Synchronizing local compiler cache.

.gitattributes

@@ -5335,3 +5335,5 @@ neuronxcc-2.15.128.0+56dc5a86/MODULE_8a9089ea3d3598a213b5+39f12043/model.neff fi
 neuronxcc-2.15.128.0+56dc5a86/MODULE_dc8c775621fadb1249e0+39f12043/model.neff filter=lfs diff=lfs merge=lfs -text
 neuronxcc-2.15.128.0+56dc5a86/MODULE_df6ea342aff6a93d5c1a+39f12043/model.neff filter=lfs diff=lfs merge=lfs -text
 neuronxcc-2.15.128.0+56dc5a86/MODULE_f122ab1946e5da931afe+39f12043/model.neff filter=lfs diff=lfs merge=lfs -text
+neuronxcc-2.15.128.0+56dc5a86/MODULE_33299df8f92d4e4a5e71+39f12043/model.neff filter=lfs diff=lfs merge=lfs -text
+neuronxcc-2.15.128.0+56dc5a86/MODULE_8993a0d9f11bf2f76885+39f12043/model.neff filter=lfs diff=lfs merge=lfs -text

neuronxcc-2.15.128.0+56dc5a86/0_REGISTRY/0.0.25.dev0/inference/llama/meta-llama/Meta-Llama-3.1-8B/493b6ca8d87ccdaf398d.json

@@ -0,0 +1 @@
+{"architectures": ["LlamaForCausalLM"], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "neuron": {"auto_cast_type": "bf16", "batch_size": 4, "checkpoint_id": "meta-llama/Meta-Llama-3.1-8B", "checkpoint_revision": "8d10549bcf802355f2d6203a33ed27e81b15b9e5", "compiler_type": "neuronx-cc", "compiler_version": "2.15.128.0+56dc5a86", "num_cores": 2, "sequence_length": 4096, "task": "text-generation"}, "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": {"factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3"}, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "use_cache": true, "vocab_size": 128256}

neuronxcc-2.15.128.0+56dc5a86/MODULE_33299df8f92d4e4a5e71+39f12043/compile_flags.json

@@ -0,0 +1 @@
+["--model-type=transformer", "--auto-cast=none", "--execute-repetition=1"]

neuronxcc-2.15.128.0+56dc5a86/MODULE_33299df8f92d4e4a5e71+39f12043/model.hlo_module.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:897c7775a15eb9c250a2376c32da490a3d7fe7856cecafb1b5b3dc2cca285950
+size 451078

neuronxcc-2.15.128.0+56dc5a86/MODULE_33299df8f92d4e4a5e71+39f12043/model.neff

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:351cb529a8a4f0161e90abb579b5a1b4745d811ca95a74b7bdb6e3ad40b3a7ef
+size 6063104

neuronxcc-2.15.128.0+56dc5a86/MODULE_8993a0d9f11bf2f76885+39f12043/compile_flags.json

@@ -0,0 +1 @@
+["--model-type=transformer", "--auto-cast=none", "--execute-repetition=1"]

neuronxcc-2.15.128.0+56dc5a86/MODULE_8993a0d9f11bf2f76885+39f12043/model.hlo_module.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:625543ae9eed97fe79767bd486072291b04f615582a54fb54d367b890c228cc8
+size 363692

neuronxcc-2.15.128.0+56dc5a86/MODULE_8993a0d9f11bf2f76885+39f12043/model.neff

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e513b003361fb9d897cce436fecb095cd70ec22c4bdb1b15569c06151bd5332b
+size 10456064