mirror of
https://github.com/kvcache-ai/ktransformers.git
synced 2026-04-29 12:19:50 +00:00
f36699affd
Complete SFT (Supervised Fine-Tuning) backend for MoE models using AMX SIMD: Core C++ implementation: - sft_moe.hpp: Forward/backward with LoRA fused operations (~5500 lines) - moe-sft-tp.hpp: Tensor-parallel wrapper for multi-NUMA - amx/moe-sft-tp.hpp: AMX-specific TP implementation - avx_kernels.hpp: AVX512 SIMD kernels for LoRA GEMM - amx_kernels.hpp: AMX tile kernels for Panel5 rank-outer optimization - worker_pool: RDTSC profiling, Chrome trace output, SFT timer infrastructure - ext_bindings.cpp: SFT MOE pybind bindings (BF16/INT8/INT4 + SkipLoRA variants) Python sft/ submodule (kt_kernel.sft): - base.py: BaseSFTMoEWrapper with buffer management (template method pattern) - amx.py: AMXSFTMoEWrapper (weight loading, C++ task construction) - autograd.py: KTMoEFunction (torch.autograd.Function for distributed training) - layer.py: KTMoELayerWrapper (nn.Module replacing HF MoE layers) - arch.py: MOEArchConfig (Qwen3/DeepSeek/Mixtral architecture detection) - weights.py: Expert weight extraction and checkpoint loading - lora.py: PEFT LoRA adaptation (view buffers, grad buffers, save/load adapter) - wrapper.py: wrap_moe_layers_with_kt_wrapper, load_kt_model, build_kt_device_map - config.py: KTConfig dataclass (DeepSpeed-style opaque config passthrough) - dist_utils.py: Distributed gather/scatter, checkpoint-phase detection Design decisions: - Rank-0-only expert pattern: only rank 0 holds C++ wrapper and expert weights - DeepSpeed-style integration: accelerate keeps only KTransformersPlugin (framework interaction fields), all logic in kt_kernel.sft - Inference isolation: importing kt_kernel does not load sft/ submodule - Old field name compatibility: _get_kt_config() converts kt_xxx→xxx automatically Verified: Qwen3-235B-A22B 4GPU AMXBF16 training, loss converges normally.
94 lines
3.2 KiB
Python
94 lines
3.2 KiB
Python
# KT-Kernel: High-performance kernel operations for KTransformers
|
|
# SPDX-License-Identifier: Apache-2.0
|
|
|
|
"""
|
|
KT-Kernel provides high-performance kernel operations for KTransformers,
|
|
including CPU-optimized MoE inference with AMX, AVX, and KML support.
|
|
|
|
The package automatically detects your CPU capabilities and loads the optimal
|
|
kernel variant (AMX, AVX512, or AVX2) at runtime.
|
|
|
|
Example usage:
|
|
>>> from kt_kernel import KTMoEWrapper
|
|
>>> wrapper = KTMoEWrapper(
|
|
... layer_idx=0,
|
|
... num_experts=8,
|
|
... num_experts_per_tok=2,
|
|
... hidden_size=4096,
|
|
... moe_intermediate_size=14336,
|
|
... num_gpu_experts=2,
|
|
... cpuinfer_threads=32,
|
|
... threadpool_count=2,
|
|
... weight_path="/path/to/weights",
|
|
... chunked_prefill_size=512,
|
|
... method="AMXINT4"
|
|
... )
|
|
|
|
Check which CPU variant is loaded:
|
|
>>> import kt_kernel
|
|
>>> print(kt_kernel.__cpu_variant__) # 'amx', 'avx512', or 'avx2'
|
|
|
|
Environment Variables:
|
|
KT_KERNEL_CPU_VARIANT: Override automatic detection ('amx', 'avx512', 'avx2')
|
|
KT_KERNEL_DEBUG: Enable debug output ('1' to enable)
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
# Detect CPU and load optimal extension variant
|
|
from ._cpu_detect import initialize as _initialize_cpu
|
|
|
|
_kt_kernel_ext, __cpu_variant__ = _initialize_cpu()
|
|
|
|
# Make the extension module available to other modules in this package
|
|
import sys
|
|
|
|
sys.modules["kt_kernel_ext"] = _kt_kernel_ext
|
|
|
|
# Also expose kt_kernel_ext as an attribute for backward compatibility
|
|
kt_kernel_ext = _kt_kernel_ext
|
|
|
|
# Import main API
|
|
from .experts import KTMoEWrapper
|
|
|
|
def __getattr__(name):
|
|
if name == "AMXSFTMoEWrapper":
|
|
try:
|
|
from .sft.amx import AMXSFTMoEWrapper
|
|
return AMXSFTMoEWrapper
|
|
except (ImportError, AttributeError):
|
|
return None
|
|
raise AttributeError(f"module 'kt_kernel' has no attribute {name!r}")
|
|
|
|
# Read version from package metadata (preferred) or fallback to project root
|
|
try:
|
|
# Try to get version from installed package metadata (works in installed environment)
|
|
from importlib.metadata import version, PackageNotFoundError
|
|
|
|
try:
|
|
__version__ = version("kt-kernel")
|
|
except PackageNotFoundError:
|
|
# Package not installed, try to read from source tree version.py
|
|
import os
|
|
|
|
_root_version_file = os.path.join(os.path.dirname(os.path.dirname(os.path.dirname(__file__))), "version.py")
|
|
if os.path.exists(_root_version_file):
|
|
_version_ns = {}
|
|
with open(_root_version_file, "r", encoding="utf-8") as f:
|
|
exec(f.read(), _version_ns)
|
|
__version__ = _version_ns.get("__version__", "0.4.3")
|
|
else:
|
|
__version__ = "0.4.3"
|
|
except ImportError:
|
|
# Python < 3.8, fallback to pkg_resources or hardcoded version
|
|
try:
|
|
from pkg_resources import get_distribution, DistributionNotFound
|
|
|
|
try:
|
|
__version__ = get_distribution("kt-kernel").version
|
|
except DistributionNotFound:
|
|
__version__ = "0.4.3"
|
|
except ImportError:
|
|
__version__ = "0.4.3"
|
|
|
|
__all__ = ["KTMoEWrapper", "AMXSFTMoEWrapper", "kt_kernel_ext", "__cpu_variant__", "__version__"]
|