dacorvo HF Staff commited on
Commit
d5dbec9
·
verified ·
1 Parent(s): 5f8c738

Synchronizing local compiler cache.

Browse files
.gitattributes CHANGED
@@ -12447,3 +12447,6 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_51a05ae109f810658ae5+24129607/model.neff
12447
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_6344e5c7baff82429569+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12448
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_6344e5c7baff82429569+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12449
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_6cb882531e16256891e9+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
 
 
 
 
12447
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_6344e5c7baff82429569+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12448
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_6344e5c7baff82429569+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12449
  neuronxcc-2.21.18209.0+043b1bf7/MODULE_6cb882531e16256891e9+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
12450
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_6b8d8e17358ff492e38a+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
12451
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_6b8d8e17358ff492e38a+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
12452
+ neuronxcc-2.21.18209.0+043b1bf7/MODULE_c07fbd091d817b7fd101+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.21.18209.0+043b1bf7/0_REGISTRY/0.4.1.dev0/llama/meta-llama/Llama-3.3-70B-Instruct/338d5297393576e0ef65.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "meta-llama/Llama-3.3-70B-Instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "LlamaForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "head_dim": 128,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 8192,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 28672,
15
+ "max_position_embeddings": 131072,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "neuron": {
19
+ "_serialized_key": "NxDNeuronConfig",
20
+ "batch_size": 1,
21
+ "capacity_factor": null,
22
+ "checkpoint_id": "meta-llama/Llama-3.3-70B-Instruct",
23
+ "checkpoint_revision": "6f6073b423013f6a7d4d9f39144961bfbfbc386b",
24
+ "continuous_batching": false,
25
+ "ep_degree": 1,
26
+ "fused_qkv": true,
27
+ "glu_mlp": true,
28
+ "local_ranks_size": 32,
29
+ "max_batch_size": 1,
30
+ "max_context_length": 4096,
31
+ "max_topk": 256,
32
+ "n_active_tokens": 4096,
33
+ "neuronxcc_version": "2.21.18209.0+043b1bf7",
34
+ "on_device_sampling": true,
35
+ "optimum_neuron_version": "0.4.1.dev0",
36
+ "output_logits": false,
37
+ "pp_degree": 1,
38
+ "sequence_length": 4096,
39
+ "speculation_length": 0,
40
+ "start_rank_id": 0,
41
+ "target": "trn1",
42
+ "torch_dtype": "bfloat16",
43
+ "tp_degree": 32
44
+ },
45
+ "num_attention_heads": 64,
46
+ "num_hidden_layers": 80,
47
+ "num_key_value_heads": 8,
48
+ "pretraining_tp": 1,
49
+ "rms_norm_eps": 1e-05,
50
+ "rope_scaling": {
51
+ "factor": 8.0,
52
+ "high_freq_factor": 4.0,
53
+ "low_freq_factor": 1.0,
54
+ "original_max_position_embeddings": 8192,
55
+ "rope_type": "llama3"
56
+ },
57
+ "rope_theta": 500000.0,
58
+ "tie_word_embeddings": false,
59
+ "use_cache": true,
60
+ "vocab_size": 128256
61
+ }
neuronxcc-2.21.18209.0+043b1bf7/MODULE_6b8d8e17358ff492e38a+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_6b8d8e17358ff492e38a+a02c3a36/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_6b8d8e17358ff492e38a+a02c3a36/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b87c4d8e38f738610b1dc24ffca9c87f31c8e8dd5548a3c3d02ab343f8f7efa9
3
+ size 1942919
neuronxcc-2.21.18209.0+043b1bf7/MODULE_6b8d8e17358ff492e38a+a02c3a36/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f9b206a557975fdf391256314d14ff8962e4c344b5aa9e8af51c6295c56e5c5
3
+ size 6288384
neuronxcc-2.21.18209.0+043b1bf7/MODULE_6b8d8e17358ff492e38a+a02c3a36/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e5ac2056338e0c9c6784d4b285d2bff2886f63495a336308180b59232f8b4af
3
+ size 6658732
neuronxcc-2.21.18209.0+043b1bf7/MODULE_c07fbd091d817b7fd101+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_c07fbd091d817b7fd101+24129607/model.done ADDED
File without changes
neuronxcc-2.21.18209.0+043b1bf7/MODULE_c07fbd091d817b7fd101+24129607/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e280a334c61a0297d1c9cc17e9033ae9dc8950a13e75b8940e4e46074100a830
3
+ size 2202991
neuronxcc-2.21.18209.0+043b1bf7/MODULE_c07fbd091d817b7fd101+24129607/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7e4f0cdae91788b4370a37e067eae63df741077e3b00e3bdd162ff36d3fb1894
3
+ size 2345984