mirror of
https://github.com/kvcache-ai/ktransformers.git
synced 2025-09-04 03:29:49 +00:00
5ec33d046d
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
59 lines
No EOL
1.7 KiB
YAML
59 lines
No EOL
1.7 KiB
YAML
- match:
|
|
class: ktransformers.models.modeling_mixtral.MixtralRotaryEmbedding
|
|
replace:
|
|
class: ktransformers.operators.RoPE.RotaryEmbedding
|
|
kwargs:
|
|
generate_device: "cuda"
|
|
prefill_device: "cuda"
|
|
- match:
|
|
name: "^model\\.layers\\..*$"
|
|
class: torch.nn.Linear # only match modules matching name and class simultaneously
|
|
replace:
|
|
class: ktransformers.operators.linear.KTransformersLinear # optimized Kernel on quantized data types
|
|
kwargs:
|
|
generate_device: "cuda"
|
|
prefill_device: "cuda"
|
|
generate_op: "KLinearMarlin"
|
|
prefill_op: "KLinearTorch"
|
|
- match:
|
|
name: "^lm_head"
|
|
class: torch.nn.Linear
|
|
replace:
|
|
class: ktransformers.operators.linear.KTransformersLinear
|
|
kwargs:
|
|
generate_device: "cuda"
|
|
prefill_device: "cuda"
|
|
generate_op: "KLinearMarlin"
|
|
prefill_op: "KLinearTorch"
|
|
- match:
|
|
name: "^model\\.layers\\..*\\.block_sparse_moe$"
|
|
class: ktransformers.models.modeling_mixtral.MixtralSparseMoeBlock
|
|
replace:
|
|
class: ktransformers.operators.experts.KMistralSparseMoEBlock
|
|
- match:
|
|
name: "^model\\.layers\\..*\\.block_sparse_moe\\.experts$"
|
|
replace:
|
|
class: ktransformers.operators.experts.KTransformersExperts
|
|
kwargs:
|
|
prefill_device: "cuda"
|
|
prefill_op: "KExpertsTorch"
|
|
generate_device: "cpu"
|
|
generate_op: "KExpertsCPU"
|
|
out_device: "cuda"
|
|
recursive: False # don't recursively inject submodules of this module
|
|
|
|
- match:
|
|
name: "^model.embed_tokens"
|
|
replace:
|
|
class: "default"
|
|
kwargs:
|
|
generate_device: "cpu"
|
|
prefill_device: "cpu"
|
|
|
|
- match:
|
|
name: "^model\\.layers\\..*\\."
|
|
replace:
|
|
class: "default"
|
|
kwargs:
|
|
generate_device: "cuda"
|
|
prefill_device: "cuda" |