除冗余代码 Merge pull request !537 from MingzhenWang/masterdelete	2 年前
pretrain_base.sh	!2676 Add triton patch Merge pull request !2676 from Jializheng/master	1 年前
pretrain_skip_train.sh	!383 change gpu to npu and fix some semantic error Merge pull request !383 from yangcheng/master	2 年前