Commit
30019a3
·
verified ·
1 Parent(s): b57b986

Synchronizing local compiler cache.

Browse files
.gitattributes CHANGED
@@ -5913,3 +5913,6 @@ neuronxcc-2.22.12471.0+b4a00d10/MODULE_dd85f09172e533f22911+519f203d/model.neff
5913
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_4194ae73faf2eec0c203+84337dd9/model.neff filter=lfs diff=lfs merge=lfs -text
5914
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_4194ae73faf2eec0c203+84337dd9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5915
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_7b3a5bcf0a5c3648a3ef+519f203d/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
5913
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_4194ae73faf2eec0c203+84337dd9/model.neff filter=lfs diff=lfs merge=lfs -text
5914
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_4194ae73faf2eec0c203+84337dd9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5915
  neuronxcc-2.22.12471.0+b4a00d10/MODULE_7b3a5bcf0a5c3648a3ef+519f203d/model.neff filter=lfs diff=lfs merge=lfs -text
5916
+ neuronxcc-2.22.12471.0+b4a00d10/MODULE_4ba9e4fd8d52ab368b5e+84337dd9/model.neff filter=lfs diff=lfs merge=lfs -text
5917
+ neuronxcc-2.22.12471.0+b4a00d10/MODULE_4ba9e4fd8d52ab368b5e+84337dd9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5918
+ neuronxcc-2.22.12471.0+b4a00d10/MODULE_f6ee548e8c515626bba8+519f203d/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.22.12471.0+b4a00d10/0_REGISTRY/0.4.4.dev2/granite/ibm-granite/granite-3.1-2b-instruct/414cb6d116ea07a522b5.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
+ "dtype": "bfloat16",
12
+ "embedding_multiplier": 12.0,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 2048,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 8192,
17
+ "logits_scaling": 8.0,
18
+ "max_position_embeddings": 131072,
19
+ "mlp_bias": false,
20
+ "model_type": "granite",
21
+ "neuron": {
22
+ "_serialized_key": "NxDNeuronConfig",
23
+ "batch_size": 4,
24
+ "capacity_factor": null,
25
+ "checkpoint_id": "ibm-granite/granite-3.1-2b-instruct",
26
+ "checkpoint_revision": "bbc2aed595bd38bd770263dc3ab831db9794441d",
27
+ "continuous_batching": true,
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": "2.22.12471.0+b4a00d10",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.4.dev2",
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-2.22.12471.0+b4a00d10/MODULE_4ba9e4fd8d52ab368b5e+84337dd9/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", "--verbose=35", "--logfile=/tmp/nxd_model/token_generation/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.22.12471.0+b4a00d10/MODULE_4ba9e4fd8d52ab368b5e+84337dd9/model.done ADDED
File without changes
neuronxcc-2.22.12471.0+b4a00d10/MODULE_4ba9e4fd8d52ab368b5e+84337dd9/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:369f3e8495530c7db7879504c7843cb117b8a42d0f7b22820bb989dc0b88d860
3
+ size 4385291
neuronxcc-2.22.12471.0+b4a00d10/MODULE_4ba9e4fd8d52ab368b5e+84337dd9/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfa7444c473767da29072267b21f5b6f4703d5eca5a7de16402375c413a62371
3
+ size 7363584
neuronxcc-2.22.12471.0+b4a00d10/MODULE_4ba9e4fd8d52ab368b5e+84337dd9/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9f9d355b301348642d558581152246a02647b7924d870e476deea954814373e
3
+ size 7548323
neuronxcc-2.22.12471.0+b4a00d10/MODULE_f6ee548e8c515626bba8+519f203d/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", "--verbose=35", "--logfile=/tmp/nxd_model/context_encoding/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.22.12471.0+b4a00d10/MODULE_f6ee548e8c515626bba8+519f203d/model.done ADDED
File without changes
neuronxcc-2.22.12471.0+b4a00d10/MODULE_f6ee548e8c515626bba8+519f203d/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b722f7add7922454bfd87ef970e71791303146a1d8bd476dde15a8104a76e918
3
+ size 3908423
neuronxcc-2.22.12471.0+b4a00d10/MODULE_f6ee548e8c515626bba8+519f203d/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a13288ed00fd7b1219d04ab81a6956f6d762b91eb6a6d09b66fce20229a09618
3
+ size 3677184