vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 01:24:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fixing memory bugs

Browse source
This commit is contained in:
Concedo 2023-06-23 18:41:23 +08:00
parent e6ddb15c3a
commit df9135e3a9
5 changed files with 12 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

8
gpttype_adapter.cpp
View file

@ -308,8 +308,12 @@ ModelLoadResult gpttype_load_model(const load_model_inputs inputs, FileFormat in
params.memory_f16 = inputs.f16_kv;
params.n_ctx = inputs.max_context_length;
neox_ctx_v2.hparams.n_ctx = gptj_ctx_v1.hparams.n_ctx = gptj_ctx_v2.hparams.n_ctx = gpt2_ctx_v1.hparams.n_ctx = gpt2_ctx_v2.hparams.n_ctx
= neox_ctx_v3.hparams.n_ctx = gptj_ctx_v3.hparams.n_ctx = gptj_ctx_v3.hparams.n_ctx = mpt_ctx_v3.hparams.n_ctx = params.n_ctx;
neox_ctx_v2.hparams.n_ctx = neox_ctx_v3.hparams.n_ctx
= gptj_ctx_v1.hparams.n_ctx = gptj_ctx_v2.hparams.n_ctx = gptj_ctx_v3.hparams.n_ctx
= gpt2_ctx_v1.hparams.n_ctx = gpt2_ctx_v2.hparams.n_ctx = gpt2_ctx_v3.hparams.n_ctx
= mpt_ctx_v3.hparams.n_ctx = params.n_ctx;
bool calc_mem_with_scratch = ggml_cpu_has_gpublas();
printf("System Info: %s\n", llama_print_system_info());
SetQuantsUnshuffled(false);