{
  "_serialized_key": "NxDNeuronConfig",
  "async_mode": false,
  "attn_kernel_enabled": false,
  "batch_size": 4,
  "capacity_factor": null,
  "cc_pipeline_tiling_factor": 2,
  "checkpoint_id": "unsloth/Llama-3.2-1B-Instruct",
  "checkpoint_revision": "5a8abab4a5d6f164389b1079fb721cfab8d7126c",
  "continuous_batching": true,
  "enable_bucketing": false,
  "ep_degree": 1,
  "flash_decoding_enabled": false,
  "fused_qkv": true,
  "glu_mlp": true,
  "is_chunked_prefill": false,
  "local_ranks_size": 2,
  "logical_nc_config": 1,
  "max_batch_size": 4,
  "max_context_length": 2048,
  "max_topk": 256,
  "mlp_kernel_enabled": false,
  "mlp_kernel_fuse_residual_add": false,
  "n_active_tokens": 2048,
  "neuronxcc_version": "2.17.194.0+d312836f",
  "num_cores_per_group": 1,
  "on_device_sampling": false,
  "optimum_neuron_version": "0.2.0",
  "output_logits": false,
  "padding_side": "right",
  "pp_degree": 1,
  "qk_layernorm": false,
  "qkv_kernel_enabled": false,
  "rpl_reduce_dtype": "float16",
  "sequence_length": 2048,
  "sequence_parallel_enabled": false,
  "speculation_length": 0,
  "start_rank_id": 0,
  "target": null,
  "torch_dtype": "float16",
  "tp_degree": 2,
  "vocab_parallel": false
}