diff --git a/build/torch210-cxx11-cu126-x86_64-linux/__init__.py b/build/torch210-cxx11-cu126-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch210-cxx11-cu126-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch210-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..3e1b316855495addd39ea7ef89e2a628788b4478
--- /dev/null
+++ b/build/torch210-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49fd317d18b8b13367c70f037d1e8e3077aad8318d6dc40cd3050ab6f4e1d091
+size 712114272
diff --git a/build/torch210-cxx11-cu126-x86_64-linux/_ops.py b/build/torch210-cxx11-cu126-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch210-cxx11-cu126-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu126-x86_64-linux/layer_norm/__init__.py b/build/torch210-cxx11-cu126-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch210-cxx11-cu126-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch210-cxx11-cu126-x86_64-linux/layers.py b/build/torch210-cxx11-cu126-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch210-cxx11-cu126-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu126-x86_64-linux/metadata.json b/build/torch210-cxx11-cu126-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch210-cxx11-cu126-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu128-x86_64-linux/__init__.py b/build/torch210-cxx11-cu128-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch210-cxx11-cu128-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch210-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..bd5edf6c104e59f60b0b7984d42f3734f7e04e69
--- /dev/null
+++ b/build/torch210-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8f9c486fa147def1328121949fe502ba856d73e599a00844acf78faa8129cee
+size 1231439976
diff --git a/build/torch210-cxx11-cu128-x86_64-linux/_ops.py b/build/torch210-cxx11-cu128-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch210-cxx11-cu128-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu128-x86_64-linux/layer_norm/__init__.py b/build/torch210-cxx11-cu128-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch210-cxx11-cu128-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch210-cxx11-cu128-x86_64-linux/layers.py b/build/torch210-cxx11-cu128-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch210-cxx11-cu128-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu128-x86_64-linux/metadata.json b/build/torch210-cxx11-cu128-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch210-cxx11-cu128-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu130-x86_64-linux/__init__.py b/build/torch210-cxx11-cu130-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch210-cxx11-cu130-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu130-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch210-cxx11-cu130-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..74da187913202ae33734c1557bd953a88f3abfae
--- /dev/null
+++ b/build/torch210-cxx11-cu130-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:657b35fbbd096c4e34b804790484286941b781ef936fb920f9f1d10f7b0d4281
+size 1238357112
diff --git a/build/torch210-cxx11-cu130-x86_64-linux/_ops.py b/build/torch210-cxx11-cu130-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch210-cxx11-cu130-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu130-x86_64-linux/layer_norm/__init__.py b/build/torch210-cxx11-cu130-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch210-cxx11-cu130-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch210-cxx11-cu130-x86_64-linux/layers.py b/build/torch210-cxx11-cu130-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch210-cxx11-cu130-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch210-cxx11-cu130-x86_64-linux/metadata.json b/build/torch210-cxx11-cu130-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch210-cxx11-cu130-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu126-x86_64-linux/__init__.py b/build/torch28-cxx11-cu126-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch28-cxx11-cu126-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch28-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..998d160f15c98c3d90fe4a5c6a8c229c28ec5349
--- /dev/null
+++ b/build/torch28-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4c4fce45ad6f08cfa1a3e2c7851c0964524975543a3e16b72406b6c8187bba4
+size 712034088
diff --git a/build/torch28-cxx11-cu126-x86_64-linux/_ops.py b/build/torch28-cxx11-cu126-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch28-cxx11-cu126-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu126-x86_64-linux/layer_norm/__init__.py b/build/torch28-cxx11-cu126-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch28-cxx11-cu126-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch28-cxx11-cu126-x86_64-linux/layers.py b/build/torch28-cxx11-cu126-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch28-cxx11-cu126-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu126-x86_64-linux/metadata.json b/build/torch28-cxx11-cu126-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch28-cxx11-cu126-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu128-x86_64-linux/__init__.py b/build/torch28-cxx11-cu128-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch28-cxx11-cu128-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch28-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..76d9391c84580a3423af8099fbc29253cef21975
--- /dev/null
+++ b/build/torch28-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5821346938e86e0308c60fd072d54b57aba427aac75e354d3132dddc755ba125
+size 1231343024
diff --git a/build/torch28-cxx11-cu128-x86_64-linux/_ops.py b/build/torch28-cxx11-cu128-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch28-cxx11-cu128-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu128-x86_64-linux/layer_norm/__init__.py b/build/torch28-cxx11-cu128-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch28-cxx11-cu128-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch28-cxx11-cu128-x86_64-linux/layers.py b/build/torch28-cxx11-cu128-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch28-cxx11-cu128-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu128-x86_64-linux/metadata.json b/build/torch28-cxx11-cu128-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch28-cxx11-cu128-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu129-x86_64-linux/__init__.py b/build/torch28-cxx11-cu129-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch28-cxx11-cu129-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu129-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch28-cxx11-cu129-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..6de541efe55ef1a858397aa4f3ab8f239f3e7c70
