apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 21 Packages Projects Releases Wiki Activity

[TI-consistent] support quant use pow2scale (#7308)

Browse Source 
* support quant use pow2scale

* fix

* fix
This commit is contained in:
liuruyan
2026-04-13 15:01:53 +08:00
committed by GitHub
parent 6213ad5340
commit b34708604c
4 changed files with 24 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/envs.py
+14 -11
View File
@@ -212,17 +212,6 @@ environment_variables: dict[str, Callable[[], Any]] = {
"FD_XPU_MOE_FFN_QUANT_TYPE_MAP": lambda: os.getenv("FD_XPU_MOE_FFN_QUANT_TYPE_MAP", ""),
# Whether to enable low latency in mixed scenario
"FD_XPU_ENABLE_MIXED_EP_MODE": lambda: bool(int(os.getenv("FD_XPU_ENABLE_MIXED_EP_MODE", "0"))),
# Whether to use phi FP8 quantization,if 1,use paddle default.
"FD_USE_PHI_FP8_QUANT": lambda: bool(int(os.getenv("FD_USE_PHI_FP8_QUANT", "1"))),
# Enables the Paddle/phi combined TopK operator only when topk_method == noaux_tc,
# intended for training alignment. Defaults to 0 (disabled).
"FD_USE_PHI_MOE_TOPK": lambda: bool(int(os.getenv("FD_USE_PHI_MOE_TOPK", "0"))),
# Whether to use phi MOE permute,if 1,use paddle op.
"FD_USE_PHI_MOE_PERMUTE": lambda: bool(int(os.getenv("FD_USE_PHI_MOE_PERMUTE", "0"))),
# Whether to use phi rms_norm,if 1,use paddle op.
"FD_USE_PHI_RMSNORM": lambda: bool(int(os.getenv("FD_USE_PHI_RMSNORM", "0"))),
# Control class SiluAndMul to use swiglu or fusid_bias_act operator in the forward_cuda function
"FD_SiluAndMul_USE_PHI_SWIGLU": lambda: bool(int(os.getenv("FD_SiluAndMul_USE_PHI_SWIGLU", "0"))),
# Reserve output blocks for decoding requests when schedule new prefill requests
"FD_RESERVE_OUTPUT_BLOCK_NUM_FOR_DECODE_WHEN_SCHEDULE_NEW_PREFILL": lambda: int(
os.getenv("FD_RESERVE_OUTPUT_BLOCK_NUM_FOR_DECODE_WHEN_SCHEDULE_NEW_PREFILL", "16")
@@ -264,8 +253,22 @@ environment_variables: dict[str, Callable[[], Any]] = {
"FD_SAVE_OUTPUT_CACHE_FOR_PREEMPTED_REQUEST": lambda: bool(
int(os.getenv("FD_SAVE_OUTPUT_CACHE_FOR_PREEMPTED_REQUEST", "1"))
),
# train-infer consistency, used in RL
# Whether to align RoPE and moe gate precision with training
"FD_ENABLE_RL": lambda: int(os.getenv("FD_ENABLE_RL", "0")),
# Whether to use phi FP8 quantization,if 1,use paddle default.
"FD_USE_PHI_FP8_QUANT": lambda: bool(int(os.getenv("FD_USE_PHI_FP8_QUANT", "1"))),
# Enables the Paddle/phi combined TopK operator only when topk_method == noaux_tc,
# intended for training alignment. Defaults to 0 (disabled).
"FD_USE_PHI_MOE_TOPK": lambda: bool(int(os.getenv("FD_USE_PHI_MOE_TOPK", "0"))),
# Whether to use phi MOE permute,if 1,use paddle op.
"FD_USE_PHI_MOE_PERMUTE": lambda: bool(int(os.getenv("FD_USE_PHI_MOE_PERMUTE", "0"))),
# Whether to use phi rms_norm,if 1,use paddle op.
"FD_USE_PHI_RMSNORM": lambda: bool(int(os.getenv("FD_USE_PHI_RMSNORM", "0"))),
# Control class SiluAndMul to use swiglu or fusid_bias_act operator in the forward_cuda function
"FD_SiluAndMul_USE_PHI_SWIGLU": lambda: bool(int(os.getenv("FD_SiluAndMul_USE_PHI_SWIGLU", "0"))),
# Whether to enable FP8 quantization with pow2scale.
"FD_FP8_QUANT_WITH_POW2SCALE": lambda: bool(int(os.getenv("FD_FP8_QUANT_WITH_POW2SCALE", "0"))),
}