feat: update to include rev in kernel for reproducible symbols

README.md

@@ -2,9 +2,6 @@
 tags:
 - kernel
 ---
-
-![Status](https://hubwebhook.dholtz.com/shield?repo=kernels-community/activation)
-
 ## Activation
 
 Activation kernels from [vLLM](https://github.com/vllm-project/vllm/blob/main/csrc/activation_kernels.cu).

build.toml

@@ -8,8 +8,7 @@ src = [
 ]
 
 [kernel.activation]
-#language = "cuda-hipify"
-#rocm-archs = [ "gfx906", "gfx908", "gfx90a", "gfx940", "gfx941", "gfx942", "gfx1030", "gfx1100", "gfx1101" ]
+cuda-capabilities = [ "7.0", "7.2", "7.5", "8.0", "8.6", "8.7", "8.9", "9.0" ]
 src = [
   "activation/activation_kernels.cu",
   "activation/cuda_compat.h",

build/torch25-cxx11-cu118-x86_64-linux/activation/{_activation_78448fa.abi3.so → _activation_o63kkyjirmkf4.abi3.so}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:323dbf69b89390fd46b207abc1314a4cbe27491e1bb9f026c840bc3bff43b7d3
-size 2447952
+oid sha256:d50cdabfbed1df74e921ac34ff00bca0555977b14ef8082ddae7b1f30985a494
+size 2370160

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_o63kkyjirmkf4
+ops = torch.ops._activation_o63kkyjirmkf4
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_o63kkyjirmkf4::{op_name}"

build/torch25-cxx11-cu121-x86_64-linux/activation/{_activation_78448fa.abi3.so → _activation_vrl36m2ejer54.abi3.so}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6146ac6e77cbd458560bf67c46d93217833f2caf08260cc80a4aa62ba5645ee9
-size 2471056
+oid sha256:2bd0709ef09c8f0c18d1dc4a36c8096c59459bece61f5f5dbea95d1e73f54d44
+size 2393264

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_vrl36m2ejer54
+ops = torch.ops._activation_vrl36m2ejer54
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_vrl36m2ejer54::{op_name}"

build/torch25-cxx11-cu124-x86_64-linux/activation/{_activation_78448fa.abi3.so → _activation_va3moa75vw7c2.abi3.so}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:28eea3907055742f99bc9d7d4260add848adc2f6464e97029f37cd42a5c6bd0a
-size 2509832
+oid sha256:8353447f64e7d2df1a6a341d9c53bced53abef267f079923ae774170d0d57c53
+size 2427936

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_va3moa75vw7c2
+ops = torch.ops._activation_va3moa75vw7c2
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_va3moa75vw7c2::{op_name}"

build/torch25-cxx98-cu118-x86_64-linux/activation/{_activation_78448fa.abi3.so → _activation_qr3gs3eckeig4.abi3.so}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d5609ad07903b98c83c297bfb64f0d944df5edfe1c611fee23ec6c8fbd952604
-size 2440392
+oid sha256:df184a6315118d787a1bd6b435cb45f1ca7828445a1f1c0e55c57645cfbba43a
+size 2362600

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_qr3gs3eckeig4
+ops = torch.ops._activation_qr3gs3eckeig4
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_qr3gs3eckeig4::{op_name}"

build/torch25-cxx98-cu121-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7e6475ed603ad2cb565bd19ad2554484bd6c00d0d3f02decff60f2285df2546f
-size 2463232

build/torch25-cxx98-cu121-x86_64-linux/activation/_activation_p7gbzt25w3zg2.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccb13cfc2e45cf483e8b9f77f1760f28b48bcf185508d51b32d45bc759c4e8bb
+size 2385440

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_p7gbzt25w3zg2
+ops = torch.ops._activation_p7gbzt25w3zg2
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_p7gbzt25w3zg2::{op_name}"

build/torch25-cxx98-cu124-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a0767f6dba00c543d3cb77e2044bccd32ef569abc55b921231112c8a1ddfb187
-size 2502088

build/torch25-cxx98-cu124-x86_64-linux/activation/_activation_jg7yaigtn7wco.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f8048853e8cb06e8574a9a9497800d2be438f7989d79f44dcf2e0ced38a75a9
+size 2420192

