Build uploaded using `kernels`.

build/torch210-cxx11-cu126-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch210-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:49fd317d18b8b13367c70f037d1e8e3077aad8318d6dc40cd3050ab6f4e1d091
-size 712114272

build/torch210-cxx11-cu126-x86_64-linux/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _layer_norm_fd07706
-ops = torch.ops._layer_norm_fd07706
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_layer_norm_fd07706::{op_name}"

build/torch210-cxx11-cu126-x86_64-linux/layer_norm/__init__.py

@@ -1,26 +0,0 @@
-import ctypes
-import sys
-
-import importlib
-from pathlib import Path
-from types import ModuleType
-
-def _import_from_path(file_path: Path) -> ModuleType:
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path.absolute())).value)
-    module_name = path_hash
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None:
-        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    if module is None:
-        raise ImportError(f"Cannot load module {module_name} from spec")
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)  # type: ignore
-    return module
-
-
-globals().update(vars(_import_from_path(Path(__file__).parent.parent / "__init__.py")))

build/torch210-cxx11-cu126-x86_64-linux/layers.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class LayerNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = False,
-        )
-        return output[0].view(hidden_states.shape)
-
-class LlamaRMSNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = True,
-        )
-        return output[0].view(hidden_states.shape)

build/torch210-cxx11-cu126-x86_64-linux/metadata.json

@@ -1 +0,0 @@
-{"python-depends":[]}

build/torch210-cxx11-cu128-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch210-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a8f9c486fa147def1328121949fe502ba856d73e599a00844acf78faa8129cee
-size 1231439976

build/torch210-cxx11-cu128-x86_64-linux/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _layer_norm_fd07706
-ops = torch.ops._layer_norm_fd07706
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_layer_norm_fd07706::{op_name}"

build/torch210-cxx11-cu128-x86_64-linux/layer_norm/__init__.py

@@ -1,26 +0,0 @@
-import ctypes
-import sys
-
-import importlib
-from pathlib import Path
-from types import ModuleType
-
-def _import_from_path(file_path: Path) -> ModuleType:
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path.absolute())).value)
-    module_name = path_hash
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None:
-        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    if module is None:
-        raise ImportError(f"Cannot load module {module_name} from spec")
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)  # type: ignore
-    return module
-
-
-globals().update(vars(_import_from_path(Path(__file__).parent.parent / "__init__.py")))

build/torch210-cxx11-cu128-x86_64-linux/layers.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class LayerNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = False,
-        )
-        return output[0].view(hidden_states.shape)
-
-class LlamaRMSNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = True,
-        )
-        return output[0].view(hidden_states.shape)

build/torch210-cxx11-cu128-x86_64-linux/metadata.json

@@ -1 +0,0 @@
-{"python-depends":[]}

build/torch210-cxx11-cu130-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch210-cxx11-cu130-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:657b35fbbd096c4e34b804790484286941b781ef936fb920f9f1d10f7b0d4281
-size 1238357112

build/torch210-cxx11-cu130-x86_64-linux/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _layer_norm_fd07706
-ops = torch.ops._layer_norm_fd07706
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_layer_norm_fd07706::{op_name}"

build/torch210-cxx11-cu130-x86_64-linux/layer_norm/__init__.py

@@ -1,26 +0,0 @@
-import ctypes
-import sys
-
-import importlib
-from pathlib import Path
-from types import ModuleType
-
-def _import_from_path(file_path: Path) -> ModuleType:
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path.absolute())).value)
-    module_name = path_hash
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None:
-        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    if module is None:
-        raise ImportError(f"Cannot load module {module_name} from spec")
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)  # type: ignore
-    return module
-
-
-globals().update(vars(_import_from_path(Path(__file__).parent.parent / "__init__.py")))

build/torch210-cxx11-cu130-x86_64-linux/layers.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class LayerNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = False,
-        )
-        return output[0].view(hidden_states.shape)
-
-class LlamaRMSNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = True,
-        )
-        return output[0].view(hidden_states.shape)

build/torch210-cxx11-cu130-x86_64-linux/metadata.json

@@ -1 +0,0 @@
-{"python-depends":[]}

build/torch28-cxx11-cu126-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch28-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f4c4fce45ad6f08cfa1a3e2c7851c0964524975543a3e16b72406b6c8187bba4
-size 712034088

