Synchronizing local compiler cache.
Browse files
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.3.dev0/llama4_text/meta-llama/Llama-4-Scout-17B-16E-Instruct/ca9ca8c51442b972dc25.json
ADDED
|
@@ -0,0 +1,220 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "meta-llama/Llama-4-Scout-17B-16E-Instruct",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_chunk_size": 8192,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attn_scale": 0.1,
|
| 9 |
+
"attn_temperature_tuning": true,
|
| 10 |
+
"dtype": "bfloat16",
|
| 11 |
+
"floor_scale": 8192,
|
| 12 |
+
"for_llm_compressor": false,
|
| 13 |
+
"head_dim": 128,
|
| 14 |
+
"hidden_act": "silu",
|
| 15 |
+
"hidden_size": 5120,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"interleave_moe_layer_step": 1,
|
| 18 |
+
"intermediate_size": 8192,
|
| 19 |
+
"intermediate_size_mlp": 16384,
|
| 20 |
+
"layer_types": [
|
| 21 |
+
"chunked_attention",
|
| 22 |
+
"chunked_attention",
|
| 23 |
+
"chunked_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"chunked_attention",
|
| 26 |
+
"chunked_attention",
|
| 27 |
+
"chunked_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"chunked_attention",
|
| 30 |
+
"chunked_attention",
|
| 31 |
+
"chunked_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"chunked_attention",
|
| 34 |
+
"chunked_attention",
|
| 35 |
+
"chunked_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"chunked_attention",
|
| 38 |
+
"chunked_attention",
|
| 39 |
+
"chunked_attention",
|
| 40 |
+
"full_attention",
|
| 41 |
+
"chunked_attention",
|
| 42 |
+
"chunked_attention",
|
| 43 |
+
"chunked_attention",
|
| 44 |
+
"full_attention",
|
| 45 |
+
"chunked_attention",
|
| 46 |
+
"chunked_attention",
|
| 47 |
+
"chunked_attention",
|
| 48 |
+
"full_attention",
|
| 49 |
+
"chunked_attention",
|
| 50 |
+
"chunked_attention",
|
| 51 |
+
"chunked_attention",
|
| 52 |
+
"full_attention",
|
| 53 |
+
"chunked_attention",
|
| 54 |
+
"chunked_attention",
|
| 55 |
+
"chunked_attention",
|
| 56 |
+
"full_attention",
|
| 57 |
+
"chunked_attention",
|
| 58 |
+
"chunked_attention",
|
| 59 |
+
"chunked_attention",
|
| 60 |
+
"full_attention",
|
| 61 |
+
"chunked_attention",
|
| 62 |
+
"chunked_attention",
|
| 63 |
+
"chunked_attention",
|
| 64 |
+
"full_attention",
|
| 65 |
+
"chunked_attention",
|
| 66 |
+
"chunked_attention",
|
| 67 |
+
"chunked_attention",
|
| 68 |
+
"full_attention"
|
| 69 |
+
],
|
| 70 |
+
"max_position_embeddings": 10485760,
|
| 71 |
+
"model_type": "llama4_text",
|
| 72 |
+
"moe_layers": [
|
| 73 |
+
0,
|
| 74 |
+
1,
|
| 75 |
+
2,
|
| 76 |
+
3,
|
| 77 |
+
4,
|
| 78 |
+
5,
|
| 79 |
+
6,
|
| 80 |
+
7,
|
| 81 |
+
8,
|
| 82 |
+
9,
|
| 83 |
+
10,
|
| 84 |
+
11,
|
| 85 |
+
12,
|
| 86 |
+
13,
|
| 87 |
+
14,
|
| 88 |
+
15,
|
| 89 |
+
16,
|
| 90 |
+
17,
|
| 91 |
+
18,
|
| 92 |
+
19,
|
| 93 |
+
20,
|
| 94 |
+
21,
|
| 95 |
+
22,
|
| 96 |
+
23,
|
| 97 |
+
24,
|
| 98 |
+
25,
|
| 99 |
+
26,
|
| 100 |
