vrr/kvcache-ai-ktransformers
2
0
Fork 0
mirror of https://github.com/kvcache-ai/ktransformers.git synced 2025-09-09 13:55:27 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Enable support for Intel XPU devices, add support for DeepSeek V2/V3 first

Browse source
This commit is contained in:
rnwang04 2025-05-14 14:28:22 +00:00
parent 333351c7c8
commit 142fb7ce6c
22 changed files with 673 additions and 81 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ktransformers/models/custom_cache.py
View file

@ -213,7 +213,7 @@ class KDeepSeekV3Cache(nn.Module):
self.v_caches = []
def load(self, inference_context: "sched_ext.InferenceContext"):
def load(self, inference_context: "sched_ext.InferenceContext"):
for i in range(self.config.num_hidden_layers):
self.k_caches.append(
@ -293,7 +293,7 @@ class KGQACache(nn.Module):
self.v_caches = []
def load(self, inference_context: sched_ext.InferenceContext):
def load(self, inference_context: "sched_ext.InferenceContext"):
print(self.config.num_hidden_layers)
for i in range(self.config.num_hidden_layers):
self.k_caches.append(