matt/ollama37
1
0
Fork 0
mirror of https://github.com/dogkeeper886/ollama37.git synced 2025-12-10 07:46:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

add gemma vision encoder

Browse Source
This commit is contained in:
Michael Yang
2025-03-06 12:16:54 -08:00
parent 5f74d1fd47
commit 4b037a97dc
10 changed files with 337 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ml/backend.go
View File

@@ -135,7 +135,9 @@ type Tensor interface {
RMSNorm(ctx Context, weight Tensor, eps float32) Tensor
Scale(ctx Context, s float64) Tensor
AvgPool1D(ctx Context, k, s, p int) Tensor
Conv2D(ctx Context, weight Tensor, s0, s1, p0, p1, d0, d1 int) Tensor
RoPE(ctx Context, positionIDs, ropeFactors Tensor, dim, ropeType uint32, base, scale float32) Tensor
Tanh(ctx Context) Tensor

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

@@ -947,6 +947,13 @@ func (t *Tensor) Conv2D(ctx ml.Context, t2 ml.Tensor, s0, s1, p0, p1, d0, d1 int
}
}
func (t *Tensor) AvgPool1D(ctx ml.Context, k, s, p int) ml.Tensor {
return &Tensor{
b: t.b,
t: C.ggml_pool_1d(ctx.(*Context).ctx, t.t, C.GGML_OP_POOL_AVG, C.int(k), C.int(s), C.int(p)),
}
}
func (t *Tensor) ScaledDotProductAttention(ctx ml.Context, key, value, mask ml.Tensor, scale float64) ml.Tensor {
var kqMask *C.struct_ggml_tensor
if mask != nil {