build/torch28-cxx11-cu126-x86_64-linux/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _layer_norm_fd07706
-ops = torch.ops._layer_norm_fd07706
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_layer_norm_fd07706::{op_name}"

build/torch28-cxx11-cu126-x86_64-linux/layer_norm/__init__.py

@@ -1,26 +0,0 @@
-import ctypes
-import sys
-
-import importlib
-from pathlib import Path
-from types import ModuleType
-
-def _import_from_path(file_path: Path) -> ModuleType:
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path.absolute())).value)
-    module_name = path_hash
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None:
-        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    if module is None:
-        raise ImportError(f"Cannot load module {module_name} from spec")
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)  # type: ignore
-    return module
-
-
-globals().update(vars(_import_from_path(Path(__file__).parent.parent / "__init__.py")))

build/torch28-cxx11-cu126-x86_64-linux/layers.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class LayerNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = False,
-        )
-        return output[0].view(hidden_states.shape)
-
-class LlamaRMSNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = True,
-        )
-        return output[0].view(hidden_states.shape)

build/torch28-cxx11-cu126-x86_64-linux/metadata.json

@@ -1 +0,0 @@
-{"python-depends":[]}

build/torch28-cxx11-cu128-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch28-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5821346938e86e0308c60fd072d54b57aba427aac75e354d3132dddc755ba125
-size 1231343024

build/torch28-cxx11-cu128-x86_64-linux/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _layer_norm_fd07706
-ops = torch.ops._layer_norm_fd07706
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_layer_norm_fd07706::{op_name}"

build/torch28-cxx11-cu128-x86_64-linux/layer_norm/__init__.py

@@ -1,26 +0,0 @@
-import ctypes
-import sys
-
-import importlib
-from pathlib import Path
-from types import ModuleType
-
-def _import_from_path(file_path: Path) -> ModuleType:
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path.absolute())).value)
-    module_name = path_hash
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None:
-        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    if module is None:
-        raise ImportError(f"Cannot load module {module_name} from spec")
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)  # type: ignore
-    return module
-
-
-globals().update(vars(_import_from_path(Path(__file__).parent.parent / "__init__.py")))

build/torch28-cxx11-cu128-x86_64-linux/layers.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class LayerNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = False,
-        )
-        return output[0].view(hidden_states.shape)
-
-class LlamaRMSNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = True,
-        )
-        return output[0].view(hidden_states.shape)

build/torch28-cxx11-cu128-x86_64-linux/metadata.json

@@ -1 +0,0 @@
-{"python-depends":[]}

build/torch28-cxx11-cu129-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch28-cxx11-cu129-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:43c278069ef7e766a8eae76c27b4c91a3e84065c4714f7d9e0d6ff8413732e7a
-size 1283038336

build/torch28-cxx11-cu129-x86_64-linux/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _layer_norm_fd07706
-ops = torch.ops._layer_norm_fd07706
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_layer_norm_fd07706::{op_name}"

build/torch28-cxx11-cu129-x86_64-linux/layer_norm/__init__.py

@@ -1,26 +0,0 @@
-import ctypes
-import sys
-
-import importlib
-from pathlib import Path
-from types import ModuleType
-
-def _import_from_path(file_path: Path) -> ModuleType:
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path.absolute())).value)
-    module_name = path_hash
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None:
-        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    if module is None:
-        raise ImportError(f"Cannot load module {module_name} from spec")
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)  # type: ignore
-    return module
-
-
-globals().update(vars(_import_from_path(Path(__file__).parent.parent / "__init__.py")))

build/torch28-cxx11-cu129-x86_64-linux/layers.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class LayerNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = False,
-        )
-        return output[0].view(hidden_states.shape)
-
-class LlamaRMSNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = True,
-        )
-        return output[0].view(hidden_states.shape)

build/torch28-cxx11-cu129-x86_64-linux/metadata.json

@@ -1 +0,0 @@
-{"python-depends":[]}

build/torch29-cxx11-cu126-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch29-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bc404a5e076466f49a0be4fa53652f2a7b40f1c611478ba8d1c4ef07c524815a
-size 712034248

build/torch29-cxx11-cu126-x86_64-linux/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _layer_norm_fd07706
-ops = torch.ops._layer_norm_fd07706
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_layer_norm_fd07706::{op_name}"

