mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2025-11-02 09:12:03 +00:00
server : add SWA checkpoints (#15293)
* server : add SWA checkpoints ggml-ci * cont : server clean-up * server : handle state restore fails * llama : add extended llama_state_seq_ API * server : do not make checkpoints if --swa-full ggml-ci * llama : remove flags value for NONE * server : configure number of SWA checkpoints with CLI arg ggml-ci * args : fix scope of new argument
This commit is contained in:
@@ -111,9 +111,9 @@ struct llama_context {
|
||||
size_t state_get_data( uint8_t * dst, size_t size);
|
||||
size_t state_set_data(const uint8_t * src, size_t size);
|
||||
|
||||
size_t state_seq_get_size(llama_seq_id seq_id);
|
||||
size_t state_seq_get_data(llama_seq_id seq_id, uint8_t * dst, size_t size);
|
||||
size_t state_seq_set_data(llama_seq_id seq_id, const uint8_t * src, size_t size);
|
||||
size_t state_seq_get_size(llama_seq_id seq_id, llama_state_seq_flags flags);
|
||||
size_t state_seq_get_data(llama_seq_id seq_id, uint8_t * dst, size_t size, llama_state_seq_flags flags);
|
||||
size_t state_seq_set_data(llama_seq_id seq_id, const uint8_t * src, size_t size, llama_state_seq_flags flags);
|
||||
|
||||
bool state_load_file(
|
||||
const char * filepath,
|
||||
@@ -213,8 +213,8 @@ private:
|
||||
size_t state_write_data(llama_io_write_i & io);
|
||||
size_t state_read_data (llama_io_read_i & io);
|
||||
|
||||
size_t state_seq_write_data(llama_io_write_i & io, llama_seq_id seq_id);
|
||||
size_t state_seq_read_data (llama_io_read_i & io, llama_seq_id seq_id);
|
||||
size_t state_seq_write_data(llama_io_write_i & io, llama_seq_id seq_id, llama_state_seq_flags flags);
|
||||
size_t state_seq_read_data (llama_io_read_i & io, llama_seq_id seq_id, llama_state_seq_flags flags);
|
||||
|
||||
//
|
||||
// members
|
||||
|
||||
Reference in New Issue
Block a user