minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 13 Packages Projects Releases Wiki Activity

[MLX] Support Qwen3.5 (dense) Model (#25754)

Browse Source 
Signed-off-by: Xiaodong Ye <yeahdongcn@gmail.com>
Co-authored-by: Alex Nails <alex.nails@radixark.ai>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
R0CKSTAR
2026-05-30 17:05:02 +08:00
committed by GitHub
co-authored by Alex Nails Claude Opus 4.6
parent 7c5708cba7
commit a952e9174f
23 changed files with 2943 additions and 291 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/bench_one_batch.py
+1 -1
View File
@@ -536,7 +536,7 @@ class _MlxBenchRunner:
if server_args.max_total_tokens is not None:
init_kwargs["pool_size"] = server_args.max_total_tokens
self.mlx_runner = MlxModelRunner(**init_kwargs)
self.mlx_runner.init_kv_pool(req_to_token_pool=None)
self.mlx_runner.init_cache_pools(req_to_token_pool=None)
self.fake_torch_runner = model_runner
def clear(self):