tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-04 18:16:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

bitnet : replace 1.58b with b1.58, as in the paper

Browse Source
This commit is contained in:
Francis Couture-Harpin
2024-06-28 20:38:12 -04:00
parent 0996149911
commit bfd2f21fb4
5 changed files with 7 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
convert-hf-to-gguf.py
View File

@ -300,7 +300,7 @@ class Model:
if self.ftype != gguf.LlamaFileType.ALL_F32 and extra_f16 and not extra_f32:
# TODO: cleaner model-specific per-tensor types
# NOTE: Q1_3 is only relevant for BitNet 1.58b
# NOTE: Q1_3 is only relevant for BitNet b1.58
if (
self.ftype == gguf.LlamaFileType.MOSTLY_Q1_3
and gguf.can_quantize_to_q1_3(data)