Kernels:
Trusted publisher
Remove builds incompatible with kernels >= 0.14
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/__init__.py +0 -3
- build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/_ops.py +0 -9
- build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/__init__.py +0 -3
- build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/_ops.py +0 -9
- build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch26-cxx11-cu126-aarch64-linux/quantization_eetq/__init__.py +0 -3
- build/torch26-cxx11-cu126-aarch64-linux/quantization_eetq/_ops.py +0 -9
- build/torch26-cxx11-cu126-aarch64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch26-cxx11-cu126-aarch64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/__init__.py +0 -3
- build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/_ops.py +0 -9
- build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/__init__.py +0 -3
- build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/_ops.py +0 -9
- build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/__init__.py +0 -3
- build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/_ops.py +0 -9
- build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch26-cxx98-cu126-aarch64-linux/quantization_eetq/__init__.py +0 -3
- build/torch26-cxx98-cu126-aarch64-linux/quantization_eetq/_ops.py +0 -9
- build/torch26-cxx98-cu126-aarch64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch26-cxx98-cu126-aarch64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/__init__.py +0 -3
- build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/_ops.py +0 -9
- build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/__init__.py +0 -3
- build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/__pycache__/__init__.cpython-313.pyc +0 -0
- build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/__pycache__/_ops.cpython-313.pyc +0 -0
- build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/__pycache__/custom_ops.cpython-313.pyc +0 -0
- build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/_ops.py +0 -9
- build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_92f3139.abi3.so +0 -3
- build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch27-cxx11-cu126-aarch64-linux/quantization_eetq/__init__.py +0 -3
- build/torch27-cxx11-cu126-aarch64-linux/quantization_eetq/_ops.py +0 -9
- build/torch27-cxx11-cu126-aarch64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so +0 -3
- build/torch27-cxx11-cu126-aarch64-linux/quantization_eetq/custom_ops.py +0 -36
- build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/__init__.py +0 -3
- build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/__pycache__/__init__.cpython-313.pyc +0 -0
- build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/__pycache__/_ops.cpython-313.pyc +0 -0
- build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/__pycache__/custom_ops.cpython-313.pyc +0 -0
- build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/_ops.py +0 -9
- build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/_quantization_eetq_92f3139.abi3.so +0 -3
- build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/custom_ops.py +0 -36
build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:09177b6bbe99ca3dadec283fcd1fbed831b5e150c6afc3ab10704ddceea95576
|
| 3 |
-
size 31359200
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu118-x86_64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:071573e21e6054145a3182514e04116df6866ae748c227fd2ea436153e4c6275
|
| 3 |
-
size 31002600
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu124-x86_64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu126-aarch64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu126-aarch64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu126-aarch64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:3c312731f3cfedd24fa364a2ea37ffbdbef283026d9970ece49558ab990b64dd
|
| 3 |
-
size 30886928
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu126-aarch64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:0910d73c92282fee4a95a03b12f5c7f5a014d27e452deed21425c6c959da509a
|
| 3 |
-
size 31029576
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx11-cu126-x86_64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:fce19321affc8e4734484925ac93862ba41fa5d192ef133710c07e82f1e344e4
|
| 3 |
-
size 31354584
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu118-x86_64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:7315f176ff319da68450bc7396fada660484741687b7b0c907baf4520c511219
|
| 3 |
-
size 31166344
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu124-x86_64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu126-aarch64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu126-aarch64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu126-aarch64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:e4aa8aceae17bbed8be4287c4cbc01f8249b385e260a139172a34aa12efecb89
