vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-07-10 01:18:32 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix incorrect mtp layers setting (+1 squashed commits)

Browse source 
Squashed commits:

[8dad1a5c0] fix incorrect mtp layers setting
This commit is contained in:
Concedo 2026-06-18 21:09:35 +08:00
parent 6591c33667
commit 635c45e1a0
2 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

2
gpttype_adapter.cpp
View file

@ -893,7 +893,7 @@ static void mtp_decoding_setup(llama_model * main_model, llama_context * main_ct
}
draft_is_mtp = true;
speculative_state_setup(main_ctx, mtp_ctx_params, 0, COMMON_SPECULATIVE_TYPE_DRAFT_MTP);
speculative_state_setup(main_ctx, mtp_ctx_params, static_cast<int32_t>(main_model->n_gpu_layers()), COMMON_SPECULATIVE_TYPE_DRAFT_MTP);
}
//loads a model for speculative decoding.

2
koboldcpp.py
View file

@ -78,7 +78,7 @@ dry_seq_break_max = 128
extra_images_max = 4 # for kontext/qwen img
# global vars
KcppVersion = "1.115.2"
KcppVersion = "1.116"
showdebug = True
kcpp_instance = None #global running instance
global_memory = {"tunnel_url": "", "restart_target":"", "input_to_exit":False, "load_complete":False, "restart_override_base_config":"", "last_active_timestamp":datetime.now(), "triggered_sleeping":False, "current_model":"initial_model", "base_config":"", "swapReqType": None, "autoswapmode": False}