| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
docs(megatron):evaluation documents fix Co-authored-by: guihaowen666<guihaowen@huawei.com> # message auto-generated for no-merge-commit merge: !4838 merge br_master_doc_fix_0721 into master docs(megatron):evaluation documents fix Created-by: guihaowen666 Commit-by: guihaowen666 Merged-by: ascend-robot Description: ## What this PR does / why we need it? The evaluation module-related documents have some formatting issues, and this PR has corrected them. ## Does this PR introduce any user-facing change? Modified the functional documentation related to the evaluation module, without altering the business code. Thus there is no user-facing change. ## How was this patch tested? cross-review passed and CI gate check passed See merge request: Ascend/MindSpeed-LLM!4838 | 10 天前 | |
refactor(pytorch): delete ckpt-v1 Co-authored-by: qyzqyz<quyueze@h-partners.com> # message auto-generated for no-merge-commit merge: !4831 merge master into master refactor(pytorch): delete ckpt-v1 Created-by: qyzqyz Commit-by: qyzqyz Merged-by: ascend-robot Description: ## What this PR does / why we need it? This PR removes the legacy Weight Conversion V1 implementation and completes the migration to Weight Conversion V2. The main changes include: * Remove the deprecated Weight Conversion V1 codebase. * Migrate the remaining scripts that still depended on V1 to Weight Conversion V2. * Update related README and documentation to reflect the new conversion workflow. * Clean up obsolete files and references associated with the V1 implementation. This change simplifies code maintenance and ensures that all weight conversion workflows are unified under the actively maintained V2 framework. --- ## Does this PR introduce any user-facing change? Yes. Weight Conversion V1 is no longer supported. Users should use Weight Conversion V2 for all checkpoint conversion workflows. The related README and documentation have been updated accordingly to guide users to the new conversion process. --- ## How was this patch tested? The following verification was performed: * Verified that all remaining weight conversion scripts have been successfully migrated to Weight Conversion V2. * Verified that the updated conversion workflow functions correctly after removing the V1 implementation. * Verified that related documentation and examples are consistent with the current implementation. * Regression tests and pipeline ST passed successfully. See merge request: Ascend/MindSpeed-LLM!4831 | 8 小时前 | |
docs: correct model name capitalization and normalize code block formatting Co-authored-by: Rostellaria<xiaoshaoan@huawei.com> # message auto-generated for no-merge-commit merge: !4817 merge master into master docs: correct model name capitalization and normalize code block formatting Created-by: Rostellaria Commit-by: Rostellaria Merged-by: ascend-robot Description: [#1644](https://gitcode.com/Ascend/MindSpeed-LLM/issues/1644) ## What this PR does / why we need it? This PR fixes typos and capitalization errors in the inference and deterministic computation docs (e.g., Chatglm3 → ChatGLM3 , Lama3.1 → Llama3.1 , ROCE → RoCE ) and normalizes code block indentation and parameter formatting in chat.md . ## Does this PR introduce any user-facing change? No functional change; only documentation wording/format is corrected. Affected files: docs/zh/pytorch/tools/deterministic_computation.md , docs/zh/pytorch/training/inference/chat.md , docs/zh/pytorch/training/inference/inference.md . ## How was this patch tested? Verified by manual review that changes are limited to documentation text and formatting with no code logic involved, so no runtime tests are required. See merge request: Ascend/MindSpeed-LLM!4817 | 16 天前 | |
docs: rectify documentation usability issues Co-authored-by: daixzh<daixuezheng@huawei.com> # message auto-generated for no-merge-commit merge: !4804 merge doc-fix into master docs: rectify documentation usability issues Created-by: daixzh Commit-by: daixzh Merged-by: ascend-robot Description: ## What this PR does / why we need it? Fix documentation issues, including spacing conventions, punctuation, and other formatting inconsistencies. ## Does this PR introduce any user-facing change? Yes, multiple documentation files are updated for better readability and accuracy. ## How was this patch tested? Reviewed the updated markdown content for correctness and formatting. See merge request: Ascend/MindSpeed-LLM!4804 | 16 天前 | |
