Build

build/torch25-cxx11-cu118-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch25-cxx11-cu118-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_v7rnpcck3kry4
+ops = torch.ops._quantization_eetq_v7rnpcck3kry4
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_v7rnpcck3kry4::{op_name}"

build/torch25-cxx11-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_v7rnpcck3kry4.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff073c997a4857cf7f88f5ca5b6a0e0aed3feb9739ba4187a3c1ec6fd2f1b64b
+size 28364752

build/torch25-cxx11-cu118-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch25-cxx11-cu121-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch25-cxx11-cu121-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_zcfiojfkx55be
+ops = torch.ops._quantization_eetq_zcfiojfkx55be
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_zcfiojfkx55be::{op_name}"

build/torch25-cxx11-cu121-x86_64-linux/quantization_eetq/_quantization_eetq_zcfiojfkx55be.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93f7cce612f7efed34ccfadc57ca6a17899feba7ba4780f33d4ff8f828171bfc
+size 27919784

build/torch25-cxx11-cu121-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch25-cxx11-cu124-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch25-cxx11-cu124-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_btymam4x7xvs6
+ops = torch.ops._quantization_eetq_btymam4x7xvs6
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_btymam4x7xvs6::{op_name}"

build/torch25-cxx11-cu124-x86_64-linux/quantization_eetq/_quantization_eetq_btymam4x7xvs6.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3e693ee1f395169b985b0dadab379cc5d2be0858ebd0caa3dedc5c6b8bd7a7d
+size 27950768

build/torch25-cxx11-cu124-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch25-cxx98-cu118-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch25-cxx98-cu118-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_yy3p6bsf622sq
+ops = torch.ops._quantization_eetq_yy3p6bsf622sq
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_yy3p6bsf622sq::{op_name}"

build/torch25-cxx98-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_yy3p6bsf622sq.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3c69a2ef282a7d50795a5656514227bbdef8e5153a2bbd2f277c78ea39de4cd
+size 28368280

build/torch25-cxx98-cu118-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch25-cxx98-cu121-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch25-cxx98-cu121-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_imijtykkseqze
+ops = torch.ops._quantization_eetq_imijtykkseqze
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_imijtykkseqze::{op_name}"

build/torch25-cxx98-cu121-x86_64-linux/quantization_eetq/_quantization_eetq_imijtykkseqze.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4111fb3304cba2975f11275f8444505208d4fc4b31da4ff3e3d508a3838ef64
+size 28063536

build/torch25-cxx98-cu121-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch25-cxx98-cu124-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch25-cxx98-cu124-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_4qerj3t7ddiry
+ops = torch.ops._quantization_eetq_4qerj3t7ddiry
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_4qerj3t7ddiry::{op_name}"

build/torch25-cxx98-cu124-x86_64-linux/quantization_eetq/_quantization_eetq_4qerj3t7ddiry.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:933e5f40627ba92496a0fbd14c4a602d0ea81cbb66ec2e4e870e692d87601b3f
+size 28110408

build/torch25-cxx98-cu124-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_j23ltbqvrnixg
+ops = torch.ops._quantization_eetq_j23ltbqvrnixg
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_j23ltbqvrnixg::{op_name}"

build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_j23ltbqvrnixg.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9de877da7bf0e968bb9d307539860efa86b12bcd15ac017acee2377f0241495a
+size 28369152

build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_p5neqtnhdgxv2
+ops = torch.ops._quantization_eetq_p5neqtnhdgxv2
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_p5neqtnhdgxv2::{op_name}"

build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/_quantization_eetq_p5neqtnhdgxv2.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20e49f06c5ea24b8f93e1d1200b87ab284cbdbab87495e3ae34a162375b5634e
+size 27951064

build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_idk3dezy35dfk
+ops = torch.ops._quantization_eetq_idk3dezy35dfk
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_idk3dezy35dfk::{op_name}"

build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/_quantization_eetq_idk3dezy35dfk.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0da474df2d93afd24827cd9ce04844e32bd0d742cbf60504c14f670e884ddc3c
+size 27998512

build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_fpjoxzd7nm2qa
+ops = torch.ops._quantization_eetq_fpjoxzd7nm2qa
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_fpjoxzd7nm2qa::{op_name}"

build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_fpjoxzd7nm2qa.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7fb594a0a3bcca4af967680d7b5865e6fcbed12bb9d59f8374bf4601516f7bc8
+size 28364528

build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_k7mlunxe2ye4s
+ops = torch.ops._quantization_eetq_k7mlunxe2ye4s
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_k7mlunxe2ye4s::{op_name}"

build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/_quantization_eetq_k7mlunxe2ye4s.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:865d3f89b2b8bbe79c7493416999cda076a106f0b3ecb5feb2664a2dee4d0fa5
+size 28114800

build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )

build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/__init__.py

@@ -0,0 +1,3 @@
+from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
+
+__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]

build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/_ops.py

@@ -0,0 +1,9 @@
+import torch
+from . import _quantization_eetq_7m7hz3sbwkaio
+ops = torch.ops._quantization_eetq_7m7hz3sbwkaio
+
+def add_op_namespace_prefix(op_name: str):
+    """
+    Prefix op by namespace.
+    """
+    return f"_quantization_eetq_7m7hz3sbwkaio::{op_name}"

build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/_quantization_eetq_7m7hz3sbwkaio.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f443eff55bcac99507a8a622333461922f6bcee610cab9334882628e58a54da6
+size 28162256

build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/custom_ops.py

@@ -0,0 +1,36 @@
+from typing import List
+import torch
+
+from ._ops import ops
+
+
+def w8_a16_gemm(
+    input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
+) -> torch.Tensor:
+    return ops.w8_a16_gemm(input, weight, scale)
+
+
+def w8_a16_gemm_(
+    input: torch.Tensor,
+    weight: torch.Tensor,
+    scale: torch.Tensor,
+    output: torch.Tensor,
+    m: int,
+    n: int,
+    k: int,
+) -> torch.Tensor:
+    return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
+
+
+def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
+    return ops.preprocess_weights(origin_weight, is_int4)
+
+
+def quant_weights(
+    origin_weight: torch.Tensor,
+    quant_type: torch.dtype,
+    return_unprocessed_quantized_tensor: bool,
+) -> List[torch.Tensor]:
+    return ops.quant_weights(
+        origin_weight, quant_type, return_unprocessed_quantized_tensor
+    )