[CPU] Introduce Int4TinyGemmCpuTensor to replace Int4CPULayout in AQT

test/quantization/quantize_/workflows/int4/test_int4_tinygemm_cpu_tensor.py

@@ -0,0 +1,85 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD 3-Clause license found in the
+# LICENSE file in the root directory of this source tree.
+
+import tempfile
+import unittest
+
+import torch
+from torch.testing._internal.common_utils import (
+    TestCase,
+    instantiate_parametrized_tests,
+    parametrize,
+    run_tests,
+)
+
+from torchao.quantization import (
+    Int4WeightOnlyConfig,
+    quantize_,
+)
+from torchao.quantization.utils import compute_error
+from torchao.utils import (
+    torch_version_at_least,
+)
+
+
+def get_config(group_size):
+    return Int4WeightOnlyConfig(
+        group_size=group_size,
+        packing_format="int4_tinygemm_cpu",
+        version=2,
+    )
+
+
+@unittest.skipIf(not torch_version_at_least("2.6.0"), "Need pytorch 2.6+")
+class TestInt4TinyGemmCpuTensor(TestCase):
+    @parametrize(
+        "sizes",
+        [
+            ((128,), 256, 128),
+            ((32, 128), 512, 128),
+            ((2, 32, 128), 256, 12),
+        ],
+    )
+    @parametrize("dtype", [torch.float, torch.bfloat16, torch.half])
+    @parametrize("group_size", [32, 64, 128])
+    def test_linear(self, sizes, dtype, group_size):
+        device = "cpu"
+        M, N, K = sizes
+        input = torch.randn(*M, K, dtype=dtype, device=device)
+        linear = torch.nn.Linear(K, N, dtype=dtype, device=device)
+        original = linear(input)
+        quantize_(linear, get_config(group_size))
+        quantized = linear(input)
+        self.assertTrue(compute_error(original, quantized) > 20)
+
+        compiled_linear = torch.compile(linear)
+        quantized_and_compiled = compiled_linear(input)
+        self.assertTrue(compute_error(original, quantized_and_compiled) > 20)
+
+    @parametrize("dtype", [torch.float, torch.bfloat16, torch.half])
+    def test_module_path(self, dtype):
+        linear = torch.nn.Linear(128, 256, dtype=dtype)
+        quantize_(linear, get_config(group_size=128))
+        self.assertEqual(
+            str(type(linear.weight)),
+            "<class 'torchao.quantization.Int4TinyGemmCpuTensor'>",
+        )
+
+        with tempfile.NamedTemporaryFile() as f:
+            torch.save(linear.state_dict(), f)
+            f.seek(0)
+            state_dict = torch.load(f)
+            self.assertEqual(
+                str(type(state_dict["weight"])),
+                "<class 'torchao.quantization.Int4TinyGemmCpuTensor'>",
+            )
+
+
+instantiate_parametrized_tests(TestInt4TinyGemmCpuTensor)
+
+
+if __name__ == "__main__":
+    run_tests()

torchao/quantization/__init__.py

@@ -93,6 +93,7 @@
     Int4MarlinSparseTensor,
     Int4PreshuffledTensor,
     Int4Tensor,
+    Int4TinyGemmCpuTensor,
     IntxUnpackedTensor,
 )
 from .smoothquant import (
@@ -164,6 +165,7 @@
     "Int4MarlinSparseTensor",
     "IntxUnpackedTensor",
     "Float8Tensor",
+    "Int4TinyGemmCpuTensor",
     # smooth quant - subject to change
     "get_scale",
     "SmoothFakeDynQuantMixin",

torchao/quantization/quant_api.py

@@ -75,6 +75,7 @@
     Int4MarlinSparseTensor,
     Int4PreshuffledTensor,
     Int4Tensor,
+    Int4TinyGemmCpuTensor,
     IntxUnpackedTensor,
     QuantizeTensorToFloat8Kwargs,
 )
@@ -1080,6 +1081,12 @@ def _int4_weight_only_quantize_tensor(weight, config):
                 block_size,
             )
             return new_weight
+        elif packing_format == PackingFormat.INT4_TINYGEMM_CPU:
+            new_weight = Int4TinyGemmCpuTensor.from_hp(
+                weight,
+                block_size,
+            )
+            return new_weight
         else:
             raise ValueError(f"Unsupported packing format: {packing_format}")
 