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_jg7yaigtn7wco
+ops = torch.ops._activation_jg7yaigtn7wco
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_jg7yaigtn7wco::{op_name}"

build/torch26-cxx11-cu118-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e0c04d860454cc565113a3c93ff755fe9cbba0578c4604b89ad89e47c2503932
-size 2448056

build/torch26-cxx11-cu118-x86_64-linux/activation/_activation_ncisyrun7guwk.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cde5439e78ba0e1aaa1937d798b214b46d38cbab8e4384b93a22239fed1a4dd4
+size 2370264

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_ncisyrun7guwk
+ops = torch.ops._activation_ncisyrun7guwk
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_ncisyrun7guwk::{op_name}"

build/torch26-cxx11-cu124-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:48d7b0d190af1dd0366dbaeb0690b9c7cd1dfdc9aeda9b0b23bce56c70f5cbae
-size 2509928

build/torch26-cxx11-cu124-x86_64-linux/activation/_activation_ochhfvlnc3vyc.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6bd20d411c51fc8729b15cab6a60c5c9185222474aa035489e1bff299d76682
+size 2428040

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_ochhfvlnc3vyc
+ops = torch.ops._activation_ochhfvlnc3vyc
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_ochhfvlnc3vyc::{op_name}"

build/torch26-cxx11-cu126-aarch64-linux/activation/__init__.py

@@ -1,52 +0,0 @@
-import torch
-
-from ._ops import ops
-
-from . import layers
-
-
-def silu_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.silu_and_mul(out, x)
-    return out
-
-
-def gelu_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_and_mul(out, x)
-    return out
-
-
-def gelu_tanh_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_tanh_and_mul(out, x)
-    return out
-
-
-def fatrelu_and_mul(out: torch.Tensor, x: torch.Tensor, threshold: float = 0.0) -> None:
-    ops.fatrelu_and_mul(out, x, threshold)
-    return out
-
-
-def gelu_fast(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_fast(out, x)
-    return out
-
-
-def gelu_new(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_new(out, x)
-    return out
-
-
-def gelu_quick(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_quick(out, x)
-    return out
-
-
-__all__ = [
-    "silu_and_mul",
-    "gelu_and_mul",
-    "gelu_tanh_and_mul",
-    "fatrelu_and_mul",
-    "gelu_fast",
-    "gelu_new",
-    "gelu_quick",
-    "layers",
-]

build/torch26-cxx11-cu126-aarch64-linux/activation/_activation_bbdc1b4_dirty.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:558e4499ad3c09d02633488cfdc802a228b78a8cd51d963c92239d44744298c7
-size 2631936

build/torch26-cxx11-cu126-aarch64-linux/activation/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _activation_bbdc1b4_dirty
-ops = torch.ops._activation_bbdc1b4_dirty
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_activation_bbdc1b4_dirty::{op_name}"

build/torch26-cxx11-cu126-aarch64-linux/activation/layers.py

@@ -1,65 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class SiluAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.silu_and_mul(out, x)
-        return out
-
-
-class GeluAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.gelu_and_mul(out, x)
-        return out
-
-
-class GeluTanhAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.gelu_tanh_and_mul(out, x)
-        return out
-
-
-class FatreluAndMul(nn.Module):
-    def __init__(self, threshold: float = 0.0):
-        super().__init__()
-        self.threshold = threshold
-
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.fatrelu_and_mul(out, x, self.threshold)
-        return out
-
-
-class FastGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_fast(out, x)
-        return out
-
-
-class NewGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_new(out, x)
-        return out
-
-
-class QuickGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_quick(out, x)
-        return out

build/torch26-cxx11-cu126-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:11a11d0f4119edc5c637bab04ebd5669750a0e4f4000f58ab1bf5be2d8d9ab0b
-size 2518568

build/torch26-cxx11-cu126-x86_64-linux/activation/_activation_u6vnqubnicksq.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41c18b20c2bf8c49d2d3088a9bc1aad4293df0b57eafc9b141a9e8e595fe551a
+size 2436672

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_u6vnqubnicksq
+ops = torch.ops._activation_u6vnqubnicksq
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_u6vnqubnicksq::{op_name}"

build/torch26-cxx98-cu118-x86_64-linux/activation/_activation_2vn6ty3gfqfb6.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfbcd5da358cd5cb7982d19c8880cf4db6f08b46622a7a953f755ad59e4e1492
+size 2362752

build/torch26-cxx98-cu118-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:56dcc985761e309cbef3fc2a201f26e800583128d6e5a3fc1b23800fb0b8b48c
-size 2440544

