mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2025-12-24 13:28:13 +08:00
13 lines
299 B
YAML
13 lines
299 B
YAML
max_model_len: 32768
|
|
max_num_seqs: 256
|
|
tensor_parallel_size: 8
|
|
quantization: block_wise_fp8
|
|
gpu_memory_utilization: 0.9
|
|
kv_cache_ratio: 0.8
|
|
enable_chunked_prefill: True
|
|
max_num_batched_tokens: 1024
|
|
max_num_partial_prefills: 3
|
|
max_long_partial_prefills: 3
|
|
enable_prefix_caching: True
|
|
swap_space: 200
|