Megatron-Bridge/src/megatron/bridge/recipes/llama/llama3.py at main · NVIDIA-NeMo/Megatron-Bridge · 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
131
132
133
134
135
136
137
138
# Copyright (c) 2026, NVIDIA CORPORATION.  All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# 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.
# ruff: noqa: F401
"""Compatibility aliases for legacy recipe names."""

from __future__ import annotations

from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_8b_peft_1gpu_h100_bf16_config as llama3_8b_peft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_8b_pretrain_2gpu_h100_bf16_config as llama3_8b_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_8b_pretrain_2gpu_h100_fp8cs_config as llama3_8b_low_precision_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_8b_pretrain_16gpu_h100_bf16_16k_config as llama3_8b_16k_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_8b_pretrain_32gpu_h100_bf16_64k_config as llama3_8b_64k_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_8b_pretrain_64gpu_h100_bf16_128k_config as llama3_8b_128k_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_8b_sft_2gpu_h100_bf16_config as llama3_8b_sft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_70b_peft_8gpu_h100_bf16_config as llama3_70b_peft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_70b_pretrain_32gpu_h100_bf16_16k_config as llama3_70b_16k_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_70b_pretrain_32gpu_h100_bf16_config as llama3_70b_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_70b_pretrain_32gpu_h100_bf16_deterministic_config as llama3_70b_pretrain_deterministic_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_70b_pretrain_256gpu_h100_bf16_64k_config as llama3_70b_64k_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama3_70b_sft_32gpu_h100_bf16_config as llama3_70b_sft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_8b_peft_1gpu_h100_bf16_config as llama31_8b_peft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_8b_pretrain_2gpu_h100_bf16_config as llama31_8b_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_8b_sft_2gpu_h100_bf16_config as llama31_8b_sft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_70b_peft_8gpu_h100_bf16_config as llama31_70b_peft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_70b_pretrain_32gpu_h100_bf16_config as llama31_70b_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_70b_sft_32gpu_h100_bf16_config as llama31_70b_sft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_405b_peft_32gpu_h100_bf16_config as llama31_405b_peft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_405b_pretrain_256gpu_h100_bf16_config as llama31_405b_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_405b_pretrain_256gpu_h100_bf16_deterministic_config as llama31_405b_pretrain_deterministic_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama31_405b_sft_128gpu_h100_bf16_config as llama31_405b_sft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama32_1b_peft_1gpu_h100_bf16_config as llama32_1b_peft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama32_1b_pretrain_1gpu_h100_bf16_config as llama32_1b_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama32_1b_sft_1gpu_h100_bf16_config as llama32_1b_sft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama32_3b_peft_1gpu_h100_bf16_config as llama32_3b_peft_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama32_3b_pretrain_1gpu_h100_bf16_config as llama32_3b_pretrain_config,
)
from megatron.bridge.recipes.llama.h100.llama3 import (
    llama32_3b_sft_1gpu_h100_bf16_config as llama32_3b_sft_config,
)


__all__ = [
    "llama31_405b_peft_config",
    "llama31_405b_pretrain_config",
    "llama31_405b_pretrain_deterministic_config",
    "llama31_405b_sft_config",
    "llama31_70b_peft_config",
    "llama31_70b_pretrain_config",
    "llama31_70b_sft_config",
    "llama31_8b_peft_config",
    "llama31_8b_pretrain_config",
    "llama31_8b_sft_config",
    "llama32_1b_peft_config",
    "llama32_1b_pretrain_config",
    "llama32_1b_sft_config",
    "llama32_3b_peft_config",
    "llama32_3b_pretrain_config",
    "llama32_3b_sft_config",
    "llama3_70b_16k_pretrain_config",
    "llama3_70b_64k_pretrain_config",
    "llama3_70b_peft_config",
    "llama3_70b_pretrain_config",
    "llama3_70b_pretrain_deterministic_config",
    "llama3_70b_sft_config",
    "llama3_8b_128k_pretrain_config",
    "llama3_8b_16k_pretrain_config",
    "llama3_8b_64k_pretrain_config",
    "llama3_8b_peft_config",
    "llama3_8b_pretrain_config",
    "llama3_8b_sft_config",
    "llama3_8b_low_precision_pretrain_config",
]