tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-26 19:55:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

Use correct type of pooling for embedding models (#5500)

Browse Source 
Use correct type of pooling for embedding models
This commit is contained in:
Douglas Hanley
2024-02-15 11:21:49 -06:00
committed by GitHub
parent c06e45d729
commit 4524290e87
5 changed files with 94 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
gguf-py/gguf/gguf_writer.py
View File

@ -19,6 +19,7 @@ from .constants import (
GGUFValueType,
Keys,
RopeScalingType,
PoolingType,
TokenType,
)
@ -360,8 +361,8 @@ class GGUFWriter:
def add_causal_attention(self, value: bool) -> None:
self.add_bool(Keys.Attention.CAUSAL.format(arch=self.arch), value)
def add_pooling_layer(self, value: bool) -> None:
self.add_bool(Keys.LLM.POOLING_LAYER.format(arch=self.arch), value)
def add_pooling_type(self, value: PoolingType) -> None:
self.add_uint32(Keys.LLM.POOLING_TYPE.format(arch=self.arch), value)
def add_rope_dimension_count(self, count: int) -> None:
self.add_uint32(Keys.Rope.DIMENSION_COUNT.format(arch=self.arch), count)