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

64
ktransformers/ktransformers_ext/cuda/binding.cpp
View file

@ -1,10 +1,8 @@
/**
* @Description :
* @Author : Azure-Tang
* @Author : Azure-Tang, Boxin Zhang
* @Date : 2024-07-25 13:38:30
* @Version : 1.0.0
* @LastEditors : kkk1nak0
* @LastEditTime : 2024-08-12 03:05:04
* @Version : 0.2.2
* @Copyright (c) 2024 by KVCache.AI, All Rights Reserved.
**/
@ -19,22 +17,44 @@
// namespace py = pybind11;
PYBIND11_MODULE(KTransformersOps, m) {
m.def("dequantize_q8_0", &dequantize_q8_0, "Function to dequantize q8_0 data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q6_k", &dequantize_q6_k, "Function to dequantize q6_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q5_k", &dequantize_q5_k, "Function to dequantize q5_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q4_k", &dequantize_q4_k, "Function to dequantize q4_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q3_k", &dequantize_q3_k, "Function to dequantize q3_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q2_k", &dequantize_q2_k, "Function to dequantize q2_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_iq4_xs", &dequantize_iq4_xs, "Function to dequantize iq4_xs data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("device"), py::arg("target_dtype"));
m.def("gptq_marlin_gemm", &gptq_marlin_gemm, "Function to perform GEMM using Marlin quantization.",
py::arg("a"), py::arg("b_q_weight"), py::arg("b_scales"), py::arg("g_idx"),
py::arg("perm"), py::arg("workspace"), py::arg("num_bits"), py::arg("size_m"),
py::arg("size_n"), py::arg("size_k"), py::arg("is_k_full"));
m.def("dequantize_q8_0", [](const intptr_t data, int num_bytes, int blk_size, const int ele_per_blk, torch::Device device, torch::Dtype target_dtype) {
return dequantize_q8_0((int8_t*)data, num_bytes, blk_size, ele_per_blk, device, target_dtype);
}, "Function to dequantize q8_0 data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("ele_per_blk"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q6_k", [](const intptr_t data, int num_bytes, int blk_size, const int ele_per_blk, torch::Device device, torch::Dtype target_dtype) {
return dequantize_q6_k((int8_t*)data, num_bytes, blk_size, ele_per_blk, device, target_dtype);
}, "Function to dequantize q6_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("ele_per_blk"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q5_k", [](const intptr_t data, int num_bytes, int blk_size, const int ele_per_blk, torch::Device device, torch::Dtype target_dtype) {
return dequantize_q5_k((int8_t*)data, num_bytes, blk_size, ele_per_blk, device, target_dtype);
}, "Function to dequantize q5_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("ele_per_blk"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q4_k", [](const intptr_t data, int num_bytes, int blk_size, const int ele_per_blk, torch::Device device, torch::Dtype target_dtype) {
return dequantize_q4_k((int8_t*)data, num_bytes, blk_size, ele_per_blk, device, target_dtype);
}, "Function to dequantize q4_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("ele_per_blk"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q3_k", [](const intptr_t data, int num_bytes, int blk_size, const int ele_per_blk, torch::Device device, torch::Dtype target_dtype) {
return dequantize_q3_k((int8_t*)data, num_bytes, blk_size, ele_per_blk, device, target_dtype);
}, "Function to dequantize q3_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("ele_per_blk"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_q2_k", [](const intptr_t data, int num_bytes, int blk_size, const int ele_per_blk, torch::Device device, torch::Dtype target_dtype) {
return dequantize_q2_k((int8_t*)data, num_bytes, blk_size, ele_per_blk, device, target_dtype);
}, "Function to dequantize q2_k data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("ele_per_blk"), py::arg("device"), py::arg("target_dtype"));
m.def("dequantize_iq4_xs", [](const intptr_t data, int num_bytes, int blk_size, const int ele_per_blk, torch::Device device, torch::Dtype target_dtype) {
return dequantize_iq4_xs((int8_t*)data, num_bytes, blk_size, ele_per_blk, device, target_dtype);
}, "Function to dequantize iq4_xs data.",
py::arg("data"), py::arg("num_bytes"), py::arg("blk_size"), py::arg("ele_per_blk"), py::arg("device"), py::arg("target_dtype"));
m.def("gptq_marlin_gemm", &gptq_marlin_gemm, "Function to perform GEMM using Marlin quantization.",
py::arg("a"), py::arg("b_q_weight"), py::arg("b_scales"), py::arg("g_idx"),
py::arg("perm"), py::arg("workspace"), py::arg("num_bits"), py::arg("size_m"),
py::arg("size_n"), py::arg("size_k"), py::arg("is_k_full"));
}