Update .env
This commit is contained in:
@@ -6,7 +6,7 @@ SERVED_NAME=deepseek-coder
|
|||||||
TP_SIZE=4
|
TP_SIZE=4
|
||||||
DTYPE=float16
|
DTYPE=float16
|
||||||
MAX_MODEL_LEN=65536
|
MAX_MODEL_LEN=65536
|
||||||
GPU_MEM_UTIL=0.60 # 0.88
|
GPU_MEM_UTIL=0.60
|
||||||
MAX_NUM_SEQS=4
|
MAX_NUM_SEQS=4
|
||||||
VLLM_EXTRA_ARGS=--quantization gptq --disable-custom-all-reduce --enforce-eager
|
VLLM_EXTRA_ARGS=--quantization gptq --disable-custom-all-reduce --enforce-eager
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user