File size: 8,303 Bytes
d9adc0d
 
ab0856f
ed3e1e1
 
 
 
 
5143203
ed3e1e1
 
ab0856f
 
 
2bad6cc
ed3e1e1
 
5143203
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ed3e1e1
 
 
 
 
 
 
 
d9adc0d
 
 
 
 
faa378a
 
d9adc0d
faa378a
 
 
d9adc0d
 
faa378a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d9adc0d
 
 
faa378a
 
 
d9adc0d
 
 
 
 
 
faa378a
 
d9adc0d
faa378a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d9adc0d
 
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
{
  "metadata": {
    "total_size": 140323266560,
    "format": "safetensors",
    "model_name": "DeepXR/Helion-2.5-Rnd",
    "version": "2.5.0-rnd",
    "precision": "float16",
    "parameters": "70B",
    "total_shards": 83,
    "created_at": "2025-01-30T00:00:00Z",
    "sha256_checksums_available": true,
    "shard_size_bytes": 1690280320,
    "shard_size_gb": 1.69,
    "shard_size_gib": 1.57,
    "note": "Model weights distributed across 83 SafeTensors shards in FP16 precision"
  },
  "weight_map": {
    "model.embed_tokens.weight": "shard_00.safetensors",
    "model.layers.0.self_attn.q_proj.weight": "shard_01.safetensors",
    "model.layers.0.self_attn.k_proj.weight": "shard_01.safetensors",
    "model.layers.0.self_attn.v_proj.weight": "shard_02.safetensors",
    "model.layers.0.self_attn.o_proj.weight": "shard_02.safetensors",
    "model.layers.0.mlp.gate_proj.weight": "shard_03.safetensors",
    "model.layers.0.mlp.up_proj.weight": "shard_03.safetensors",
    "model.layers.0.mlp.down_proj.weight": "shard_04.safetensors",
    "model.layers.0.input_layernorm.weight": "shard_04.safetensors",
    "model.layers.0.post_attention_layernorm.weight": "shard_04.safetensors",
    "model.layers.1.self_attn.q_proj.weight": "shard_05.safetensors",
    "model.layers.1.self_attn.k_proj.weight": "shard_05.safetensors",
    "model.layers.1.self_attn.v_proj.weight": "shard_06.safetensors",
    "model.layers.1.self_attn.o_proj.weight": "shard_06.safetensors",
    "model.layers.1.mlp.gate_proj.weight": "shard_07.safetensors",
    "model.layers.1.mlp.up_proj.weight": "shard_07.safetensors",
    "model.layers.1.mlp.down_proj.weight": "shard_08.safetensors",
    "model.layers.2.self_attn.q_proj.weight": "shard_09.safetensors",
    "model.layers.2.self_attn.k_proj.weight": "shard_09.safetensors",
    "model.layers.2.self_attn.v_proj.weight": "shard_10.safetensors",
    "model.layers.2.self_attn.o_proj.weight": "shard_10.safetensors",
    "model.layers.2.mlp.gate_proj.weight": "shard_11.safetensors",
    "model.layers.2.mlp.up_proj.weight": "shard_11.safetensors",
    "model.layers.2.mlp.down_proj.weight": "shard_12.safetensors",
    "model.layers.3.self_attn.q_proj.weight": "shard_13.safetensors",
    "model.layers.3.self_attn.k_proj.weight": "shard_13.safetensors",
    "model.layers.3.self_attn.v_proj.weight": "shard_14.safetensors",
    "model.layers.3.self_attn.o_proj.weight": "shard_14.safetensors",
    "model.layers.3.mlp.gate_proj.weight": "shard_15.safetensors",
    "model.layers.3.mlp.up_proj.weight": "shard_15.safetensors",
    "model.layers.3.mlp.down_proj.weight": "shard_16.safetensors",
    "model.layers.29.self_attn.q_proj.weight": "shard_73.safetensors",
    "model.layers.29.self_attn.k_proj.weight": "shard_73.safetensors",
    "model.layers.29.self_attn.v_proj.weight": "shard_74.safetensors",
    "model.layers.29.self_attn.o_proj.weight": "shard_74.safetensors",
    "model.layers.29.mlp.gate_proj.weight": "shard_75.safetensors",
    "model.layers.29.mlp.up_proj.weight": "shard_75.safetensors",
    "model.layers.29.mlp.down_proj.weight": "shard_76.safetensors",
    "model.layers.30.self_attn.q_proj.weight": "shard_77.safetensors",
    "model.layers.30.self_attn.k_proj.weight": "shard_77.safetensors",
    "model.layers.30.self_attn.v_proj.weight": "shard_78.safetensors",
    "model.layers.30.self_attn.o_proj.weight": "shard_78.safetensors",
    "model.layers.30.mlp.gate_proj.weight": "shard_79.safetensors",
    "model.layers.30.mlp.up_proj.weight": "shard_79.safetensors",
    "model.layers.30.mlp.down_proj.weight": "shard_80.safetensors",
    "model.layers.31.self_attn.q_proj.weight": "shard_80.safetensors",
    "model.layers.31.self_attn.k_proj.weight": "shard_80.safetensors",
    "model.layers.31.self_attn.v_proj.weight": "shard_81.safetensors",
    "model.layers.31.self_attn.o_proj.weight": "shard_81.safetensors",