--- /dev/null
+++ b/build/torch28-cxx11-cu129-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43c278069ef7e766a8eae76c27b4c91a3e84065c4714f7d9e0d6ff8413732e7a
+size 1283038336
diff --git a/build/torch28-cxx11-cu129-x86_64-linux/_ops.py b/build/torch28-cxx11-cu129-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch28-cxx11-cu129-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu129-x86_64-linux/layer_norm/__init__.py b/build/torch28-cxx11-cu129-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch28-cxx11-cu129-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch28-cxx11-cu129-x86_64-linux/layers.py b/build/torch28-cxx11-cu129-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch28-cxx11-cu129-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch28-cxx11-cu129-x86_64-linux/metadata.json b/build/torch28-cxx11-cu129-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch28-cxx11-cu129-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu126-x86_64-linux/__init__.py b/build/torch29-cxx11-cu126-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch29-cxx11-cu126-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch29-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..dec1e0b9a762965582fbf18968fb76454b839f2e
--- /dev/null
+++ b/build/torch29-cxx11-cu126-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc404a5e076466f49a0be4fa53652f2a7b40f1c611478ba8d1c4ef07c524815a
+size 712034248
diff --git a/build/torch29-cxx11-cu126-x86_64-linux/_ops.py b/build/torch29-cxx11-cu126-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch29-cxx11-cu126-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu126-x86_64-linux/layer_norm/__init__.py b/build/torch29-cxx11-cu126-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch29-cxx11-cu126-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch29-cxx11-cu126-x86_64-linux/layers.py b/build/torch29-cxx11-cu126-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch29-cxx11-cu126-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu126-x86_64-linux/metadata.json b/build/torch29-cxx11-cu126-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch29-cxx11-cu126-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu128-x86_64-linux/__init__.py b/build/torch29-cxx11-cu128-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch29-cxx11-cu128-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch29-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..09248c23d30665178d409dcb53bfcd274d522aaf
--- /dev/null
+++ b/build/torch29-cxx11-cu128-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8da63d5fa4aeca09b5b5f1b3355c401fc516a15622637a2c65a03081fc55fdb3
+size 1231343160
diff --git a/build/torch29-cxx11-cu128-x86_64-linux/_ops.py b/build/torch29-cxx11-cu128-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch29-cxx11-cu128-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu128-x86_64-linux/layer_norm/__init__.py b/build/torch29-cxx11-cu128-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch29-cxx11-cu128-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch29-cxx11-cu128-x86_64-linux/layers.py b/build/torch29-cxx11-cu128-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch29-cxx11-cu128-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu128-x86_64-linux/metadata.json b/build/torch29-cxx11-cu128-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch29-cxx11-cu128-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu130-x86_64-linux/__init__.py b/build/torch29-cxx11-cu130-x86_64-linux/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..946160f16b9dc91fefeea037fb7ac84fd6afd802
--- /dev/null
+++ b/build/torch29-cxx11-cu130-x86_64-linux/__init__.py
@@ -0,0 +1,26 @@
+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",
+]
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu130-x86_64-linux/_layer_norm_fd07706.abi3.so b/build/torch29-cxx11-cu130-x86_64-linux/_layer_norm_fd07706.abi3.so
new file mode 100644
index 0000000000000000000000000000000000000000..065ce5210251c015968eccb0e361a66d19a2fe31
--- /dev/null
+++ b/build/torch29-cxx11-cu130-x86_64-linux/_layer_norm_fd07706.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bf6e51b89bda807e770de087312693e67a4f215e8b036c39e92b6bd7de12ebb
+size 1238272584
diff --git a/build/torch29-cxx11-cu130-x86_64-linux/_ops.py b/build/torch29-cxx11-cu130-x86_64-linux/_ops.py
new file mode 100644
index 0000000000000000000000000000000000000000..58eafd610133ab10f4d6afed289b76a069f413c9
--- /dev/null
+++ b/build/torch29-cxx11-cu130-x86_64-linux/_ops.py
@@ -0,0 +1,9 @@
+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}"
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu130-x86_64-linux/layer_norm/__init__.py b/build/torch29-cxx11-cu130-x86_64-linux/layer_norm/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..03dbc1afe1cf156661a2b1b22003cd5f599a0309
--- /dev/null
+++ b/build/torch29-cxx11-cu130-x86_64-linux/layer_norm/__init__.py
@@ -0,0 +1,26 @@
+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")))
diff --git a/build/torch29-cxx11-cu130-x86_64-linux/layers.py b/build/torch29-cxx11-cu130-x86_64-linux/layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..7ed883f42ead452f8b60f498ec11302c53d3cf74
--- /dev/null
+++ b/build/torch29-cxx11-cu130-x86_64-linux/layers.py
@@ -0,0 +1,51 @@
+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)
\ No newline at end of file
diff --git a/build/torch29-cxx11-cu130-x86_64-linux/metadata.json b/build/torch29-cxx11-cu130-x86_64-linux/metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..76bafa5f33b6818aa6bb4cab04be811b87519b44
--- /dev/null
+++ b/build/torch29-cxx11-cu130-x86_64-linux/metadata.json
@@ -0,0 +1 @@
+{"python-depends":[]}
\ No newline at end of file