optimum-internal-testing-user commited on
Commit
2410aad
·
verified ·
1 Parent(s): dede982

Synchronizing local compiler cache.

Browse files
.gitattributes CHANGED
@@ -5067,3 +5067,6 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_73b50e8e61409e1f8fe1+ed72d204/model.neff
5067
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_50d30ce495feb34fe5c1+a9d440f5/model.neff filter=lfs diff=lfs merge=lfs -text
5068
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_50d30ce495feb34fe5c1+a9d440f5/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5069
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_cae14ab78f54c8b4e73f+ed72d204/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
5067
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_50d30ce495feb34fe5c1+a9d440f5/model.neff filter=lfs diff=lfs merge=lfs -text
5068
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_50d30ce495feb34fe5c1+a9d440f5/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5069
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_cae14ab78f54c8b4e73f+ed72d204/model.neff filter=lfs diff=lfs merge=lfs -text
5070
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_5b7172cc10110a470457+ed72d204/model.neff filter=lfs diff=lfs merge=lfs -text
5071
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_9361fb1ac241d0d047f1+a9d440f5/model.neff filter=lfs diff=lfs merge=lfs -text
5072
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_9361fb1ac241d0d047f1+a9d440f5/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.0.dev0/granite/ibm-granite/granite-3.1-2b-instruct/ce959172dbc6a16df53b.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "ibm-granite/granite-3.1-2b-instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "GraniteForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.1,
10
+ "attention_multiplier": 0.015625,
11
+ "embedding_multiplier": 12.0,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 2048,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 8192,
16
+ "logits_scaling": 8.0,
17
+ "max_position_embeddings": 131072,
18
+ "mlp_bias": false,
19
+ "model_type": "granite",
20
+ "neuron": {
21
+ "_serialized_key": "NxDNeuronConfig",
22
+ "batch_size": 4,
23
+ "capacity_factor": null,
24
+ "checkpoint_id": "ibm-granite/granite-3.1-2b-instruct",
25
+ "checkpoint_revision": "bbc2aed595bd38bd770263dc3ab831db9794441d",
26
+ "continuous_batching": true,
27
+ "enable_bucketing": false,
28
+ "ep_degree": 1,
29
+ "fused_qkv": true,
30
+ "glu_mlp": true,
31
+ "local_ranks_size": 2,
32
+ "max_batch_size": 4,
33
+ "max_context_length": 4096,
34
+ "max_topk": 256,
35
+ "n_active_tokens": 4096,
36
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.0.dev0",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "sequence_length": 4096,
42
+ "speculation_length": 0,
43
+ "start_rank_id": 0,
44
+ "target": "trn1",
45
+ "torch_dtype": "bfloat16",
46
+ "tp_degree": 2
47
+ },
48
+ "num_attention_heads": 32,
49
+ "num_hidden_layers": 40,
50
+ "num_key_value_heads": 8,
51
+ "residual_multiplier": 0.22,
52
+ "rms_norm_eps": 1e-05,
53
+ "rope_scaling": null,
54
+ "rope_theta": 5000000.0,
55
+ "tie_word_embeddings": true,
56
+ "use_cache": true,
57
+ "vocab_size": 49155
58
+ }
neuronxcc-2.21.18209.0+043b1bf7/MODULE_5b7172cc10110a470457+ed72d204/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_model/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_5b7172cc10110a470457+ed72d204/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_5b7172cc10110a470457+ed72d204/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e44bfe923734a9e692c83a3f6f8bf4e15fc148249c77d1aef3fde47ff55a5d1
3
+ size 1050679
neuronxcc-2.21.18209.0+043b1bf7/MODULE_5b7172cc10110a470457+ed72d204/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e28886991d12741447b30efa8893e0977fe2de2c7d53584535b6c26f97bc50df
3
+ size 5827584
neuronxcc-2.21.18209.0+043b1bf7/MODULE_9361fb1ac241d0d047f1+a9d440f5/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_model/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_9361fb1ac241d0d047f1+a9d440f5/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_9361fb1ac241d0d047f1+a9d440f5/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea7ea4024512779624a08b9442447c93f47d17ef5b03b786fe5ea61a5b733834
3
+ size 1072107
neuronxcc-2.21.18209.0+043b1bf7/MODULE_9361fb1ac241d0d047f1+a9d440f5/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d94268e27a81f1ee6de2f8082852f047b480ebd477a491c886d40b594c26b385
3
+ size 6974464
neuronxcc-2.21.18209.0+043b1bf7/MODULE_9361fb1ac241d0d047f1+a9d440f5/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bc1719e07caf3b44da4a897ec365164231cb566c062d8c7a3a02bef630183db
3
+ size 7158800