Hugging Face's logo

aws-neuron
/
optimum-neuron-cache

Model card Files
xet
Community
647
dacorvo HF Staff commited on
Commit
6629716
·
verified ·
1 Parent(s): cd30361

Synchronizing local compiler cache.

Browse files
Files changed (12) hide show
  1. .gitattributes +3 -0
  2. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev2/qwen3/Qwen/Qwen3-Embedding-0.6B/9eb34185e11a0b7c5ac2.json +87 -0
  3. neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev2/qwen3/Qwen/Qwen3-Embedding-0.6B/d652fbacbf3a73cb1caa.json +87 -0
  4. neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/compile_flags.json +1 -0
  5. neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/model.done +0 -0
  6. neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/model.hlo_module.pb +3 -0
  7. neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/model.neff +3 -0
  8. neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/wrapped_neff.hlo +3 -0
  9. neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3156313afcf6f9c1c2b+24129607/compile_flags.json +1 -0
  10. neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3156313afcf6f9c1c2b+24129607/model.done +0 -0
  11. neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3156313afcf6f9c1c2b+24129607/model.hlo_module.pb +3 -0
  12. neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3156313afcf6f9c1c2b+24129607/model.neff +3 -0
.gitattributes CHANGED
@@ -12925,3 +12925,6 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_f668e8059b4ad00c098b+a02c3a36/wrapped_nef
12925
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_ee0ccce0e045a597f99a+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12926
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_ee0ccce0e045a597f99a+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12927
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_f34dea1dd0e39d38d79e+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
12925
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_ee0ccce0e045a597f99a+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12926
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_ee0ccce0e045a597f99a+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12927
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_f34dea1dd0e39d38d79e+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12928
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12929
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12930
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3156313afcf6f9c1c2b+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev2/qwen3/Qwen/Qwen3-Embedding-0.6B/9eb34185e11a0b7c5ac2.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-Embedding-0.6B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 1024,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 3072,
16
+ "layer_types": [
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
+ ],
46
+ "max_position_embeddings": 32768,
47
+ "max_window_layers": 28,
48
+ "model_type": "qwen3",
49
+ "neuron": {
50
+ "_serialized_key": "NxDNeuronConfig",
51
+ "batch_size": 4,
52
+ "capacity_factor": null,
53
+ "checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
54
+ "checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
55
+ "continuous_batching": false,
56
+ "ep_degree": 1,
57
+ "fused_qkv": true,
58
+ "glu_mlp": true,
59
+ "local_ranks_size": 24,
60
+ "max_batch_size": 4,
61
+ "max_context_length": 1024,
62
+ "max_topk": 256,
63
+ "n_active_tokens": 1024,
64
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
65
+ "on_device_sampling": false,
66
+ "optimum_neuron_version": "0.4.2.dev2",
67
+ "output_logits": false,
68
+ "pp_degree": 1,
69
+ "sequence_length": 1024,
70
+ "speculation_length": 0,
71
+ "start_rank_id": 0,
72
+ "target": "trn1",
73
+ "torch_dtype": "bfloat16",
74
+ "tp_degree": 24
75
+ },
76
+ "num_attention_heads": 16,
77
+ "num_hidden_layers": 28,
78
+ "num_key_value_heads": 8,
79
+ "rms_norm_eps": 1e-06,
80
+ "rope_scaling": null,
81
+ "rope_theta": 1000000,
82
+ "sliding_window": null,
83
+ "tie_word_embeddings": true,
84
+ "use_cache": true,
85
+ "use_sliding_window": false,
86
+ "vocab_size": 151669
87
+ }
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.2.dev2/qwen3/Qwen/Qwen3-Embedding-0.6B/d652fbacbf3a73cb1caa.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "Qwen/Qwen3-Embedding-0.6B",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 1024,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 3072,
16
+ "layer_types": [
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
+ ],
46
+ "max_position_embeddings": 32768,
47
+ "max_window_layers": 28,
48
+ "model_type": "qwen3",
49
+ "neuron": {
50
+ "_serialized_key": "NxDNeuronConfig",
51
+ "batch_size": 6,
52
+ "capacity_factor": null,
53
+ "checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
54
+ "checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
55
+ "continuous_batching": false,
56
+ "ep_degree": 1,
57
+ "fused_qkv": true,
58
+ "glu_mlp": true,
59
+ "local_ranks_size": 24,
60
+ "max_batch_size": 6,
61
+ "max_context_length": 1024,
62
+ "max_topk": 256,
63
+ "n_active_tokens": 1024,
64
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
65
+ "on_device_sampling": false,
66
+ "optimum_neuron_version": "0.4.2.dev2",
67
+ "output_logits": false,
68
+ "pp_degree": 1,
69
+ "sequence_length": 1024,
70
+ "speculation_length": 0,
71
+ "start_rank_id": 0,
72
+ "target": "trn1",
73
+ "torch_dtype": "bfloat16",
74
+ "tp_degree": 24
75
+ },
76
+ "num_attention_heads": 16,
77
+ "num_hidden_layers": 28,
78
+ "num_key_value_heads": 8,
79
+ "rms_norm_eps": 1e-06,
80
+ "rope_scaling": null,
81
+ "rope_theta": 1000000,
82
+ "sliding_window": null,
83
+ "tie_word_embeddings": true,
84
+ "use_cache": true,
85
+ "use_sliding_window": false,
86
+ "vocab_size": 151669
87
+ }
neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+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_07e856db17444803b299+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a97bb411dcd64871e1b1784620de52aab98fa041ae87f39fee46a8518b9f118
3
+ size 78518
neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:656e4cb5b7bfb6a5602c38d8502dcb0b3d9d9a6361ba2268a808b09ef261935f
3
+ size 287744
neuronxcc-2.21.18209.0+043b1bf7/MODULE_07e856db17444803b299+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad223b09145789b49f073d6b6a2a1e0544f5ed69dd4cdcb65272eae873b3395d
3
+ size 296094
neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3156313afcf6f9c1c2b+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_b3156313afcf6f9c1c2b+24129607/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3156313afcf6f9c1c2b+24129607/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f61a1b41eb6388be1fb89f02fa8277176b1be0d7dfd3e649990642aa78e195bc
3
+ size 79552
neuronxcc-2.21.18209.0+043b1bf7/MODULE_b3156313afcf6f9c1c2b+24129607/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2d32c0b5819cd53fb4f9d4b3c1e2bd992110382bc3ffee2b9e8cf65327fa5e41
3
+ size 338944