|
| 3 |
-
size 31083056
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu126-aarch64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:c98e478af0de310d3588e72c9cb10e7c787413f3caae65843df3f6fa16a473e2
|
| 3 |
-
size 31189216
|
|
|
|
|
|
|
|
|
|
|
|
build/torch26-cxx98-cu126-x86_64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/__pycache__/__init__.cpython-313.pyc
DELETED
|
Binary file (334 Bytes)
|
|
|
build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/__pycache__/_ops.cpython-313.pyc
DELETED
|
Binary file (547 Bytes)
|
|
|
build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/__pycache__/custom_ops.cpython-313.pyc
DELETED
|
Binary file (1.84 kB)
|
|
|
build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_92f3139
|
| 3 |
-
ops = torch.ops._quantization_eetq_92f3139
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_92f3139::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/_quantization_eetq_92f3139.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:a87d3bf237f1e897b2b58d95a686ee4b5584329b6d04243eaf970681be00bb67
|
| 3 |
-
size 31977872
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu118-x86_64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu126-aarch64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu126-aarch64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_6337ee0
|
| 3 |
-
ops = torch.ops._quantization_eetq_6337ee0
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_6337ee0::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu126-aarch64-linux/quantization_eetq/_quantization_eetq_6337ee0.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:85f2f79a15fee959addb101180c59e6f67f9739b795d6bf5ce696db9760b2807
|
| 3 |
-
size 30887136
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu126-aarch64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/__init__.py
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
from .custom_ops import w8_a16_gemm, w8_a16_gemm_, preprocess_weights, quant_weights
|
| 2 |
-
|
| 3 |
-
__all__ = ["w8_a16_gemm", "w8_a16_gemm_", "preprocess_weights", "quant_weights"]
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/__pycache__/__init__.cpython-313.pyc
DELETED
|
Binary file (334 Bytes)
|
|
|
build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/__pycache__/_ops.cpython-313.pyc
DELETED
|
Binary file (547 Bytes)
|
|
|
build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/__pycache__/custom_ops.cpython-313.pyc
DELETED
|
Binary file (1.84 kB)
|
|
|
build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/_ops.py
DELETED
|
@@ -1,9 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
from . import _quantization_eetq_92f3139
|
| 3 |
-
ops = torch.ops._quantization_eetq_92f3139
|
| 4 |
-
|
| 5 |
-
def add_op_namespace_prefix(op_name: str):
|
| 6 |
-
"""
|
| 7 |
-
Prefix op by namespace.
|
| 8 |
-
"""
|
| 9 |
-
return f"_quantization_eetq_92f3139::{op_name}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/_quantization_eetq_92f3139.abi3.so
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:3a0d40ed3d499de55afde610e6bdb0c08e3ed40128c6f82984a1fdfbd2a5baae
|
| 3 |
-
size 31639912
|
|
|
|
|
|
|
|
|
|
|
|
build/torch27-cxx11-cu126-x86_64-linux/quantization_eetq/custom_ops.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
from typing import List
|
| 2 |
-
import torch
|
| 3 |
-
|
| 4 |
-
from ._ops import ops
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
def w8_a16_gemm(
|
| 8 |
-
input: torch.Tensor, weight: torch.Tensor, scale: torch.Tensor
|
| 9 |
-
) -> torch.Tensor:
|
| 10 |
-
return ops.w8_a16_gemm(input, weight, scale)
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
def w8_a16_gemm_(
|
| 14 |
-
input: torch.Tensor,
|
| 15 |
-
weight: torch.Tensor,
|
| 16 |
-
scale: torch.Tensor,
|
| 17 |
-
output: torch.Tensor,
|
| 18 |
-
m: int,
|
| 19 |
-
n: int,
|
| 20 |
-
k: int,
|
| 21 |
-
) -> torch.Tensor:
|
| 22 |
-
return ops.w8_a16_gemm_(input, weight, scale, output, m, n, k)
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
def preprocess_weights(origin_weight: torch.Tensor, is_int4: bool) -> torch.Tensor:
|
| 26 |
-
return ops.preprocess_weights(origin_weight, is_int4)
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def quant_weights(
|
| 30 |
-
origin_weight: torch.Tensor,
|
| 31 |
-
quant_type: torch.dtype,
|
| 32 |
-
return_unprocessed_quantized_tensor: bool,
|
| 33 |
-
) -> List[torch.Tensor]:
|
| 34 |
-
return ops.quant_weights(
|
| 35 |
-
origin_weight, quant_type, return_unprocessed_quantized_tensor
|
| 36 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|