docs: clarify FSDP2 quick start vs training guide navigation Co-authored-by: lishuai<lishuai_math@163.com> # message auto-generated for no-merge-commit merge: !4652 merge docs/fsdp2-doc-navigation into master docs: clarify FSDP2 quick start vs training guide navigation Created-by: lishuai26 Commit-by: lishuai Merged-by: ascend-robot Description: ## What this PR does / why we need it? Readers can confuse the FSDP2 quick-start tutorial with the FSDP2 training/config guide. This PR clarifies navigation and adds cross-links so newcomers follow the correct path. Specific changes: - [Docs] Update docs/zh/docs_guide.md to distinguish FSDP2 quick start (fsdp2_quick_start.md) from the FSDP2 training guide (finetune_fsdp2.md), including directory-tree annotations. - [Docs] Add bi-directional notes/links between fsdp2_quick_start.md and finetune/fsdp2/finetune_fsdp2.md. Note: MindSpore navigation changes were dropped after maintainer feedback that MindSpore docs/code are being sunset. Related to #1346 Fixes #1541 ## Does this PR introduce any user-facing change? Yes. Documentation navigation and links for FSDP2 readers are updated. No training code, scripts, or runtime behavior changes. Impacted docs: - docs/zh/docs_guide.md - docs/zh/pytorch/training/fsdp2_quick_start.md - docs/zh/pytorch/training/finetune/fsdp2/finetune_fsdp2.md ## How was this patch tested? - Verified relative markdown links resolve to existing files in the repo. - Ascend docs pipeline (markdownlint / link validity) passed on the updated branch. See merge request: Ascend/MindSpeed-LLM!4652 | 12 天前 | |
docs(pytorch): updated doc for AIDD Co-authored-by: LQ1206<liuqian164@h-partners.com> # message auto-generated for no-merge-commit merge: !4768 merge master into master docs(pytorch): updated doc for AIDD Created-by: LQ1206 Commit-by: LQ1206 Merged-by: ascend-robot Description: ## What this PR does / why we need it? updated doc for AIDD ## Does this PR introduce any user-facing change? fix doc error ## How was this patch tested? No test See merge request: Ascend/MindSpeed-LLM!4768 | 18 天前 | |
docs: fix AIDD review documentation issues Co-authored-by: xuwenyue<xuwenyue1@huawei.com> # message auto-generated for no-merge-commit merge: !4791 merge docs/fix-issue-1610-master into master docs: fix AIDD review documentation issues Created-by: xuwenyue Commit-by: xuwenyue Merged-by: ascend-robot Description: ## What this PR does / why we need it? Fix documentation formatting issues including FAQ markdown formatting, precision_issue table alignment, supported_models content updates, and quick_start improvements. ## Does this PR introduce any user-facing change? Yes, multiple documentation files are updated for better readability and accuracy. ## How was this patch tested? Reviewed the updated markdown content for correctness and formatting. Related issue: #1610 See merge request: Ascend/MindSpeed-LLM!4791 | 16 天前 | |
docs: fix AIDD review documentation issues Co-authored-by: LinShua<707894133@qq.com> # message auto-generated for no-merge-commit merge: !4786 merge master_docs_aidd_0716 into master docs: fix AIDD review documentation issues Created-by: LinShua Commit-by: LinShua Merged-by: ascend-robot Description: Fixes #1610 ## What this PR does / why we need it? 1. Corrects AIDD-reviewed documentation issues in the DeepSeek-V3 example, including LoRA terminology, script naming, heading numbering, and checkpoint terminology. 2. Converts the GatedDeltaNet training-guide reference into a clickable relative Markdown link. 3. Removes the reported Chinese-English spacing and link-label spacing issues, and standardizes the YaRN display name. 4. This PR contains documentation-only changes; the review workbook and temporary artifacts are not included. ## Does this PR introduce any user-facing change? No. It improves documentation wording, navigation, and copy-paste clarity without changing training logic or runtime behavior. ## How was this patch tested? 1. git diff --check passed for the changed documentation files. 2. Repository-configured pre-commit was attempted on the five changed files, but hook-environment initialization could not complete because the GitCode hook repository connection stalled. The remote Git hooks passed during git push. 3. Verified the referenced local documentation targets exist and confirmed the pushed commit contains only the five intended documentation files. See merge request: Ascend/MindSpeed-LLM!4786 | 17 天前 |
| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
| 10 天前 | ||
| 8 小时前 | ||
| 16 天前 | ||
| 16 天前 | ||
| 12 天前 | ||
| 18 天前 | ||
| 16 天前 | ||
| 17 天前 |