modular/max/nn/__init__.py at main · ScottzCodez/modular · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
# ===----------------------------------------------------------------------=== #
# Copyright (c) 2025, Modular Inc. All rights reserved.
#
# Licensed under the Apache License v2.0 with LLVM Exceptions:
# https://llvm.org/LICENSE.txt
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ===----------------------------------------------------------------------=== #

from .attention import (
    Attention,
    AttentionImpl,
    AttentionImplQKV,
    AttentionQKV,
    AttentionWithRope,
    AttentionWithRopeQKV,
    AttentionWithRopeV1,
    DistributedAttentionImpl,
    DistributedAttentionWithRope,
    GGUFQAttentionWithRope,
    GPTQAttentionWithRope,
    RaggedAttention,
)
from .clamp import clamp
from .comm import Allreduce, Signals
from .conv import Conv1D, Conv1DV1, Conv2DV1, Conv3D, Conv3DV1
from .conv_transpose import ConvTranspose1d, WeightNormConvTranspose1d
from .embedding import Embedding, EmbeddingV1, VocabParallelEmbedding
from .kernels import MHAMaskVariant
from .layer import Layer, LayerList, Module
from .linear import (
    MLP,
    MLPV1,
    ColumnParallelLinear,
    DistributedMLP,
    Float8Config,
    Float8InputScaleSpec,
    Float8ScaleGranularity,
    Float8ScaleOrigin,
    Float8WeightScaleSpec,
    GPTQLinear,
    Linear,
    LinearV1,
)
from .norm import (
    DistributedRMSNorm,
    GroupNorm,
    LayerNorm,
    LayerNormV1,
    RMSNorm,
    RMSNormV1,
)
from .rotary_embedding import (
    LinearScalingParams,
    Llama3RopeScalingParams,
    Llama3RotaryEmbedding,
    OptimizedRotaryEmbedding,
    RotaryEmbedding,
)
from .sequential import Sequential
from .transformer import (
    DistributedTransformer,
    DistributedTransformerBlock,
    ReturnLogits,
    Transformer,
    TransformerBlock,
)

__all__ = [
    "Allreduce",
    "Attention",
    "AttentionImpl",
    "AttentionImplQKV",
    "AttentionQKV",
    "AttentionWithRopeV1",
    "AttentionWithRopeQKV",
    "AttentionWithRope",
    "RaggedAttention",
    "Conv1DV1",
    "Conv2DV1",
    "Conv3DV1",
    "Conv3D",
    "Conv1D",
    "ConvTranspose1d",
    "WeightNormConvTranspose1d",
    "DistributedAttentionImpl",
    "DistributedAttentionWithRope",
    "ColumnParallelLinear",
    "DistributedMLP",
    "DistributedRMSNorm",
    "DistributedTransformer",
    "DistributedTransformerBlock",
    "EmbeddingV1",
    "Embedding",
    "Float8Config",
    "Float8ScaleGranularity",
    "Float8ScaleOrigin",
    "Float8InputScaleSpec",
    "Float8WeightScaleSpec",
    "GGUFQAttentionWithRope",
    "GPTQAttentionWithRope",
    "GPTQLinear",
    "GroupNorm",
    "Layer",
    "LayerNormV1",
    "LayerNorm",
    "LinearV1",
    "Linear",
    "LinearScalingParams",
    "Llama3RopeScalingParams",
    "Llama3RotaryEmbedding",
    "MHAMaskVariant",
    "MLPV1",
    "MLP",
    "Module",
    "LayerList",
    "OptimizedRotaryEmbedding",
    "RMSNormV1",
    "RMSNorm",
    "RotaryEmbedding",
    "ReturnLogits",
    "Sequential",
    "Signals",
    "Transformer",
    "TransformerBlock",
]