build/torch29-cxx11-cu126-x86_64-linux/layer_norm/__init__.py

@@ -1,26 +0,0 @@
-import ctypes
-import sys
-
-import importlib
-from pathlib import Path
-from types import ModuleType
-
-def _import_from_path(file_path: Path) -> ModuleType:
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path.absolute())).value)
-    module_name = path_hash
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None:
-        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    if module is None:
-        raise ImportError(f"Cannot load module {module_name} from spec")
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)  # type: ignore
-    return module
-
-
-globals().update(vars(_import_from_path(Path(__file__).parent.parent / "__init__.py")))

build/torch29-cxx11-cu126-x86_64-linux/layers.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class LayerNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = False,
-        )
-        return output[0].view(hidden_states.shape)
-
-class LlamaRMSNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = True,
-        )
-        return output[0].view(hidden_states.shape)

build/torch29-cxx11-cu126-x86_64-linux/metadata.json

@@ -1 +0,0 @@
-{"python-depends":[]}

build/torch29-cxx11-cu128-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch29-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:8da63d5fa4aeca09b5b5f1b3355c401fc516a15622637a2c65a03081fc55fdb3
-size 1231343160

build/torch29-cxx11-cu128-x86_64-linux/_ops.py

@@ -1,9 +0,0 @@
-import torch
-from . import _layer_norm_fd07706
-ops = torch.ops._layer_norm_fd07706
-
-def add_op_namespace_prefix(op_name: str):
-    """
-    Prefix op by namespace.
-    """
-    return f"_layer_norm_fd07706::{op_name}"

build/torch29-cxx11-cu128-x86_64-linux/layer_norm/__init__.py

@@ -1,26 +0,0 @@
-import ctypes
-import sys
-
-import importlib
-from pathlib import Path
-from types import ModuleType
-
-def _import_from_path(file_path: Path) -> ModuleType:
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path.absolute())).value)
-    module_name = path_hash
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None:
-        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    if module is None:
-        raise ImportError(f"Cannot load module {module_name} from spec")
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)  # type: ignore
-    return module
-
-
-globals().update(vars(_import_from_path(Path(__file__).parent.parent / "__init__.py")))

build/torch29-cxx11-cu128-x86_64-linux/layers.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-
-class LayerNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = False,
-        )
-        return output[0].view(hidden_states.shape)
-
-class LlamaRMSNorm(nn.Module):
-    weight: torch.Tensor
-    variance_epsilon: float
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        output = ops.dropout_add_ln_fwd(
-            hidden_states.view(-1, hidden_states.shape[-1]),
-            gamma = self.weight,
-            beta = None,
-            rowscale = None,
-            colscale = None,
-            x0_subset = None,
-            z_subset = None,
-            dropout_p = 0,
-            epsilon = self.variance_epsilon,
-            rowscale_const = 1.0,
-            z_numrows = hidden_states.shape[1],
-            gen = None,
-            residual_in_fp32 = False,
-            is_rms_norm = True,
-        )
-        return output[0].view(hidden_states.shape)

build/torch29-cxx11-cu128-x86_64-linux/metadata.json

@@ -1 +0,0 @@
-{"python-depends":[]}

build/torch29-cxx11-cu130-x86_64-linux/__init__.py

@@ -1,26 +0,0 @@
-import torch
-import torch.nn as nn
-
-from ._ops import ops
-
-from . import layers
-
-def dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_fwd(input, gamma, beta, rowscale, colscale, x0_subset, z_subset, dropout_p, epsilon, rowscale_const, z_numrows, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_bwd(dz, dx, x, mu, rsigma, gamma, rowscale, colscale, x0_subset, z_subset, dropout_p, rowscale_const, x0_numrows, has_residual, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_fwd(input, gamma0, beta0, gamma1, beta1, dropout_p, epsilon, gen, residual_in_fp32, is_rms_norm)
-
-def dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm):
-    return ops.dropout_add_ln_parallel_residual_bwd(dz0, dz1, dx, x, mu, rsigma, gamma0, gamma1, dropout_p, has_x1, has_residual, is_rms_norm)
-
-__all__ = [
-    "layers",
-    "dropout_add_ln_fwd",
-    "dropout_add_ln_bwd",
-    "dropout_add_ln_parallel_residual_fwd",
-    "dropout_add_ln_parallel_residual_bwd",
-]

build/torch29-cxx11-cu130-x86_64-linux/_layer_norm_fd07706.abi3.so

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7bf6e51b89bda807e770de087312693e67a4f215e8b036c39e92b6bd7de12ebb
-size 1238272584