chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity

Cleanup ModelRunner (#15802)

Browse Source
This commit is contained in:
Liangsheng Yin
2025-12-25 18:13:30 +08:00
committed by GitHub
parent e6ce16a4c2
commit f4e835af2f
4 changed files with 62 additions and 74 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
docs/references/environment_variables.md
View File

@@ -67,7 +67,7 @@ SGLang supports various environment variables that can be used to configure its
| `SGLANG_DEBUG_MEMORY_POOL` | Enable memory pool debugging | `false` |
| `SGLANG_CLIP_MAX_NEW_TOKENS_ESTIMATION` | Clip max new tokens estimation for memory planning | `4096` |
| `SGLANG_DETOKENIZER_MAX_STATES` | Maximum states for detokenizer | Default value based on system |
| `SGL_DISABLE_TP_MEMORY_INBALANCE_CHECK` | Disable checks for memory imbalance across Tensor Parallel ranks | Not set (defaults to enabled check) |
| `SGLANG_ENABLE_TP_MEMORY_INBALANCE_CHECK` | Enable checks for memory imbalance across Tensor Parallel ranks | `true` |
## Model-Specific Options
@@ -112,7 +112,6 @@ SGLang supports various environment variables that can be used to configure its
| `SGLANG_TEST_RETRACT_NO_PREFILL_BS` | When SGLANG_TEST_RETRACT is enabled, no prefill is performed if the batch size exceeds SGLANG_TEST_RETRACT_NO_PREFILL_BS. | `2 ** 31` |
| `SGLANG_RECORD_STEP_TIME` | Record step time for profiling | `false` |
| `SGLANG_TEST_REQUEST_TIME_STATS` | Test request time statistics | `false` |
| `SGLANG_CI_SMALL_KV_SIZE` | Use small KV cache size in CI | `-1` |
## Profiling & Benchmarking