Skip to content

Commit 7784a89

Browse files
authored
[docs] update qwen3.5 best practice (#9039)
1 parent 501a0c9 commit 7784a89

2 files changed

Lines changed: 0 additions & 3 deletions

File tree

docs/source/BestPractices/Qwen3_5-Best-Practice.md

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -308,7 +308,6 @@ swift infer \
308308
```
309309

310310
Megatron-SWIFT训练Qwen3.5的提示:
311-
- 全参数训练:参考[这个例子](https://github.com/modelscope/ms-swift/tree/main/examples/models/qwen3_5/mcore_full.sh)
312311
- 关于MTP训练:"mcore-bridge>=1.1.0"支持了多模态MTP的训练(暂时需安装[main分支](https://github.com/modelscope/mcore-bridge/pull/14)),请安装对应版本。
313312
- TP 限制解除:使用 "megatron-core>=0.16" 可解除 TP 受到的 `num_query_groups` 限制。
314313
- CP支持:"mcore-bridge>=1.1.0"支持了GDN的CP训练(暂时需安装[main分支](https://github.com/modelscope/mcore-bridge/pull/16)),此外需安装megatron-core dev分支。

docs/source_en/BestPractices/Qwen3_5-Best-Practice.md

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -305,8 +305,6 @@ swift infer \
305305
```
306306

307307
Tips for training Qwen3.5 with Megatron-SWIFT:
308-
309-
- Full parameter training: Refer to [this example](https://github.com/modelscope/ms-swift/tree/main/examples/models/qwen3_5/mcore_full.sh).
310308
- Regarding MTP training: `mcore-bridge>=1.1.0` supports multimodal MTP training (currently requires installing the [main branch](https://github.com/modelscope/mcore-bridge/pull/14)). Please install the corresponding version.
311309
- TP Limitation Removed: Using `megatron-core>=0.16` removes the `num_query_groups` limitation on TP.
312310
- CP support: "mcore-bridge>=1.1.0" supports CP training for GDN (currently requires installing the [main branch](https://github.com/modelscope/mcore-bridge/pull/16)). Additionally, the megatron-core dev branch needs to be installed.

0 commit comments

Comments
 (0)