[AMD] Cap DSV4 Flash max_total_num_tokens (#30313)
Co-authored-by: YC Yen-Ching Tseng <yctseng@amd.com>
This commit is contained in:
co-authored by
YC Yen-Ching Tseng
parent
99db3b0fa5
commit
dabd4cfcfd
@@ -33,6 +33,7 @@ runtime:
|
||||
mem_fraction_static: 0.90
|
||||
page_size: 256
|
||||
max_running_requests: 256
|
||||
max_total_tokens: 8551168
|
||||
chunked_prefill_size: 8192
|
||||
swa_full_tokens_ratio: 0.1
|
||||
|
||||
|
||||
@@ -33,6 +33,7 @@ runtime:
|
||||
mem_fraction_static: 0.90
|
||||
page_size: 256
|
||||
max_running_requests: 256
|
||||
max_total_tokens: 8551168
|
||||
chunked_prefill_size: 8192
|
||||
swa_full_tokens_ratio: 0.1
|
||||
|
||||
|
||||
@@ -33,6 +33,7 @@ runtime:
|
||||
mem_fraction_static: 0.90
|
||||
page_size: 256
|
||||
max_running_requests: 256
|
||||
max_total_tokens: 8551168
|
||||
chunked_prefill_size: 8192
|
||||
swa_full_tokens_ratio: 0.1
|
||||
|
||||
|
||||
@@ -35,6 +35,7 @@ runtime:
|
||||
mem_fraction_static: 0.90
|
||||
page_size: 256
|
||||
max_running_requests: 256
|
||||
max_total_tokens: 8551168
|
||||
chunked_prefill_size: 8192
|
||||
swa_full_tokens_ratio: 0.1
|
||||
|
||||
|
||||
@@ -33,6 +33,7 @@ runtime:
|
||||
mem_fraction_static: 0.90
|
||||
page_size: 256
|
||||
max_running_requests: 256
|
||||
max_total_tokens: 8551168
|
||||
chunked_prefill_size: 8192
|
||||
swa_full_tokens_ratio: 0.1
|
||||
|
||||
|
||||
@@ -33,6 +33,7 @@ runtime:
|
||||
mem_fraction_static: 0.90
|
||||
page_size: 256
|
||||
max_running_requests: 256
|
||||
max_total_tokens: 8551168
|
||||
chunked_prefill_size: 8192
|
||||
swa_full_tokens_ratio: 0.1
|
||||
|
||||
|
||||
@@ -33,6 +33,7 @@ runtime:
|
||||
mem_fraction_static: 0.90
|
||||
page_size: 256
|
||||
max_running_requests: 256
|
||||
max_total_tokens: 8551168
|
||||
chunked_prefill_size: 8192
|
||||
swa_full_tokens_ratio: 0.1
|
||||
|
||||
|
||||
@@ -33,6 +33,7 @@ runtime:
|
||||
mem_fraction_static: 0.90
|
||||
page_size: 256
|
||||
max_running_requests: 256
|
||||
max_total_tokens: 8551168
|
||||
chunked_prefill_size: 8192
|
||||
swa_full_tokens_ratio: 0.1
|
||||
|
||||
|
||||
Reference in New Issue
Block a user