docs:资料文档core版本更新以及安装指导修改
Co-authored-by: Your Name<you@example.com>
# message auto-generated for no-merge-commit merge:
!2521 merge master into master
docs:资料文档core版本更新以及安装指导修改
Created-by: zs-derrick1
Commit-by: Your Name
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
资料文档core版本更新以及安装指导修改
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2521
[Bugfix]InternVL3_5 README: Fixing the model name subdirectory of hf_path.
Co-authored-by: ruanhao9656<ruanhao7@h-partners.com>
# message auto-generated for no-merge-commit merge:
!2491 merge master into master
[Bugfix]InternVL3_5 README: Fixing the model name subdirectory of hf_path.
Created-by: ruanhao9656
Commit-by: ruanhao9656
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
Please describe the background and detailed changes of the PR. If it is a bugfix, please attach the related issue.
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2491
[Feature] fsdp2 config construct update
Co-authored-by: yangx_sy<sunyang49@huawei.com>
# message auto-generated for no-merge-commit merge:
!2482 merge fsdp2_args into master
[Feature] fsdp2 config construct update
Created-by: yangx_sy
Commit-by: yangx_sy
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
对于纯血fsdp2后端，拆分其配置文件，将一些优化配置从model项拆分出来
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2482
[Feature] fsdp2 config construct update
Co-authored-by: yangx_sy<sunyang49@huawei.com>
# message auto-generated for no-merge-commit merge:
!2482 merge fsdp2_args into master
[Feature] fsdp2 config construct update
Created-by: yangx_sy
Commit-by: yangx_sy
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
对于纯血fsdp2后端，拆分其配置文件，将一些优化配置从model项拆分出来
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2482
feat: qwen3.5moe and qwen3.5 support mtp
Co-authored-by: yaoyaoxu<xuyaoyao.824404@huawei.com>
# message auto-generated for no-merge-commit merge:
!2389 merge mtp_qwen35 into master
feat: qwen3.5moe and qwen3.5 support mtp
Created-by: yaoyaoxu
Commit-by: yaoyaoxu
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
1.修复bug：chunk loss缺少入参labels
2. qwen3.5moe和qwen3.5开发mtp模块，均支持chunk loss和cp
## Does this PR introduce any user-facing change?
用户可以在xxx_config.yaml中model下的mtp_num_layers字段配置为1来开启mtp模块，其中mtp_loss_scaling_factor字段支持配置，默认为0.1
## How was this patch tested?
通过整网进行验证
See merge request: Ascend/MindSpeed-MM!2389
fix:fix preload and feature bug and modify attn_implementation
Co-authored-by: WendongPang<pangwendong@huawei.com>
# message auto-generated for no-merge-commit merge:
!2534 merge preloader into master
fix:fix preload and feature bug and modify attn_implementation
Created-by: WendongPang
Commit-by: WendongPang
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
fix:fix preload bug and modify attn_implementation https://gitcode.com/Ascend/MindSpeed-MM/issues/322
enable_preload不需要在非纯fsdp2后端的yaml脚本中配置，添加model部分参数转移到feature参数，部分地方未改全，以及修改attention的实现方式为flash_attention_2
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2534
[Feature] fsdp2 config construct update
Co-authored-by: yangx_sy<sunyang49@huawei.com>
# message auto-generated for no-merge-commit merge:
!2482 merge fsdp2_args into master
[Feature] fsdp2 config construct update
Created-by: yangx_sy
Commit-by: yangx_sy
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
对于纯血fsdp2后端，拆分其配置文件，将一些优化配置从model项拆分出来
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2482
[Feature] fsdp2 config construct update
Co-authored-by: yangx_sy<sunyang49@huawei.com>
# message auto-generated for no-merge-commit merge:
!2482 merge fsdp2_args into master
[Feature] fsdp2 config construct update
Created-by: yangx_sy
Commit-by: yangx_sy
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
对于纯血fsdp2后端，拆分其配置文件，将一些优化配置从model项拆分出来
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2482
fix:fix preload and feature bug and modify attn_implementation
Co-authored-by: WendongPang<pangwendong@huawei.com>
# message auto-generated for no-merge-commit merge:
!2534 merge preloader into master
fix:fix preload and feature bug and modify attn_implementation
Created-by: WendongPang
Commit-by: WendongPang
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
fix:fix preload bug and modify attn_implementation https://gitcode.com/Ascend/MindSpeed-MM/issues/322
enable_preload不需要在非纯fsdp2后端的yaml脚本中配置，添加model部分参数转移到feature参数，部分地方未改全，以及修改attention的实现方式为flash_attention_2
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2534
[Docs] Modify current repository URLs to relative paths
Co-authored-by: AZe_404<wangze62@h-partners.com>
# message auto-generated for no-merge-commit merge:
!2364 merge chg_branch into master
[Docs] Modify current repository URLs to relative paths
Created-by: AZe_404
Commit-by: AZe_404
Merged-by: ascend-robot
Description: ## What this PR does / why we need it?
将MM仓库内的链接修改为相对路径访问.
## Does this PR introduce any user-facing change?
Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path.
## How was this patch tested?
Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations.
See merge request: Ascend/MindSpeed-MM!2364
