dacorvo HF Staff commited on
Commit
f52de34
·
verified ·
1 Parent(s): 672d64c

Synchronizing local compiler cache.

Browse files
.gitattributes CHANGED
@@ -5197,3 +5197,6 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_d961fefa73e341c88838+747527b0/wrapped_nef
5197
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_1a286be93a0c3641f5b3+747527b0/model.neff filter=lfs diff=lfs merge=lfs -text
5198
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_1a286be93a0c3641f5b3+747527b0/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5199
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_dcfbae5d14fe3d7bc51b+877608f3/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
5197
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_1a286be93a0c3641f5b3+747527b0/model.neff filter=lfs diff=lfs merge=lfs -text
5198
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_1a286be93a0c3641f5b3+747527b0/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5199
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_dcfbae5d14fe3d7bc51b+877608f3/model.neff filter=lfs diff=lfs merge=lfs -text
5200
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_587bc590d8801c17f369+747527b0/model.neff filter=lfs diff=lfs merge=lfs -text
5201
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_587bc590d8801c17f369+747527b0/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
5202
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_c831c878efbfbe7a9c25+877608f3/model.neff 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/64a790f83e0ac2530718.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": 4,
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": "trn2",
45
+ "torch_dtype": "bfloat16",
46
+ "tp_degree": 4
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_587bc590d8801c17f369+747527b0/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_model/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_587bc590d8801c17f369+747527b0/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_587bc590d8801c17f369+747527b0/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b696f9f7074d5455a31f46ea93f51268ea495bc86a0ad6dc3095c5f4084049a
3
+ size 2333362
neuronxcc-2.21.18209.0+043b1bf7/MODULE_587bc590d8801c17f369+747527b0/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:70ab87c0901d401248ffbdf2cab776b32d6d42cf0cf83a7a1076f36ae04790e3
3
+ size 5213184
neuronxcc-2.21.18209.0+043b1bf7/MODULE_587bc590d8801c17f369+747527b0/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d476a1c960a97ab2443b433199cab6e8aaa9d4a3df8008afbcfc8bf3a63f76f
3
+ size 5400869
neuronxcc-2.21.18209.0+043b1bf7/MODULE_c831c878efbfbe7a9c25+877608f3/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/context_encoding_model/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.21.18209.0+043b1bf7/MODULE_c831c878efbfbe7a9c25+877608f3/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_c831c878efbfbe7a9c25+877608f3/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9055335c2580e196d735081b0fded88162b24529feb85279b2685e5411ada9d
3
+ size 2018688
neuronxcc-2.21.18209.0+043b1bf7/MODULE_c831c878efbfbe7a9c25+877608f3/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d863a77cefbafdfdad8b7a79ec057f93db8a7c62b5c253b6df9ff1584ff2ab8f
3
+ size 44114944