+
27,
|
| 101 |
+
28,
|
| 102 |
+
29,
|
| 103 |
+
30,
|
| 104 |
+
31,
|
| 105 |
+
32,
|
| 106 |
+
33,
|
| 107 |
+
34,
|
| 108 |
+
35,
|
| 109 |
+
36,
|
| 110 |
+
37,
|
| 111 |
+
38,
|
| 112 |
+
39,
|
| 113 |
+
40,
|
| 114 |
+
41,
|
| 115 |
+
42,
|
| 116 |
+
43,
|
| 117 |
+
44,
|
| 118 |
+
45,
|
| 119 |
+
46,
|
| 120 |
+
47
|
| 121 |
+
],
|
| 122 |
+
"neuron": {
|
| 123 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 124 |
+
"batch_size": 4,
|
| 125 |
+
"capacity_factor": null,
|
| 126 |
+
"checkpoint_id": "meta-llama/Llama-4-Scout-17B-16E-Instruct",
|
| 127 |
+
"checkpoint_revision": "92f3b1597a195b523d8d9e5700e57e4fbb8f20d3",
|
| 128 |
+
"continuous_batching": true,
|
| 129 |
+
"ep_degree": 1,
|
| 130 |
+
"fused_qkv": false,
|
| 131 |
+
"glu_mlp": true,
|
| 132 |
+
"local_ranks_size": 16,
|
| 133 |
+
"max_batch_size": 4,
|
| 134 |
+
"max_context_length": 4096,
|
| 135 |
+
"max_topk": 256,
|
| 136 |
+
"n_active_tokens": 4096,
|
| 137 |
+
"neuronxcc_version": "2.21.18209.0+043b1bf7",
|
| 138 |
+
"on_device_sampling": true,
|
| 139 |
+
"optimum_neuron_version": "0.4.3.dev0",
|
| 140 |
+
"output_logits": false,
|
| 141 |
+
"pp_degree": 1,
|
| 142 |
+
"sequence_length": 4096,
|
| 143 |
+
"speculation_length": 0,
|
| 144 |
+
"start_rank_id": 0,
|
| 145 |
+
"target": "trn1",
|
| 146 |
+
"torch_dtype": "bfloat16",
|
| 147 |
+
"tp_degree": 16
|
| 148 |
+
},
|
| 149 |
+
"no_rope_layers": [
|
| 150 |
+
1,
|
| 151 |
+
1,
|
| 152 |
+
1,
|
| 153 |
+
0,
|
| 154 |
+
1,
|
| 155 |
+
1,
|
| 156 |
+
1,
|
| 157 |
+
0,
|
| 158 |
+
1,
|
| 159 |
+
1,
|
| 160 |
+
1,
|
| 161 |
+
0,
|
| 162 |
+
1,
|
| 163 |
+
1,
|
| 164 |
+
1,
|
| 165 |
+
0,
|
| 166 |
+
1,
|
| 167 |
+
1,
|
| 168 |
+
1,
|
| 169 |
+
0,
|
| 170 |
+
1,
|
| 171 |
+
1,
|
| 172 |
+
1,
|
| 173 |
+
0,
|
| 174 |
+
1,
|
| 175 |
+
1,
|
| 176 |
+
1,
|
| 177 |
+
0,
|
| 178 |
+
1,
|
| 179 |
+
1,
|
| 180 |
+
1,
|
| 181 |
+
0,
|
| 182 |
+
1,
|
| 183 |
+
1,
|
| 184 |
+
1,
|
| 185 |
+
0,
|
| 186 |
+
1,
|
| 187 |
+
1,
|
| 188 |
+
1,
|
| 189 |
+
0,
|
| 190 |
+
1,
|
| 191 |
+
1,
|
| 192 |
+
1,
|
| 193 |
+
0,
|
| 194 |
+
1,
|
| 195 |
+
1,
|
| 196 |
+
1,
|
| 197 |
+
0
|
| 198 |
+
],
|
| 199 |
+
"num_attention_heads": 40,
|
| 200 |
+
"num_experts_per_tok": 1,
|
| 201 |
+
"num_hidden_layers": 48,
|
| 202 |
+
"num_key_value_heads": 8,
|
| 203 |
+
"num_local_experts": 16,
|
| 204 |
+
"output_router_logits": false,
|
| 205 |
+
"rms_norm_eps": 1e-05,
|
| 206 |
+
"rope_scaling": {
|
| 207 |
+
"factor": 16.0,
|
| 208 |
+
"high_freq_factor": 1.0,
|
| 209 |
+
"low_freq_factor": 1.0,
|
| 210 |
+
"original_max_position_embeddings": 8192,
|
| 211 |
+
"rope_type": "llama3"
|
| 212 |
+
},
|
| 213 |
+
"rope_theta": 500000.0,
|
| 214 |
+
"router_aux_loss_coef": 0.001,
|
| 215 |
+
"router_jitter_noise": 0.0,
|
| 216 |
+
"tie_word_embeddings": false,
|
| 217 |
+
"use_cache": true,
|
| 218 |
+
"use_qk_norm": true,
|
| 219 |
+
"vocab_size": 202048
|
| 220 |
+
}
|