dacorvo HF Staff commited on
Commit
eb2ec40
·
verified ·
1 Parent(s): 20f1290

Synchronizing local compiler cache.

Browse files
.gitattributes CHANGED
@@ -7319,3 +7319,6 @@ neuronxcc-2.23.6484.0+3b612583/MODULE_b85127c63c8b6628b99b+84337dd9/wrapped_neff
7319
  neuronxcc-2.23.6484.0+3b612583/MODULE_09bce716582bd8588829+84337dd9/model.neff filter=lfs diff=lfs merge=lfs -text
7320
  neuronxcc-2.23.6484.0+3b612583/MODULE_09bce716582bd8588829+84337dd9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
7321
  neuronxcc-2.23.6484.0+3b612583/MODULE_42bf2b2812df73d93caa+519f203d/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
7319
  neuronxcc-2.23.6484.0+3b612583/MODULE_09bce716582bd8588829+84337dd9/model.neff filter=lfs diff=lfs merge=lfs -text
7320
  neuronxcc-2.23.6484.0+3b612583/MODULE_09bce716582bd8588829+84337dd9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
7321
  neuronxcc-2.23.6484.0+3b612583/MODULE_42bf2b2812df73d93caa+519f203d/model.neff filter=lfs diff=lfs merge=lfs -text
7322
+ neuronxcc-2.23.6484.0+3b612583/MODULE_2b240375da7319e33734+12dd9119/model.neff filter=lfs diff=lfs merge=lfs -text
7323
+ neuronxcc-2.23.6484.0+3b612583/MODULE_7d80f89ebf06772a2d78+84337dd9/model.neff filter=lfs diff=lfs merge=lfs -text
7324
+ neuronxcc-2.23.6484.0+3b612583/MODULE_7d80f89ebf06772a2d78+84337dd9/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.23.6484.0+3b612583/0_REGISTRY/0.4.6.dev3/0cc25526e2cfc37a8875a3752f33c4d7505d8a07b869d0f3f41915cf6e763b74/c41b61f9f848c9e97f14.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 1,
24
+ "capacity_factor": null,
25
+ "checkpoint_id": "ibm-granite/granite-3.1-2b-instruct",
26
+ "checkpoint_revision": "bbc2aed595bd38bd770263dc3ab831db9794441d",
27
+ "continuous_batching": false,
28
+ "ep_degree": 1,
29
+ "fused_qkv": true,
30
+ "glu_mlp": true,
31
+ "local_ranks_size": 2,
32
+ "max_batch_size": 1,
33
+ "max_context_length": 8192,
34
+ "max_topk": 256,
35
+ "n_active_tokens": 8192,
36
+ "neuronxcc_version": "2.23.6484.0+3b612583",
37
+ "on_device_sampling": true,
38
+ "optimum_neuron_version": "0.4.6.dev3",
39
+ "output_logits": false,
40
+ "pp_degree": 1,
41
+ "prefill_chunk_size": 1024,
42
+ "sequence_length": 8192,
43
+ "speculation_length": 0,
44
+ "start_rank_id": 0,
45
+ "target": "trn1",
46
+ "torch_dtype": "bfloat16",
47
+ "tp_degree": 2
48
+ },
49
+ "num_attention_heads": 32,
50
+ "num_hidden_layers": 40,
51
+ "num_key_value_heads": 8,
52
+ "residual_multiplier": 0.22,
53
+ "rms_norm_eps": 1e-05,
54
+ "rope_scaling": null,
55
+ "rope_theta": 5000000.0,
56
+ "tie_word_embeddings": true,
57
+ "use_cache": true,
58
+ "vocab_size": 49155
59
+ }
neuronxcc-2.23.6484.0+3b612583/MODULE_2b240375da7319e33734+12dd9119/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/chunked_prefill/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.23.6484.0+3b612583/MODULE_2b240375da7319e33734+12dd9119/model.done ADDED
File without changes
neuronxcc-2.23.6484.0+3b612583/MODULE_2b240375da7319e33734+12dd9119/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd57cf4a665b2f986b101b2830b706408b6b9a51c2428d8a232073cc962c5c64
3
+ size 4776088
neuronxcc-2.23.6484.0+3b612583/MODULE_2b240375da7319e33734+12dd9119/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b6a7ebdae9fc0e2e87d639fdc8f5e2c7246cde92d3a63826c511ca4ca9b7ee2
3
+ size 12268544
neuronxcc-2.23.6484.0+3b612583/MODULE_7d80f89ebf06772a2d78+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-2.23.6484.0+3b612583/MODULE_7d80f89ebf06772a2d78+84337dd9/model.done ADDED
File without changes
neuronxcc-2.23.6484.0+3b612583/MODULE_7d80f89ebf06772a2d78+84337dd9/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8279ff7a01ea004692c908f81ef1d8a6cc3251f0beedfb0ff56b0d18c575adda
3
+ size 3615341
neuronxcc-2.23.6484.0+3b612583/MODULE_7d80f89ebf06772a2d78+84337dd9/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a92af443ff8ca4a2c9357b8741440adeaf5d88ceaa25836132b73532080d75ca
3
+ size 5868544
neuronxcc-2.23.6484.0+3b612583/MODULE_7d80f89ebf06772a2d78+84337dd9/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5c87b07ca9141400052b6e540c4aeb2446fdacee6e40fff5bd92a1fd71d7a38c
3
+ size 6052723