torchao/quantization/quantize_/common/packing_format.py

@@ -40,3 +40,8 @@ class PackingFormat(str, Enum):
     Unpacked means the subbyte quantized data is stored as int8
     """
     UNPACKED_TO_INT8 = "unpacked_to_int8"
+
+    """
+    int4_tinygemm_cpu is referring to the format used by int4 weight-only quantization on CPU, which is a groupwise quantization format.
+    """
+    INT4_TINYGEMM_CPU = "int4_tinygemm_cpu"

torchao/quantization/quantize_/workflows/__init__.py

@@ -11,6 +11,9 @@
 from .int4.int4_tensor import (
     Int4Tensor,
 )
+from .int4.int4_tinygemm_cpu_tensor import (
+    Int4TinyGemmCpuTensor,
+)
 from .intx.intx_unpacked_tensor import (
     IntxUnpackedTensor,
 )
@@ -21,5 +24,6 @@
     "Int4MarlinSparseTensor",
     "Float8Tensor",
     "QuantizeTensorToFloat8Kwargs",
+    "Int4TinyGemmCpuTensor",
     "IntxUnpackedTensor",
 ]

torchao/quantization/quantize_/workflows/int4/int4_tinygemm_cpu_tensor.py

@@ -0,0 +1,194 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD 3-Clause license found in the
+# LICENSE file in the root directory of this source tree.
+
+
+from typing import List
+
+import torch
+
+from torchao.quantization.quant_primitives import (
+    MappingType,
+    _choose_qparams_affine_tinygemm,
+    _quantize_affine_tinygemm,
+)
+from torchao.utils import (
+    TorchAOBaseTensor,
+)
+
+__all__ = [
+    "Int4TinyGemmCpuTensor",
+]
+
+aten = torch.ops.aten
+
+
+class Int4TinyGemmCpuTensor(TorchAOBaseTensor):
+    """
+    int4 weight-only quantization on CPU with tinygemm (groupwise quantization only)
+
+    Tensor Attributes:
+        qdata: preshuffled and packed int4 weight for tinygemm, always viewed as a 2D (N, K/2) tensor, last dimension is packed
+               preshuffling is specific to CPU kernels, see Note below.
+        scale_and_zero: (K/group_size, N, 2), dtype is the same as the original Tensor dtype
+
+    Non-Tensor Attributes:
+        block_size: the block size for quantization, representing the granularity, for groupwise quantization, will have block_size (1, group_size).
+                    we only support group_size = 32/64/128.
+        shape: shape of the original Tensor
+
+    Note on Details for data layout for CPU tinygemm kernel:
+
+      We use AVX512 to compute TINYGEMM on CPU. We can also leverage AVX512_VNNI and AMX instructions with torch.compile and max-autotune.
+      For data locality, we preshuffle the data in plain layout (N, K/2) to (N/block_n, K, block_n/2), where block_n = 64/32/16.
+      See https://github.com/pytorch/pytorch/blob/32eee8ed225d9f10fbbcb38c24b8b44c24c0c97c/aten/src/ATen/native/cpu/int4mm_kernel.cpp#L583 for more details.
+    """
+
+    tensor_data_names = ["qdata", "scale_and_zero"]
+    tensor_attribute_names = ["block_size", "shape"]
+
+    def __new__(
+        cls,
+        qdata,
+        scale_and_zero,
+        block_size,
+        shape,
+    ):
+        kwargs = {}
+        kwargs["device"] = qdata.device
+        kwargs["dtype"] = scale_and_zero.dtype
+        kwargs["requires_grad"] = False
+        return torch.Tensor._make_wrapper_subclass(cls, shape, **kwargs)  # type: ignore[attr-defined]
+
+    def __init__(
+        self,
+        qdata: torch.Tensor,
+        scale_and_zero: torch.Tensor,
+        block_size: List[int],
+        shape: torch.Size,
+    ):
+        self.qdata = qdata
+        self.scale_and_zero = scale_and_zero
+        self.block_size = block_size
+
+    def _quantization_type(self):
+        return f"shape={self.shape}, block_size={self.block_size}, device={self.device}"
+
+    @classmethod
+    def from_hp(
+        cls,
+        w: torch.Tensor,
+        block_size: List[int],
+    ):
+        assert w.ndim == 2 and w.device.type == "cpu", (
