vrr/kvcache-ai-ktransformers
2
0
Fork 0
mirror of https://github.com/kvcache-ai/ktransformers.git synced 2025-09-10 15:29:39 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

support smt and qlm4

Browse source
This commit is contained in:
djw 2025-07-25 12:48:51 +00:00
parent 712ad1fa3c
commit 48bc6185b5
9 changed files with 65 additions and 74 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ktransformers/optimize/optimize_rules/Glm4Moe-serve.yaml
View file

@ -60,7 +60,7 @@
- match:
name: "^model\\.layers\\..*\\.self_attn$"
replace:
class: ktransformers.operators.balance_serve_attention.KSmallthinkerAttention # optimized MLA implementation
class: ktransformers.operators.balance_serve_attention.KGlm4MoeAttention # optimized MLA implementation
kwargs:
generate_device: "cuda"
prefill_device: "cuda"