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_2vn6ty3gfqfb6
+ops = torch.ops._activation_2vn6ty3gfqfb6
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_2vn6ty3gfqfb6::{op_name}"

build/torch26-cxx98-cu124-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:03c5f08322796d0736024412babe5d7f13bb1126387976ae12a80485a40d3883
-size 2502240

build/torch26-cxx98-cu124-x86_64-linux/activation/_activation_myvteedxdpqc6.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1bc928823117c800904bcd3492bf1a0c65a32f6d8a842dc039f55e29831ab49
+size 2420344

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_myvteedxdpqc6
+ops = torch.ops._activation_myvteedxdpqc6
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_myvteedxdpqc6::{op_name}"

build/torch26-cxx98-cu126-aarch64-linux/activation/__init__.py

@@ -1,52 +0,0 @@
-import torch
-
-from ._ops import ops
-
-from . import layers
-
-
-def silu_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.silu_and_mul(out, x)
-    return out
-
-
-def gelu_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_and_mul(out, x)
-    return out
-
-
-def gelu_tanh_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_tanh_and_mul(out, x)
-    return out
-
-
-def fatrelu_and_mul(out: torch.Tensor, x: torch.Tensor, threshold: float = 0.0) -> None:
-    ops.fatrelu_and_mul(out, x, threshold)
-    return out
-
-
-def gelu_fast(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_fast(out, x)
-    return out
-
-
-def gelu_new(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_new(out, x)
-    return out
-
-
-def gelu_quick(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_quick(out, x)
-    return out
-
-
-__all__ = [
-    "silu_and_mul",
-    "gelu_and_mul",
-    "gelu_tanh_and_mul",
-    "fatrelu_and_mul",
-    "gelu_fast",
-    "gelu_new",
-    "gelu_quick",
-    "layers",
-]

build/torch26-cxx98-cu126-aarch64-linux/activation/_activation_bbdc1b4_dirty.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f6afd50526ff4221cddd52cb947900cdf6bb95ad0a6bffcd1a86bda4d3f52349
-size 2628128

build/torch26-cxx98-cu126-aarch64-linux/activation/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _activation_bbdc1b4_dirty
-ops = torch.ops._activation_bbdc1b4_dirty
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_activation_bbdc1b4_dirty::{op_name}"

build/torch26-cxx98-cu126-aarch64-linux/activation/layers.py

@@ -1,65 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class SiluAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.silu_and_mul(out, x)
-        return out
-
-
-class GeluAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.gelu_and_mul(out, x)
-        return out
-
-
-class GeluTanhAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.gelu_tanh_and_mul(out, x)
-        return out
-
-
-class FatreluAndMul(nn.Module):
-    def __init__(self, threshold: float = 0.0):
-        super().__init__()
-        self.threshold = threshold
-
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.fatrelu_and_mul(out, x, self.threshold)
-        return out
-
-
-class FastGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_fast(out, x)
-        return out
-
-
-class NewGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_new(out, x)
-        return out
-
-
-class QuickGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_quick(out, x)
-        return out

build/torch26-cxx98-cu126-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f6eae5c895c564fbd2524ce488f4e91e65dc63402cd41a8bc74474b7437b2e62
-size 2506784

build/torch26-cxx98-cu126-x86_64-linux/activation/_activation_rbswus6emrhm2.abi3.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:474727e434a9cd4ec984a6da7124992ead4ca0fefce9581d0fd503e36c065aed
+size 2424888

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

@@ -1,9 +1,9 @@
 import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
+from . import _activation_rbswus6emrhm2
+ops = torch.ops._activation_rbswus6emrhm2
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_activation_78448fa::{op_name}"
+    return f"_activation_rbswus6emrhm2::{op_name}"

build/torch27-cxx11-cu118-x86_64-linux/activation/__init__.py

@@ -1,52 +0,0 @@
-import torch
-
-from ._ops import ops
-
-from . import layers
-
-
-def silu_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.silu_and_mul(out, x)
-    return out
-
-
-def gelu_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_and_mul(out, x)
-    return out
-
-
-def gelu_tanh_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_tanh_and_mul(out, x)
-    return out
-
-
-def fatrelu_and_mul(out: torch.Tensor, x: torch.Tensor, threshold: float = 0.0) -> None:
-    ops.fatrelu_and_mul(out, x, threshold)
-    return out
-
-
-def gelu_fast(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_fast(out, x)
-    return out
-
-
-def gelu_new(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_new(out, x)
-    return out
-
-
-def gelu_quick(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_quick(out, x)
-    return out
-
-
-__all__ = [
-    "silu_and_mul",
-    "gelu_and_mul",
-    "gelu_tanh_and_mul",
-    "fatrelu_and_mul",
-    "gelu_fast",
-    "gelu_new",
-    "gelu_quick",
-    "layers",
-]