+            f"Expecting 2D tensor on CPU, but got: {w.shape} on {w.device.type}"
+        )
+        assert len(block_size) == w.ndim
+        assert block_size[0] == 1 and block_size[1] in (32, 64, 128), (
+            f"Expecting groupwise quantization with group size = 32/64/128, but got block_size: {block_size}"
+        )
+        original_shape = w.shape
+        mapping_type = MappingType.ASYMMETRIC
+        target_dtype = torch.int32
+        quant_min = 0
+        quant_max = 15
+        eps = 1e-6
+        scale_dtype = None
+        zero_point_dtype = w.dtype
+        scale, zero_point = _choose_qparams_affine_tinygemm(
+            w,
+            mapping_type,
+            block_size,
+            target_dtype,
+            quant_min,
+            quant_max,
+            eps,
+            scale_dtype,
+            zero_point_dtype,
+        )
+        int_data = _quantize_affine_tinygemm(
+            w,
+            block_size,
+            scale,
+            zero_point,
+            target_dtype,
+            quant_min,
+            quant_max,
+        )
+        assert int_data.dtype == torch.int32, (
+            "torch.ops.aten._convert_weight_to_int4pack_for_cpu expects `int32` dtype"
+        )
+        packed_weight = torch.ops.aten._convert_weight_to_int4pack_for_cpu(
+            int_data,
+            1,  # innerKTiles is not needed for CPU
+        )
+
+        scale = scale.reshape(int_data.shape[0], -1)
+        zero_point = zero_point.reshape(int_data.shape[0], -1)
+        from torchao.quantization.utils import pack_tinygemm_scales_and_zeros
+
+        scale_and_zero = pack_tinygemm_scales_and_zeros(scale, zero_point, scale.dtype)
+        return Int4TinyGemmCpuTensor(
+            qdata=packed_weight,
+            scale_and_zero=scale_and_zero,
+            block_size=block_size,
+            shape=original_shape,
+        )
+
+
+implements = Int4TinyGemmCpuTensor.implements
+
+
+@implements([torch.nn.functional.linear, aten.linear.default])
+def _(func, types, args, kwargs):
+    input_tensor, weight_tensor, bias = (
+        args[0],
+        args[1],
+        args[2] if len(args) > 2 else None,
+    )
+    assert input_tensor.device.type == "cpu", (
+        f"For CPU device only but got: {input_tensor.device}"
+    )
+    assert isinstance(weight_tensor, Int4TinyGemmCpuTensor), (
+        f"Expected weight_tensor to be Int4TinyGemmCpuTensor, got: {type(weight_tensor)}"
+    )
+    assert weight_tensor.block_size[0] == 1, (
+        f"Requires groupwise quantization, got block_size: {weight_tensor.block_size}"
+    )
+    assert input_tensor.shape[-1] == weight_tensor.shape[1], (
+        f"Shapes of input and weight do not match, input:{input_tensor.shape}, weight: {weight_tensor.shape}"
+    )
+
+    act_mat = input_tensor
+    packed_weight = weight_tensor.qdata
+    scale_and_zero = weight_tensor.scale_and_zero
+
+    orig_act_size = act_mat.size()
+    orig_dtype = act_mat.dtype
+
+    # reshape to 2D
+    act_mat = act_mat.reshape(-1, act_mat.shape[-1])
+
+    # groupwise int4 quantization
+    groupsize = weight_tensor.block_size[1]
+    y = torch.ops.aten._weight_int4pack_mm_for_cpu(
+        act_mat.contiguous(), packed_weight, groupsize, scale_and_zero
+    )
+
+    # remove out_feature padding
+    assert weight_tensor.ndim == 2
+    orig_out_features = weight_tensor.shape[-2]
+    y = y[:, :orig_out_features]
+    y = y.reshape(*orig_act_size[:-1], orig_out_features)
+
+    if bias is not None:
+        y += bias
+    return y.to(orig_dtype)
+
+
+Int4TinyGemmCpuTensor.__module__ = "torchao.quantization"
+
+# Allow a model with Int4TinyGemmCpuTensor weights to be loaded with `weights_only=True`
+torch.serialization.add_safe_globals([Int4TinyGemmCpuTensor])