Skip to content

Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

model: fix rotary embeddings for ministral 3 (#13432)

Jeffrey Morgan2dfb74410d2c
diff --git a/model/models/mistral3/model_text.go b/model/models/mistral3/model_text.go
index 01eca1c50..36106107b 100644
--- a/model/models/mistral3/model_text.go
+++ b/model/models/mistral3/model_text.go
@@ -29,24 +29,13 @@ type TextOptions struct {
func (o TextOptions) applyRotaryPositionEmbeddings(ctx ml.Context, states, positions ml.Tensor) ml.Tensor {
var ropeOpts []func(*rope.Options)
if o.ropeType == "yarn" {
- getMscale := func(scale, mscale float64) float64 {
- if scale <= 1.0 {
- return 1.0
- }
- return 0.1*mscale*math.Log(scale) + 1.0
- }
-
- var attnFactor float32
if o.ropeMscale != 0 && o.ropeMscaleAllDim != 0 {
- attnFactor = float32(getMscale(float64(o.ropeScale), float64(o.ropeMscale)) / getMscale(float64(o.ropeScale), float64(o.ropeMscaleAllDim)))
- } else {
- attnFactor = float32(getMscale(float64(o.ropeScale), 1))
+ ropeOpts = append(ropeOpts, rope.WithAttentionFactor(1.0/float32(0.1*math.Log(float64(o.ropeScale))+1.0)))
}
ropeOpts = append(ropeOpts,
rope.WithOriginalContextLength(o.ropeOrigPosEmbeddings),
rope.WithExtrapolationFactor(o.ropeExtrapolation),
- rope.WithAttentionFactor(attnFactor),
rope.WithBetaFast(o.ropeBetaFast),
rope.WithBetaSlow(o.ropeBetaSlow),
)