dacorvo HF Staff commited on
Commit
e831c27
·
verified ·
1 Parent(s): 35b93f5

Synchronizing local compiler cache.

Browse files
.gitattributes CHANGED
@@ -12286,3 +12286,6 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_bea3b475cb189ba1a56f+24129607/model.neff
12286
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_3569c16aad6f0ef9ff19+ac10809c/model.neff filter=lfs diff=lfs merge=lfs -text
12287
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_3569c16aad6f0ef9ff19+ac10809c/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12288
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_77883b7063ddcf2b753d+a32116a7/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
12286
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_3569c16aad6f0ef9ff19+ac10809c/model.neff filter=lfs diff=lfs merge=lfs -text
12287
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_3569c16aad6f0ef9ff19+ac10809c/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12288
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_77883b7063ddcf2b753d+a32116a7/model.neff filter=lfs diff=lfs merge=lfs -text
12289
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e2bc35a46f9ecb66349+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12290
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_f461f5c4551deff20e74+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12291
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_f461f5c4551deff20e74+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.1.dev0/granite/ibm-granite/granite-3.1-2b-instruct/99a22f2841a59116bc82.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 32,
23
+ "capacity_factor": null,
24
+ "checkpoint_id": "ibm-granite/granite-3.1-2b-instruct",
25
+ "checkpoint_revision": "bbc2aed595bd38bd770263dc3ab831db9794441d",
26
+ "continuous_batching": true,
27
+ "ep_degree": 1,
28
+ "fused_qkv": true,
29
+ "glu_mlp": true,
30
+ "local_ranks_size": 8,
31
+ "max_batch_size": 32,
32
+ "max_context_length": 4096,
33
+ "max_topk": 256,
34
+ "n_active_tokens": 4096,
35
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
36
+ "on_device_sampling": true,
37
+ "optimum_neuron_version": "0.4.1.dev0",
38
+ "output_logits": false,
39
+ "pp_degree": 1,
40
+ "sequence_length": 4096,
41
+ "speculation_length": 0,
42
+ "start_rank_id": 0,
43
+ "target": "trn1",
44
+ "torch_dtype": "bfloat16",
45
+ "tp_degree": 8
46
+ },
47
+ "num_attention_heads": 32,
48
+ "num_hidden_layers": 40,
49
+ "num_key_value_heads": 8,
50
+ "residual_multiplier": 0.22,
51
+ "rms_norm_eps": 1e-05,
52
+ "rope_scaling": null,
53
+ "rope_theta": 5000000.0,
54
+ "tie_word_embeddings": true,
55
+ "use_cache": true,
56
+ "vocab_size": 49155
57
+ }
neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e2bc35a46f9ecb66349+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_0e2bc35a46f9ecb66349+24129607/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e2bc35a46f9ecb66349+24129607/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6dd8461041d075bef545808774c894f68ee6848086c2cc8064754b00c858a867
3
+ size 997471
neuronxcc-2.21.18209.0+043b1bf7/MODULE_0e2bc35a46f9ecb66349+24129607/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07e0a5d8e6c0bd7c15db47325d5909e57abb916b7b2131e64ba97022c7bba1ef
3
+ size 25201664
neuronxcc-2.21.18209.0+043b1bf7/MODULE_f461f5c4551deff20e74+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_f461f5c4551deff20e74+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_f461f5c4551deff20e74+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c6c5805c176e02671b777aa97df3fc3fc0e885ef8b5edfa57a80c6f0e00d86a
3
+ size 1014720
neuronxcc-2.21.18209.0+043b1bf7/MODULE_f461f5c4551deff20e74+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b4d4dd1d1732ccf14aaa71fb41ed601c079ed4f5ee461239c62cf6137ece03e
3
+ size 7158784
neuronxcc-2.21.18209.0+043b1bf7/MODULE_f461f5c4551deff20e74+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:835e29f5a1f8f021259372b21d6994fee871a0264c294f333bbc3ac2c0e27177
3
+ size 7343546