    "model.layers.31.mlp.gate_proj.weight": "shard_81.safetensors",
    "model.layers.31.mlp.up_proj.weight": "shard_82.safetensors",
    "model.layers.31.mlp.down_proj.weight": "shard_82.safetensors",
    "model.layers.31.input_layernorm.weight": "shard_82.safetensors",
    "model.layers.31.post_attention_layernorm.weight": "shard_82.safetensors",
    "model.norm.weight": "shard_82.safetensors",
    "lm_head.weight": "shard_82.safetensors"
  }
} {
    "total_size": 140737488355328,
    "format": "safetensors",
    "model_name": "DeepXR/Helion-2.5-Rnd",
    "version": "2.5.0-rnd",
    "precision": "bfloat16",
    "parameters": "70B",
    "total_shards": 96,
    "created_at": "2025-01-30T00:00:00Z",
    "sha256_checksums_available": true,
    "shard_size_avg": "1.46GB",
    "note": "SafeTensors shards will be available soon. Model weights distributed across 96 files for optimal loading."
  },
  "weight_map": {
    "model.embed_tokens.weight": "model-00001-of-00096.safetensors",
    "model.layers.0.self_attn.q_proj.weight": "model-00002-of-00096.safetensors",
    "model.layers.0.self_attn.k_proj.weight": "model-00002-of-00096.safetensors",
    "model.layers.0.self_attn.v_proj.weight": "model-00003-of-00096.safetensors",
    "model.layers.0.self_attn.o_proj.weight": "model-00003-of-00096.safetensors",
    "model.layers.0.mlp.gate_proj.weight": "model-00004-of-00096.safetensors",
    "model.layers.0.mlp.up_proj.weight": "model-00004-of-00096.safetensors",
    "model.layers.0.mlp.down_proj.weight": "model-00005-of-00096.safetensors",
    "model.layers.0.input_layernorm.weight": "model-00005-of-00096.safetensors",
    "model.layers.0.post_attention_layernorm.weight": "model-00005-of-00096.safetensors",
    "model.layers.1.self_attn.q_proj.weight": "model-00006-of-00096.safetensors",
    "model.layers.1.self_attn.k_proj.weight": "model-00006-of-00096.safetensors",
    "model.layers.1.self_attn.v_proj.weight": "model-00007-of-00096.safetensors",
    "model.layers.1.self_attn.o_proj.weight": "model-00007-of-00096.safetensors",
    "model.layers.1.mlp.gate_proj.weight": "model-00008-of-00096.safetensors",
    "model.layers.1.mlp.up_proj.weight": "model-00008-of-00096.safetensors",
    "model.layers.1.mlp.down_proj.weight": "model-00009-of-00096.safetensors",
    "model.layers.31.self_attn.q_proj.weight": "model-00092-of-00096.safetensors",
    "model.layers.31.self_attn.k_proj.weight": "model-00093-of-00096.safetensors",
    "model.layers.31.self_attn.v_proj.weight": "model-00093-of-00096.safetensors",
    "model.layers.31.self_attn.o_proj.weight": "model-00094-of-00096.safetensors",
    "model.layers.31.mlp.gate_proj.weight": "model-00094-of-00096.safetensors",
    "model.layers.31.mlp.up_proj.weight": "model-00095-of-00096.safetensors",
    "model.layers.31.mlp.down_proj.weight": "model-00095-of-00096.safetensors",
    "model.layers.31.input_layernorm.weight": "model-00096-of-00096.safetensors",
    "model.layers.31.post_attention_layernorm.weight": "model-00096-of-00096.safetensors",
    "model.norm.weight": "model-00096-of-00096.safetensors",
    "lm_head.weight": "model-00096-of-00096.safetensors"
  },
  "safetensors_info": {
    "description": "SafeTensors format provides secure, fast, and zero-copy tensor serialization",
    "total_shards": 96,
    "parameters": "70B",
    "shard_distribution": "Each transformer layer distributed across ~3 shards for balanced loading",
    "benefits": [
      "No arbitrary code execution during loading",
      "Lazy loading support for memory efficiency",
      "Fast deserialization without pickle",
      "Tensor metadata validation",
      "Cross-platform compatibility",
      "Memory-mapped file support",
      "Parallel loading across multiple GPUs"
    ],
    "verification": "Each file will include SHA256 checksum for integrity verification",
    "status": "In preparation - 96 shards to be released soon"
  },
  "shard_structure": {
    "embedding_layer": {
      "shards": ["model-00001-of-00096.safetensors"],
      "size_estimate": "~1.0GB"
    },
    "transformer_layers": {
      "layers_per_shard": "~0.33",
      "shards_range": "model-00002-of-00096 to model-00095-of-00096",
      "size_estimate": "~1.5GB per shard"
    },
    "output_layer": {
      "shards": ["model-00096-of-00096.safetensors"],
      "includes": ["model.norm.weight", "lm_head.weight"],
      "size_estimate": "~1.0GB"
    }
  }
}