Hugging Face's logo

aws-neuron
/
optimum-neuron-cache

Model card Files
xet
 Community
665
dacorvo HF Staff commited on
Commit
01b7b1e
·
verified ·
1 Parent(s): 6db461b

Synchronizing local compiler cache.

Browse files
Files changed (8) hide show
  1. .gitattributes +2 -0
  2. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev2/719fe7a9c6374369ebb3c8b70e3ca2e20d29e8d4c6f1965131b188cd23d90d00/4c604dcc782261c0a1d2.json +56 -0
  3. neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev2/llama/TinyLlama/TinyLlama-1.1B-Chat-v1.0/4c604dcc782261c0a1d2.json +56 -0
  4. neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/compile_flags.json +1 -0
  5. neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/model.done +0 -0
  6. neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/model.hlo_module.pb +3 -0
  7. neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/model.neff +3 -0
  8. neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/wrapped_neff.hlo +3 -0
.gitattributes CHANGED
@@ -16900,3 +16900,5 @@ neuronxcc-2.21.33363.0+82129205/MODULE_be467b946b0d43827975+ac10809c/wrapped_nef
16900
  neuronxcc-2.21.33363.0+82129205/MODULE_1e724904594a04ae2471+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
16901
  neuronxcc-2.21.33363.0+82129205/MODULE_1e724904594a04ae2471+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
16902
  neuronxcc-2.21.33363.0+82129205/MODULE_55d4fc8ddfc1403ba1a5+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
16900
  neuronxcc-2.21.33363.0+82129205/MODULE_1e724904594a04ae2471+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
16901
  neuronxcc-2.21.33363.0+82129205/MODULE_1e724904594a04ae2471+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
16902
  neuronxcc-2.21.33363.0+82129205/MODULE_55d4fc8ddfc1403ba1a5+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
16903
+ neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/model.neff filter=lfs diff=lfs merge=lfs -text
16904
+ neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev2/719fe7a9c6374369ebb3c8b70e3ca2e20d29e8d4c6f1965131b188cd23d90d00/4c604dcc782261c0a1d2.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "TinyLlama/TinyLlama-1.1B-Chat-v1.0",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "LlamaForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 64,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 2048,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 5632,
16
+ "max_position_embeddings": 2048,
17
+ "mlp_bias": false,
18
+ "model_type": "llama",
19
+ "neuron": {
20
+ "_serialized_key": "NxDNeuronConfig",
21
+ "batch_size": 1,
22
+ "capacity_factor": null,
23
+ "checkpoint_id": "TinyLlama/TinyLlama-1.1B-Chat-v1.0",
24
+ "checkpoint_revision": "fe8a4ea1ffedaf415f4da2f062534de366a451e6",
25
+ "continuous_batching": false,
26
+ "ep_degree": 1,
27
+ "fused_qkv": true,
28
+ "glu_mlp": true,
29
+ "local_ranks_size": 4,
30
+ "max_batch_size": 1,
31
+ "max_context_length": 2048,
32
+ "max_topk": 256,
33
+ "n_active_tokens": 2048,
34
+ "neuronxcc_version": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev2",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "sequence_length": 2048,
40
+ "speculation_length": 0,
41
+ "start_rank_id": 0,
42
+ "target": "trn2",
43
+ "torch_dtype": "bfloat16",
44
+ "tp_degree": 4
45
+ },
46
+ "num_attention_heads": 32,
47
+ "num_hidden_layers": 22,
48
+ "num_key_value_heads": 4,
49
+ "pretraining_tp": 1,
50
+ "rms_norm_eps": 1e-05,
51
+ "rope_scaling": null,
52
+ "rope_theta": 10000.0,
53
+ "tie_word_embeddings": false,
54
+ "use_cache": true,
55
+ "vocab_size": 32000
56
+ }
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.6.dev2/llama/TinyLlama/TinyLlama-1.1B-Chat-v1.0/4c604dcc782261c0a1d2.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "TinyLlama/TinyLlama-1.1B-Chat-v1.0",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "LlamaForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "dtype": "bfloat16",
11
+ "head_dim": 64,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 2048,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 5632,
16
+ "max_position_embeddings": 2048,
17
+ "mlp_bias": false,
18
+ "model_type": "llama",
19
+ "neuron": {
20
+ "_serialized_key": "NxDNeuronConfig",
21
+ "batch_size": 1,
22
+ "capacity_factor": null,
23
+ "checkpoint_id": "TinyLlama/TinyLlama-1.1B-Chat-v1.0",
24
+ "checkpoint_revision": "fe8a4ea1ffedaf415f4da2f062534de366a451e6",
25
+ "continuous_batching": false,
26
+ "ep_degree": 1,
27
+ "fused_qkv": true,
28
+ "glu_mlp": true,
29
+ "local_ranks_size": 4,
30
+ "max_batch_size": 1,
31
+ "max_context_length": 2048,
32
+ "max_topk": 256,
33
+ "n_active_tokens": 2048,
34
+ "neuronxcc_version": "2.21.33363.0+82129205",
35
+ "on_device_sampling": true,
36
+ "optimum_neuron_version": "0.4.6.dev2",
37
+ "output_logits": false,
38
+ "pp_degree": 1,
39
+ "sequence_length": 2048,
40
+ "speculation_length": 0,
41
+ "start_rank_id": 0,
42
+ "target": "trn2",
43
+ "torch_dtype": "bfloat16",
44
+ "tp_degree": 4
45
+ },
46
+ "num_attention_heads": 32,
47
+ "num_hidden_layers": 22,
48
+ "num_key_value_heads": 4,
49
+ "pretraining_tp": 1,
50
+ "rms_norm_eps": 1e-05,
51
+ "rope_scaling": null,
52
+ "rope_theta": 10000.0,
53
+ "tie_word_embeddings": false,
54
+ "use_cache": true,
55
+ "vocab_size": 32000
56
+ }
neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+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_079584b98bad9058ec73+ac10809c/model.done ADDED
File without changes
neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dcf0969af0da81fa89f8a4c3a343dfd40ea343a503ab65be610419b1ce133ed9
3
+ size 1582347
neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:829bad29b447dbfe943dbb14637a83e9290807e258423c6d4915166fa840ac67
3
+ size 1926144
neuronxcc-2.21.33363.0+82129205/MODULE_079584b98bad9058ec73+ac10809c/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79b459219ad0892e93edb933568adea083c494057632823ca5422eb297939f81
3
+ size 2028607