v2: remove remaining logit scaling hacks from batch/optimized paths
This commit is contained in:
@@ -161,12 +161,6 @@ extension E4BModel {
|
||||
cmdBuf: cmdBuf
|
||||
)
|
||||
|
||||
// Logits scaling
|
||||
if embedWeight.groupSize == 32 && embedWeight.inDim == hiddenSize {
|
||||
let logitsScale = Float(30.0 / 116.23 / sqrt(Float(hiddenSize)))
|
||||
try scaleBufferOptimized(logitsBuffer, scale: logitsScale, count: vocabSize, cmdBuf: cmdBuf)
|
||||
}
|
||||
|
||||
// Softcapping
|
||||
if let cap = finalLogitSoftcapping {
|
||||
try applyLogitSoftcappingOptimized(
|
||||
|
||||
Reference in New Issue
Block a user