mirrors/ollama-for-amd
1
0
Fork 0
mirror of https://github.com/likelovewant/ollama-for-amd.git synced 2025-12-21 22:33:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

model: implement bert in ollama engine (#9080)

Browse Source 
* fix truncate

* s/SentencePieceModel/SentencePiece/

* bert

* wordpiece

* refactor pooling

* more tokenizers

* normalize embeddings
This commit is contained in:
Michael Yang
2025-09-15 15:35:59 -07:00
committed by GitHub
parent 6f7117145f
commit 3f6642f6fc
16 changed files with 490 additions and 40 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
ml/backend/ggml/ggml.go
View File

@@ -1205,6 +1205,13 @@ func (t *Tensor) AddID(ctx ml.Context, t2, ids ml.Tensor) ml.Tensor {
}
}
func (t *Tensor) L2Norm(ctx ml.Context, eps float32) ml.Tensor {
return &Tensor{
b: t.b,
t: C.ggml_l2_norm(ctx.(*Context).ctx, t.t, C.float(eps)),
}
}
func (t *Tensor) LayerNorm(ctx ml.Context, w, b ml.Tensor, eps float32) ml.Tensor {
tt := C.ggml_norm(ctx.(*Context).ctx, t.t, C.float(eps))
if w != nil {