[NPU] Offloading docs update (#22860)

Co-authored-by: Jianzhao Xu <xujianchao@huawei.com>
This commit is contained in:
jianzhao-xu
2026-04-15 15:04:41 +08:00
committed by GitHub
co-authored by Jianzhao Xu
parent b2af34be54
commit 45a83ffbe3
@@ -297,15 +297,15 @@ click [Server Arguments](https://docs.sglang.io/advanced_features/server_argumen
|--------------------|----------|--------------------------------|:----------------:|
| `--enable-lmcache` | `False` | bool flag<br/> (set to enable) | Special for GPU |
## Offloading
## Offloading (must be used with `--disable-cuda-graph`)
| Argument | Defaults | Options | Server supported |
|---------------------------|----------|-----------|:----------------:|
| `--cpu-offload-gb` | `0` | Type: int (must be used with `--disable-cuda-graph`) | A2, A3 |
| `--offload-group-size` | `-1` | Type: int | A2, A3 |
| `--offload-num-in-group` | `1` | Type: int | A2, A3 |
| `--offload-prefetch-step` | `1` | Type: int | A2, A3 |
| `--offload-mode` | `cpu` | `cpu`,<br/> `meta`,<br/> `sharded_gpu` (must be used with `--disable-cuda-graph`, only supported for DeepSeek) | A2, A3 |
| `--cpu-offload-gb` | `0` | Type: int | A2, A3 |
| `--offload-group-size` | `-1` | Type: int (DeepSeek only) | A2, A3 |
| `--offload-num-in-group` | `1` | Type: int (DeepSeek only) | A2, A3 |
| `--offload-prefetch-step` | `1` | Type: int (DeepSeek only) | A2, A3 |
| `--offload-mode` | `cpu` | `cpu` (DeepSeek only) <br/>`meta` (DeepSeek only) <br/>`sharded_gpu` (DeepSeek only) | A2, A3 |
## Args for multi-item scoring