build/torch27-cxx11-cu118-x86_64-linux/activation/_activation_78448fa.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f8086b2d9e0f2db80385b83e0bc28f8d158725d002e1613e1a46a87732197e9f
-size 2448152

build/torch27-cxx11-cu118-x86_64-linux/activation/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _activation_78448fa
-ops = torch.ops._activation_78448fa
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_activation_78448fa::{op_name}"

build/torch27-cxx11-cu118-x86_64-linux/activation/layers.py

@@ -1,65 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class SiluAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.silu_and_mul(out, x)
-        return out
-
-
-class GeluAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.gelu_and_mul(out, x)
-        return out
-
-
-class GeluTanhAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.gelu_tanh_and_mul(out, x)
-        return out
-
-
-class FatreluAndMul(nn.Module):
-    def __init__(self, threshold: float = 0.0):
-        super().__init__()
-        self.threshold = threshold
-
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.fatrelu_and_mul(out, x, self.threshold)
-        return out
-
-
-class FastGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_fast(out, x)
-        return out
-
-
-class NewGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_new(out, x)
-        return out
-
-
-class QuickGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_quick(out, x)
-        return out

build/torch27-cxx11-cu126-aarch64-linux/activation/__init__.py

@@ -1,52 +0,0 @@
-import torch
-
-from ._ops import ops
-
-from . import layers
-
-
-def silu_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.silu_and_mul(out, x)
-    return out
-
-
-def gelu_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_and_mul(out, x)
-    return out
-
-
-def gelu_tanh_and_mul(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_tanh_and_mul(out, x)
-    return out
-
-
-def fatrelu_and_mul(out: torch.Tensor, x: torch.Tensor, threshold: float = 0.0) -> None:
-    ops.fatrelu_and_mul(out, x, threshold)
-    return out
-
-
-def gelu_fast(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_fast(out, x)
-    return out
-
-
-def gelu_new(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_new(out, x)
-    return out
-
-
-def gelu_quick(out: torch.Tensor, x: torch.Tensor) -> None:
-    ops.gelu_quick(out, x)
-    return out
-
-
-__all__ = [
-    "silu_and_mul",
-    "gelu_and_mul",
-    "gelu_tanh_and_mul",
-    "fatrelu_and_mul",
-    "gelu_fast",
-    "gelu_new",
-    "gelu_quick",
-    "layers",
-]

build/torch27-cxx11-cu126-aarch64-linux/activation/_activation_bbdc1b4_dirty.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4210a6598f00b8921ecba1a0e24603eb05437a876ca1f473d2641e11d9a67ece
-size 2632160

build/torch27-cxx11-cu126-aarch64-linux/activation/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _activation_bbdc1b4_dirty
-ops = torch.ops._activation_bbdc1b4_dirty
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_activation_bbdc1b4_dirty::{op_name}"

build/torch27-cxx11-cu126-aarch64-linux/activation/layers.py

@@ -1,65 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class SiluAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.silu_and_mul(out, x)
-        return out
-
-
-class GeluAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.gelu_and_mul(out, x)
-        return out
-
-
-class GeluTanhAndMul(nn.Module):
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.gelu_tanh_and_mul(out, x)
-        return out
-
-
-class FatreluAndMul(nn.Module):
-    def __init__(self, threshold: float = 0.0):
-        super().__init__()
-        self.threshold = threshold
-
-    def forward(self, x: torch.Tensor):
-        d = x.shape[-1] // 2
-        output_shape = x.shape[:-1] + (d,)
-        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
-        ops.fatrelu_and_mul(out, x, self.threshold)
-        return out
-
-
-class FastGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_fast(out, x)
-        return out
-
-
-class NewGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_new(out, x)
-        return out
-
-
-class QuickGELU(nn.Module):
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        out = torch.empty_like(x)
-        ops.gelu_quick(out, x)
-        return out