2.3.0

文件	最后提交记录	最后更新时间
chatglm3	!3387 [mindspore][master]add chatglm sh Merge pull request !3387 from 杨承翰/sh_master	10 个月前
deepseek2_lite	!3171 [pytorch][feature]Deprecate some moe parameters: n-group, topk-group, etc. Merge pull request !3171 from shengjy/dep_moe	11 个月前
deepseek3	[pytorch][feature] Replace reset_position_ids with reset_attention_mask, and enable ring attention support when reset_attention_mask is active. Co-authored-by: mhh001<mahonghao1@huawei.com> # message auto-generated for no-merge-commit merge: !3506 merge master into master [pytorch][feature] Replace reset_position_ids with reset_attention_mask, and enable ring attention support when reset_attention_mask is active. Created-by: mhh111 Commit-by: mhh001 Merged-by: ascend-robot Description: pack/neat-pack场景ring cp支持 reset-attention-mask 对齐megatron 定长场景的支持 See merge request: Ascend/MindSpeed-LLM!3506	8 个月前
gemma2	!3384 [mindspore][master]add gemma sh Merge pull request !3384 from 杨承翰/gemma	10 个月前
glm4	[mindspore][docs]correct spelling errors Co-authored-by: nomiz<1606135114@qq.com> # message auto-generated for no-merge-commit merge: !3603 merge master into master [mindspore][docs]correct spelling errors Created-by: nomiz Commit-by: nomiz Merged-by: ascend-robot Description: [mindspore][docs]correct spelling errors See merge request: Ascend/MindSpeed-LLM!3603	9 个月前
glm45-moe	!3187 [pytorch][bugfix]fix name in args and remove checkout commitid Merge pull request !3187 from guozhihua/fix-deepseek-args	11 个月前
grok1	!3383 [mindspore][master]add grok1 sh Merge pull request !3383 from 杨承翰/grok	10 个月前
internlm25	!3301 [mindspore][master]add internlm pass sh Merge pull request !3301 from 杨承翰/sh_master	10 个月前
llama2	[pytorch][feature] Replace reset_position_ids with reset_attention_mask, and enable ring attention support when reset_attention_mask is active. Co-authored-by: mhh001<mahonghao1@huawei.com> # message auto-generated for no-merge-commit merge: !3506 merge master into master [pytorch][feature] Replace reset_position_ids with reset_attention_mask, and enable ring attention support when reset_attention_mask is active. Created-by: mhh111 Commit-by: mhh001 Merged-by: ascend-robot Description: pack/neat-pack场景ring cp支持 reset-attention-mask 对齐megatron 定长场景的支持 See merge request: Ascend/MindSpeed-LLM!3506	8 个月前
llama3	!3382 [mindspore][master]add llama sh Merge pull request !3382 from 杨承翰/llama	10 个月前
llama31	!3307 [mindspore][master]add llama pass sh Merge pull request !3307 from 杨承翰/llama_master	10 个月前
llama32	!3307 [mindspore][master]add llama pass sh Merge pull request !3307 from 杨承翰/llama_master	10 个月前
llama33	!3307 [mindspore][master]add llama pass sh Merge pull request !3307 from 杨承翰/llama_master	10 个月前
phi35	!3306 [mindspore][master]add phi35 pass sh Merge pull request !3306 from 杨承翰/phi_master	10 个月前
qwen25	[mindspore][docs][master]update mindspore docs Co-authored-by: ybwang19<1605891897@qq.com> # message auto-generated for no-merge-commit merge: !3568 merge master into master [mindspore][docs][master]update mindspore docs Created-by: ybwang19 Commit-by: ybwang19 Merged-by: ascend-robot Description: PTA燎原行动-资料整改 See merge request: Ascend/MindSpeed-LLM!3568	9 个月前
qwen3	[mindspore][sh]add qwen3 lora sh Co-authored-by: nomiz<1606135114@qq.com> # message auto-generated for no-merge-commit merge: !3916 merge 2.3.0 into 2.3.0 [mindspore][sh]add qwen3 lora sh Created-by: nomiz Commit-by: nomiz Merged-by: ascend-robot Description: [mindspore][sh]add qwen3 lora sh 验证 tune_qwen3_0.6b_lora 精度100步对齐  See merge request: Ascend/MindSpeed-LLM!3916	7 个月前
qwen3_moe	[mindspore][sh] add qwen3_235b Co-authored-by: wanglijun55<wanglijun54@huawei.com> # message auto-generated for no-merge-commit merge: !3656 merge branch-qwen3-235b into master [mindspore][sh] add qwen3_235b Created-by: wanglijun55 Commit-by: wanglijun55 Merged-by: ascend-robot Description: 1. 精度对齐结果：确定性开启100步强对齐PTA，确定性关闭500步和PTA误差小于千分之五  2. 显存对齐结果，不低于PTA50MB  3. 性能对齐结果，超过PTA的0.9x  See merge request: Ascend/MindSpeed-LLM!3656	8 个月前