Update GLM-5.2 B300 and GB300 NVFP4 cookbook settings (#29466)

This commit is contained in:
Baizhou Zhang
2026-06-26 16:15:33 -07:00
committed by GitHub
parent c470acde2b
commit 12f76d115c
@@ -652,8 +652,8 @@ sgl-eval run aime25 \\
"--speculative-num-steps 5", "--speculative-num-steps 5",
"--speculative-eagle-topk 1", "--speculative-eagle-topk 1",
"--speculative-num-draft-tokens 6", "--speculative-num-draft-tokens 6",
"--chunked-prefill-size 131072", "--chunked-prefill-size 8192",
"--mem-fraction-static 0.70", "--mem-fraction-static 0.8",
"--host {{HOST_IP}}", "--host {{HOST_IP}}",
"--port {{PORT}}", "--port {{PORT}}",
], ],
@@ -667,8 +667,8 @@ sgl-eval run aime25 \\
"--model-path {{MODEL_NAME}}", "--model-path {{MODEL_NAME}}",
"--tp 4", "--tp 4",
"--quantization modelopt_fp4", "--quantization modelopt_fp4",
"--chunked-prefill-size 131072", "--chunked-prefill-size 8192",
"--mem-fraction-static 0.70", "--mem-fraction-static 0.8",
"--host {{HOST_IP}}", "--host {{HOST_IP}}",
"--port {{PORT}}", "--port {{PORT}}",
], ],
@@ -686,8 +686,8 @@ sgl-eval run aime25 \\
"--speculative-num-steps 5", "--speculative-num-steps 5",
"--speculative-eagle-topk 1", "--speculative-eagle-topk 1",
"--speculative-num-draft-tokens 6", "--speculative-num-draft-tokens 6",
"--chunked-prefill-size 131072", "--chunked-prefill-size 8192",
"--mem-fraction-static 0.70", "--mem-fraction-static 0.8",
"--host {{HOST_IP}}", "--host {{HOST_IP}}",
"--port {{PORT}}", "--port {{PORT}}",
], ],
@@ -701,8 +701,8 @@ sgl-eval run aime25 \\
"--model-path {{MODEL_NAME}}", "--model-path {{MODEL_NAME}}",
"--tp 4", "--tp 4",
"--quantization modelopt_fp4", "--quantization modelopt_fp4",
"--chunked-prefill-size 131072", "--chunked-prefill-size 8192",
"--mem-fraction-static 0.70", "--mem-fraction-static 0.8",
"--host {{HOST_IP}}", "--host {{HOST_IP}}",
"--port {{PORT}}", "--port {{PORT}}",
], ],