optimum-neuron-cache
/
neuronxcc-2.12.68.0+4480452af
/0_REGISTRY
/0.0.20
/inference
/llama
/m-a-p
/OpenCodeInterpreter-DS-6.7B
/370009bdce014bbc567d.json
{"vocab_size": 32256, "max_position_embeddings": 16384, "hidden_size": 4096, "intermediate_size": 11008, "num_hidden_layers": 32, "num_attention_heads": 32, "num_key_value_heads": 32, "hidden_act": "silu", "initializer_range": 0.02, "rms_norm_eps": 1e-06, "pretraining_tp": 1, "use_cache": false, "rope_theta": 100000, "rope_scaling": {"factor": 4.0, "type": "linear"}, "attention_bias": false, "attention_dropout": 0.0, "torch_dtype": "bfloat16", "tie_word_embeddings": false, "architectures": ["LlamaForCausalLM"], "bos_token_id": 32013, "eos_token_id": 32021, "model_type": "llama", "neuron": {"task": "text-generation", "batch_size": 4, "num_cores": 24, "auto_cast_type": "fp16", "sequence_length": 4096, "compiler_type": "neuronx-cc", "compiler_version": "2.12.68.0+4480452af", "checkpoint_id": "m-a-p/OpenCodeInterpreter-DS-6.7B", "checkpoint_revision": "60b89884df814590abd76757a6db4a527cbdfc91"}} |