Synchronizing local compiler cache.

.gitattributes

@@ -7147,3 +7147,5 @@ neuronxcc-2.16.372.0+4a9b2326/MODULE_8df375bd5f331802e94b/text_encoder/model.neu
 neuronxcc-2.16.372.0+4a9b2326/MODULE_8df375bd5f331802e94b/unet/model.neuron filter=lfs diff=lfs merge=lfs -text
 neuronxcc-2.16.372.0+4a9b2326/MODULE_8df375bd5f331802e94b/vae_decoder/model.neuron filter=lfs diff=lfs merge=lfs -text
 neuronxcc-2.16.372.0+4a9b2326/MODULE_8df375bd5f331802e94b/vae_encoder/model.neuron filter=lfs diff=lfs merge=lfs -text
+neuronxcc-2.16.372.0+4a9b2326/MODULE_cb3ceed5baf6cff96237+613edded/model.neff filter=lfs diff=lfs merge=lfs -text
+neuronxcc-2.16.372.0+4a9b2326/MODULE_f1c6fefbab0176eae280+613edded/model.neff filter=lfs diff=lfs merge=lfs -text

neuronxcc-2.16.372.0+4a9b2326/0_REGISTRY/0.1.0.dev1/inference/llama/meta-llama/Llama-3.1-70B-Instruct/491a48ca8c62ac951be1.json

@@ -0,0 +1 @@
+{"architectures": ["LlamaForCausalLM"], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": [128001, 128008, 128009], "head_dim": 128, "hidden_act": "silu", "hidden_size": 8192, "initializer_range": 0.02, "intermediate_size": 28672, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "neuron": {"auto_cast_type": "bf16", "batch_size": 1, "checkpoint_id": "meta-llama/Llama-3.1-70B-Instruct", "checkpoint_revision": "1605565b47bb9346c5515c34102e054115b4f98b", "compiler_type": "neuronx-cc", "compiler_version": "2.16.372.0+4a9b2326", "num_cores": 24, "sequence_length": 4096, "task": "text-generation"}, "num_attention_heads": 64, "num_hidden_layers": 80, "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-2.16.372.0+4a9b2326/MODULE_cb3ceed5baf6cff96237+613edded/compile_flags.json

@@ -0,0 +1 @@
+["--target=trn1", "--model-type=transformer", "--auto-cast=none", "--execute-repetition=1"]

neuronxcc-2.16.372.0+4a9b2326/MODULE_cb3ceed5baf6cff96237+613edded/model.hlo_module.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9681c5ebe71d6c78653309bb86bb63a49c56aeaf723ce7f3ed501f759c1f867d
+size 834051

neuronxcc-2.16.372.0+4a9b2326/MODULE_cb3ceed5baf6cff96237+613edded/model.neff

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6df2d7ecca3871ad1ee33ccad2b0c4bbc2f865057bda87adcb48973e0b2d604c
+size 8274944

neuronxcc-2.16.372.0+4a9b2326/MODULE_f1c6fefbab0176eae280+613edded/compile_flags.json

@@ -0,0 +1 @@
+["--target=trn1", "--model-type=transformer", "--auto-cast=none", "--execute-repetition=1"]

neuronxcc-2.16.372.0+4a9b2326/MODULE_f1c6fefbab0176eae280+613edded/model.hlo_module.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:493a9444012c3029aa60b731ae0ef447b7ac3d2a22051fa03a436f012b736182
+size 1002388

neuronxcc-2.16.372.0+4a9b2326/MODULE_f1c6fefbab0176eae280+613edded/model.neff

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ebdad0b9f12412c12989f678b30b4ae379adc98b154752235d3a22d5a0c750d
+size 2540544