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

optimize gguf dequant, save mem, support Q2_K

Browse source 
use marlin for lm_head, lm_head only calc last token for prefill
extend context window to 19K for DeepSeek-V3/R1 within 24GB VRAM
This commit is contained in:
Atream 2025-02-22 06:13:01 +00:00
parent 7e1fe256c8
commit 5ec33d046d
27 changed files with 435 additions and 259 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ktransformers/optimize/optimize.py
View file

@ -126,6 +126,8 @@ def optimize_and_load_gguf(module: nn.Module, rule_file: str, gguf_path: str, mo
gguf_loader=GGUFLoader(gguf_path)
with torch.device("meta"):
inject(module, optimize_config, model_config, gguf_loader)
# pre load lm_head because its big inter result
load_weights(module.lm_head, gguf_loader, "lm_head.")
load_weights(module, gguf_loader)
module.gguf_loader = gguf_loader
del_meta(module)