Files
FastDeploy/benchmarks/yaml/qwen2_7b-32k-fp8-h800-tp1-static.yaml
2025-06-29 23:29:37 +00:00

7 lines
143 B
YAML

max_model_len: 32768
max_num_seqs: 128
kv_cache_ratio: 0.75
tensor_parallel_size: 1
quantization: wfp8afp8
enable_static_graph_inference: True