tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-28 20:25:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : remove cfg smooth factor as it is only a reparameterization of the guidance scale (#2280)

Browse Source
This commit is contained in:
Guillaume "Vermeille" Sanchez
2023-07-21 12:58:36 +02:00
committed by GitHub
parent 73643f5fb1
commit ab0e26bdfb
5 changed files with 4 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
examples/common.h
View File

@ -55,7 +55,6 @@ struct gpt_params {
// https://arxiv.org/abs/2306.17806
std::string cfg_negative_prompt; // string to help guidance
float cfg_scale = 1.f; // How strong is guidance
float cfg_smooth_factor = 1.f; // Smooth factor between old and new logits
std::string model = "models/7B/ggml-model.bin"; // model path
std::string model_alias = "unknown"; // model alias