Skip to content

feat(llama.cpp): add flash_attention and no_kv_offloading #2310

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
mudler merged 1 commit into from
May 13, 2024
Merged

feat(llama.cpp): add flash_attention and no_kv_offloading #2310

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions backend/backend.proto
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -212,6 +212,9 @@ message ModelOptions {
float YarnBetaSlow = 47;

string Type = 49;

bool FlashAttention = 56;
bool NoKVOffload = 57;
}

message Result {
Expand Down
3 changes: 3 additions & 0 deletions backend/cpp/llama/grpc-server.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2254,6 +2254,9 @@ static void params_parse(const backend::ModelOptions* request,
}
params.use_mlock = request->mlock();
params.use_mmap = request->mmap();
params.flash_attn = request->flashattention();
params.no_kv_offload = request->nokvoffload();

params.embedding = request->embeddings();

if (request->ropescaling() == "none") { params.rope_scaling_type = LLAMA_ROPE_SCALING_TYPE_NONE; }
Expand Down
2 changes: 2 additions & 0 deletions core/backend/options.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,6 +77,8 @@ func gRPCModelOpts(c config.BackendConfig) *pb.ModelOptions {
MaxModelLen: int32(c.MaxModelLen),
TensorParallelSize: int32(c.TensorParallelSize),
MMProj: c.MMProj,
FlashAttention: c.FlashAttention,
NoKVOffload: c.NoKVOffloading,
YarnExtFactor: c.YarnExtFactor,
YarnAttnFactor: c.YarnAttnFactor,
YarnBetaFast: c.YarnBetaFast,
Expand Down
3 changes: 3 additions & 0 deletions core/config/backend_config.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -132,6 +132,9 @@ type LLMConfig struct {
TensorParallelSize int `yaml:"tensor_parallel_size"` // vLLM
MMProj string `yaml:"mmproj"`

FlashAttention bool `yaml:"flash_attention"`
NoKVOffloading bool `yaml:"no_kv_offloading"`

RopeScaling string `yaml:"rope_scaling"`
ModelType string `yaml:"type"`

Expand Down