vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-10 00:54:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

WIP on sdcpp integration

Browse source
This commit is contained in:
Concedo 2024-02-29 00:40:07 +08:00
parent 1e3ac7d803
commit f75e479db0
12 changed files with 154 additions and 664 deletions
Download patch file Download diff file Expand all files Collapse all files

4
gpttype_adapter.cpp
View file

@ -949,7 +949,7 @@ ModelLoadResult gpttype_load_model(const load_model_inputs inputs, FileFormat in
#if defined(GGML_USE_CLBLAST)
if(file_format==FileFormat::GGUF_GENERIC && model_params.n_gpu_layers>0)
{
if(file_format_meta.model_architecture == GGUFArch::FALCON)
if(file_format_meta.model_architecture == GGUFArch::ARCH_FALCON)
{
printf("\nOpenCL does not support GPU Layer offloading for this model architecture! GPU Offload has been disabled.\n");
model_params.n_gpu_layers = 0;
@ -2032,7 +2032,7 @@ generation_outputs gpttype_generate(const generation_inputs inputs, generation_o
printf("\n(EOS token triggered!)");
}
remaining_tokens = 0;
last_stop_reason = stop_reason::EOS_TOKEN;
last_stop_reason = stop_reason::EOS_TOKEN_HIT;
}
for (const auto &matched : stop_sequence)