[CE]add wint4 ep (#5355)

2026-04-23 00:17:25 +08:00 · 2025-12-03 15:17:47 +08:00
parent d5a9b75b4e
commit 04d35ace5e
1 changed files with 7 additions and 0 deletions
@@ -0,0 +1,7 @@
+num_gpu_blocks_override: 1024
+max_model_len: 8192
+max_num_seqs: 64
+data_parallel_size: 4
+tensor_parallel_size: 1
+enable_expert_parallel: True
+quantization: wint4