Files
FastDeploy/benchmarks/yaml/paddleocr-vl-16k-bf16.yaml
T
2026-01-19 20:07:42 +08:00

7 lines
132 B
YAML

max_model_len: 16384
max_num_seqs: 256
max_num_batched_tokens: 16384
tensor_parallel_size: 1
gpu_memory_utilization: 0.7
workers: 4