Hugging Face's logo

aws-neuron
/
optimum-neuron-cache

Model card Files
xet
 Community
679
dacorvo HF Staff commited on
Commit
4b289b9
·
verified ·
1 Parent(s): a8b2c76

Synchronizing local compiler cache.

Browse files
Files changed (11) hide show
  1. .gitattributes +3 -0
  2. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.1.dev0/qwen3/Qwen/Qwen3-8B/359ce07819647afb5d1e.json +94 -0
  3. neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/compile_flags.json +1 -0
  4. neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/model.done +0 -0
  5. neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/model.hlo_module.pb +3 -0
  6. neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/model.neff +3 -0
  7. neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/wrapped_neff.hlo +3 -0
  8. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/compile_flags.json +1 -0
  9. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/model.done +0 -0
  10. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/model.hlo_module.pb +3 -0
  11. neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/model.neff +3 -0
.gitattributes CHANGED
@@ -12256,3 +12256,6 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_5efc215068544cab9a3a+ac10809c/wrapped_nef
12256
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_3c555e49fce854b9ff0c+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12257
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_adc1b12bf88cc35d7d8e+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12258
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_adc1b12bf88cc35d7d8e+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
 
 
 
 
12256
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_3c555e49fce854b9ff0c+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12257
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_adc1b12bf88cc35d7d8e+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12258
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_adc1b12bf88cc35d7d8e+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12259
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12260
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12261
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.1.dev0/qwen3/Qwen/Qwen3-8B/359ce07819647afb5d1e.json ADDED
@@ -0,0 +1,94 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-8B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "head_dim": 128,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 4096,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 12288,
15
+ "layer_types": [
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention",
51
+ "full_attention"
52
+ ],
53
+ "max_position_embeddings": 40960,
54
+ "max_window_layers": 36,
55
+ "model_type": "qwen3",
56
+ "neuron": {
57
+ "_serialized_key": "NxDNeuronConfig",
58
+ "batch_size": 4,
59
+ "capacity_factor": null,
60
+ "checkpoint_id": "Qwen/Qwen3-8B",
61
+ "checkpoint_revision": "b968826d9c46dd6066d109eabc6255188de91218",
62
+ "continuous_batching": true,
63
+ "ep_degree": 1,
64
+ "fused_qkv": true,
65
+ "glu_mlp": true,
66
+ "local_ranks_size": 2,
67
+ "max_batch_size": 4,
68
+ "max_context_length": 4096,
69
+ "max_topk": 256,
70
+ "n_active_tokens": 4096,
71
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
72
+ "on_device_sampling": false,
73
+ "optimum_neuron_version": "0.4.1.dev0",
74
+ "output_logits": false,
75
+ "pp_degree": 1,
76
+ "sequence_length": 4096,
77
+ "speculation_length": 0,
78
+ "start_rank_id": 0,
79
+ "target": "trn1",
80
+ "torch_dtype": "bfloat16",
81
+ "tp_degree": 2
82
+ },
83
+ "num_attention_heads": 32,
84
+ "num_hidden_layers": 36,
85
+ "num_key_value_heads": 8,
86
+ "rms_norm_eps": 1e-06,
87
+ "rope_scaling": null,
88
+ "rope_theta": 1000000,
89
+ "sliding_window": null,
90
+ "tie_word_embeddings": false,
91
+ "use_cache": true,
92
+ "use_sliding_window": false,
93
+ "vocab_size": 151936
94
+ }
neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/token_generation/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f9b8f22e3c6b1346736aa2a80cc2900c950a32fd4add82434561a30be5374603
3
+ size 904963
neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3cea9cfeffe6bb91d7ded1230965fefbdb19bce0403bbb27cff944cde2fcbc04
3
+ size 11428864
neuronxcc-2.21.18209.0+043b1bf7/MODULE_32d522dd3d1f98c8a4da+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa06235d8d33393c3af4dc500ca53c98ffa3dcbafbe2aafcc01493d11d6e51e1
3
+ size 11603971
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/context_encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:61fdaefc69c3edacf6cab8e1c287b983ae09d8f2daebce846fe8118eb1e53b69
3
+ size 1110844
neuronxcc-2.21.18209.0+043b1bf7/MODULE_3dc9d494cc695a427534+24129607/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b013261cf2269ba0fbfc42b08ee486c847314f2447e631051feb439b14772392
3
+ size 9493504