Hugging Face's logo

aws-neuron
/
optimum-neuron-cache

Model card Files
xet
Community
648
dacorvo HF Staff commited on
Commit
4af9fed
·
verified ·
1 Parent(s): b06c257

Synchronizing local compiler cache.

Browse files
Files changed (11) hide show
  1. .gitattributes +3 -0
  2. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama4_text/meta-llama/Llama-4-Scout-17B-16E-Instruct/e968be299bd852db6390.json +220 -0
  3. neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/compile_flags.json +1 -0
  4. neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/model.done +0 -0
  5. neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/model.hlo_module.pb +3 -0
  6. neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/model.neff +3 -0
  7. neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/compile_flags.json +1 -0
  8. neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/model.done +0 -0
  9. neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/model.hlo_module.pb +3 -0
  10. neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/model.neff +3 -0
  11. neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/wrapped_neff.hlo +3 -0
.gitattributes CHANGED
@@ -13700,3 +13700,6 @@ neuronxcc-2.21.33363.0+82129205/MODULE_ecad49bd34b3b175cda5+24129607/model.neff
13700
  neuronxcc-2.21.33363.0+82129205/MODULE_8c8634f2f294a15c36e7+afeb10ce/model.neff filter=lfs diff=lfs merge=lfs -text
13701
  neuronxcc-2.21.33363.0+82129205/MODULE_8c8634f2f294a15c36e7+afeb10ce/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
13702
  neuronxcc-2.21.33363.0+82129205/MODULE_b829c6e3567ae63cc8cd+e0c92c88/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
13700
  neuronxcc-2.21.33363.0+82129205/MODULE_8c8634f2f294a15c36e7+afeb10ce/model.neff filter=lfs diff=lfs merge=lfs -text
13701
  neuronxcc-2.21.33363.0+82129205/MODULE_8c8634f2f294a15c36e7+afeb10ce/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
13702
  neuronxcc-2.21.33363.0+82129205/MODULE_b829c6e3567ae63cc8cd+e0c92c88/model.neff filter=lfs diff=lfs merge=lfs -text
13703
+ neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/model.neff filter=lfs diff=lfs merge=lfs -text
13704
+ neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/model.neff filter=lfs diff=lfs merge=lfs -text
13705
+ neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama4_text/meta-llama/Llama-4-Scout-17B-16E-Instruct/e968be299bd852db6390.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": 1,
125
+ "capacity_factor": null,
126
+ "checkpoint_id": "meta-llama/Llama-4-Scout-17B-16E-Instruct",
127
+ "checkpoint_revision": "92f3b1597a195b523d8d9e5700e57e4fbb8f20d3",
128
+ "continuous_batching": false,
129
+ "ep_degree": 1,
130
+ "fused_qkv": false,
131
+ "glu_mlp": true,
132
+ "local_ranks_size": 32,
133
+ "max_batch_size": 1,
134
+ "max_context_length": 4096,
135
+ "max_topk": 256,
136
+ "n_active_tokens": 4096,
137
+ "neuronxcc_version": "2.21.33363.0+82129205",
138
+ "on_device_sampling": true,
139
+ "optimum_neuron_version": "0.4.3.dev2",
140
+ "output_logits": false,
141
+ "pp_degree": 1,
142
+ "sequence_length": 4096,
143
+ "speculation_length": 0,
144
+ "start_rank_id": 0,
145
+ "target": "trn2",
146
+ "torch_dtype": "bfloat16",
147
+ "tp_degree": 32
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
+ }
neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn2", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=2", "--logfile=/tmp/nxd_model/context_encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87eca0fc7a745752b1c8b0dea385cc411b3abc6c8b5da93a686ff610d60bb80b
3
+ size 104812800
neuronxcc-2.21.33363.0+82129205/MODULE_d4d0d8cf4c8987b96c75+a32116a7/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbdfe30d233077b42b4aa4cb96edf7e12142d5ccf2123b78cf11d5ecc433ee90
3
+ size 33864704
neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn2", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=2", "--logfile=/tmp/nxd_model/token_generation/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f4f627da3e449241581bf186df6bf79b74e63326c464b20202b2232f352680d
3
+ size 103526426
neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04f3662d6f406a936d9f9d9cef50ca9285df645e46d190a82da20fa11f5a6071
3
+ size 8070144
neuronxcc-2.21.33363.0+82129205/MODULE_ea3a007bc363c7595c33+ac10809c/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55ea03bb0d3db7df3e9765fc03ada4c230a61a4aa86868040f0ebe591f32a3bf
3
+ size 8411464