Hugging Face's logo

FlyinGodzilla
/
AMOR-Mamba2-180M

Text Generation
PyTorch
Safetensors
English
amor
custom-code
pretrained
research
mamba2
Model card Files
xet
 Community
AMOR-Mamba2-180M
File size: 4,598 Bytes
f762281
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4e93134
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ea6c58d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
846351a
 
 
 
 
 
4e93134
f762281
ea6c58d
 
 
 
 
 
 
 
 
 
 
 
f762281
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
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
{
  "repo_id": "FlyinGodzilla/AMOR-Mamba2-180M",
  "source_filename": "amor_180m_step_0007503.pt",
  "source_sha256": "8edd3ba619712905e8a00aaa7e6b97de6cc96d6b8a11ec2e24b5107e102b08bd",
  "step": 7503,
  "tokens_seen": 3687948288,
  "n_params": 184396896,
  "tensor_dtypes": {
    "torch.float32": 182
  },
  "gate_buffers": {
    "0": {
      "running_threshold": 0.7798937559127808,
      "running_std": 0.08388833701610565
    },
    "1": {
      "running_threshold": 0.6911748051643372,
      "running_std": 0.12514136731624603
    },
    "2": {
      "running_threshold": 0.7131258845329285,
      "running_std": 0.11415675282478333
    }
  },
  "optimizer_included": false,
  "precision_changed": false,
  "validation": {
    "exported_tensors_equal": true,
    "strict_model_load": true,
    "generation": true,
    "device": "cpu",
    "dtype": "float32",
    "torch": "2.6.0",
    "transformers": "4.51.3",
    "tokenizer_revision": "5590cc63198b0f111f4289daf006da0a12044173",
    "prompt": "The Eiffel Tower is",
    "generated_token_ids": [
      264,
      220,
      16,
      11
    ],
    "prefill_max_abs_diff": 0.0,
    "decode_max_abs_diff": 1.71661376953125e-05,
    "seconds": 22.35716986656189,
    "canonical_forward_exact_match": true,
    "canonical_forward_max_abs_diff": 0.0,
    "cuda": {
      "versions": {
        "torch": "2.6.0+cu124",
        "transformers": "5.0.0",
        "mamba-ssm": "2.3.0",
        "causal-conv1d": "1.6.0",
        "flash-linear-attention": "0.4.2",
        "triton": "3.2.0",
        "safetensors": "0.7.0"
      },
      "tokenizer_revision": "5590cc63198b0f111f4289daf006da0a12044173",
      "strict_model_load": true,
      "generation": true,
      "device": "NVIDIA H100 NVL",
      "torch": "2.6.0+cu124",
      "autocast": "bfloat16",
      "prefill_max_abs_diff": 0.0,
      "decode_max_abs_diff": 0.0625,
      "generated_token_ids": [
        264,
        1949,
        323,
        1825,
        2592,
        2447,
        430,
        6276
      ],
      "published_cli_passed": true,
      "test_scope": "Short CUDA smoke test; not a benchmark rerun",
      "comparison_tolerance": {
        "rtol": 0.02,
        "atol": 0.1
      },
      "tested_decoder_sha256": "4b48db8e5fb67318244c533356903f40703156cdca24ccf4fd49086e72b16db7"
    },
    "router_release": {
      "exact_router_export": true,
      "base_model_association": true,
      "default_forward_max_abs_diff": 0.0,
      "modes": [
        {
          "router": false,
          "cases": [
            {
              "batch": 1,
              "length": 16,
              "prefill_max_abs_diff": 6.103515625e-05,
              "decode_max_abs_diff": 0.0625
            },
            {
              "batch": 1,
              "length": 128,
              "prefill_max_abs_diff": 0.0,
              "decode_max_abs_diff": 0.0625
            },
            {
              "batch": 2,
              "length": 64,
              "prefill_max_abs_diff": 0.0
            }
          ],
          "canonical_router_exact": null
        },
        {
          "router": true,
          "cases": [
            {
              "batch": 1,
              "length": 16,
              "prefill_max_abs_diff": 6.103515625e-05,
              "decode_max_abs_diff": 0.0625
            },
            {
              "batch": 1,
              "length": 128,
              "prefill_max_abs_diff": 0.0,
              "decode_max_abs_diff": 0.0625
            },
            {
              "batch": 2,
              "length": 64,
              "prefill_max_abs_diff": 0.0
            }
          ],
          "canonical_router_exact": true
        }
      ],
      "frozen_thresholds_preserved": true,
      "training_router_rejected": true,
      "gpu": "NVIDIA H100 NVL",
      "torch": "2.6.0+cu124",
      "generation_cli_modes_passed": [
        "entropy",
        "router"
      ]
    },
    "current_release": {
      "strict_model_load": true,
      "entropy_and_router_generation_passed": true,
      "unit_tests_passed": 17,
      "device": "NVIDIA H100 NVL"
    }
  },
  "implementation": "AMOR shared release: native entropy gate and optional fitted inference router",
  "router": {
    "variant": "distill",
    "hidden_dim": 512,
    "activation": "silu",
    "weights": "router.safetensors",
    "config": "router_config.json",
    "sha256": "5cdb826237fb235af5a63a9d56cdcc63bdeab1b4cc1027970420f5aabadb8c48",
    "base_weight_sha256": "9857e8219052f0272aaf6a9801029dd8e1f51b668cbad3e1aadfc3a109b4aacb",
    "precision": "float32",
    "inference_only": true
  }
}