feat: add per-module automatic flash attention selection (#2132)

This commit is contained in:
leejet authored and GitHub committed 2026-10-13 22:58:45 +08:00
1 parent 60423ab5f6
commit d763dae3b8
53 files changed
+805 -125

No files matched your search

+8 -9
View File
@@ -166,15 +166,14 @@ struct SDContextParams {
std::string rpc_servers;
std::string effective_backend;
std::string effective_params_backend;
bool enable_mmap = false;
bool control_net_cpu = false;
bool clip_on_cpu = false;
bool vae_on_cpu = false;
bool flash_attn = false;
bool diffusion_flash_attn = false;
bool sage_attn = false;
bool diffusion_conv_direct = false;
bool vae_conv_direct = false;
bool enable_mmap = false;
bool control_net_cpu = false;
bool clip_on_cpu = false;
bool vae_on_cpu = false;
std::string flash_attn_mode = "auto";
bool sage_attn = false;
bool diffusion_conv_direct = false;
bool vae_conv_direct = false;
prediction_t prediction = PREDICTION_COUNT;
lora_apply_mode_t lora_apply_mode = LORA_APPLY_AUTO;