Files
FastDeploy/benchmarks/yaml/eb45-32k-bf16-a30-tp1.yaml
2025-06-29 23:29:37 +00:00

6 lines
114 B
YAML

max_model_len: 32768
max_num_seqs: 128
kv_cache_ratio: 0.75
tensor_parallel_size: 1
max_num_batched_tokens: 32768