vrr/kvcache-ai-ktransformers
2
0
Fork 0
mirror of https://github.com/kvcache-ai/ktransformers.git synced 2025-09-09 22:05:30 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

support smt and glm4

Browse source
This commit is contained in:
djw 2025-07-24 08:40:58 +00:00
parent 1677e90092
commit b66d96db97
18 changed files with 3519 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

3
ktransformers/operators/gate.py
View file

@ -212,4 +212,5 @@ class KMoEGateIPEXLLM(KMoEGate):
topk_idx, topk_weight = moe_group_topk(scores, self.orig_module.e_score_correction_bias,
self.n_group, self.topk_group, self.top_k,
self.norm_topk_prob, self.routed_scaling_factor)
return topk_idx, topk_weight.to(x.dtype)
return topk_idx, topk_weight.to(x.dtype)