ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[launcher] Add elastic and fault-tolerant training support (#8286)

Browse Source 
Signed-off-by: Butui Hu <hot123tea123@gmail.com>
This commit is contained in:
Butui Hu
2025-06-05 16:40:03 +08:00
committed by GitHub
parent 69c9e379d5
commit 1a33d65a56
3 changed files with 66 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
examples/README.md
View File

@@ -165,6 +165,14 @@ FORCE_TORCHRUN=1 NNODES=2 NODE_RANK=0 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500
FORCE_TORCHRUN=1 NNODES=2 NODE_RANK=1 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
```
### Elastic and Fault-Tolerant Supervised Fine-Tuning on Multiple Nodes
To launch an elastic job with `MAX_RESTARTS` failures retries, run the following on at least `MIN_NNODES` nodes and at most `MAX_NNODES` nodes. `RDZV_ID` should be set as a unique job id (shared by all nodes participating in the job). See also [torchrun](https://docs.pytorch.org/docs/stable/elastic/run.html).
```bash
FORCE_TORCHRUN=1 MIN_NNODES=1 MAX_NNODES=3 MAX_RESTARTS=3 RDZV_ID=llamafactory MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
```
#### Multimodal Supervised Fine-Tuning
```bash

8
examples/README_zh.md
View File

@@ -106,6 +106,14 @@ FORCE_TORCHRUN=1 NNODES=2 NODE_RANK=0 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500
FORCE_TORCHRUN=1 NNODES=2 NODE_RANK=1 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_lora/llama3_lora_sft.yaml
```
### 支持弹性和容错的多机指令监督微调
要启动一个支持弹性节点和容错的多机指令微调，在每个节点上执行以下命令。弹性节点数量范围为 `MIN_NNODES:MAX_NNODES`，每个节点最多允许因为错误重启 `MAX_RESTARTS` 次。`RDZV_ID` 应设置为一个唯一的作业 ID（由参与该作业的所有节点共享）。更多新可以参考官方文档 [torchrun](https://docs.pytorch.org/docs/stable/elastic/run.html)。
```bash
FORCE_TORCHRUN=1 MIN_NNODES=1 MAX_NNODES=3 MAX_RESTARTS=3 RDZV_ID=llamafactory MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
```
#### 使用 DeepSpeed ZeRO-3 平均分配显存
```bash