config.json

{
  "architectures": [
    "VitPoseBackbone"
  ],
  "attention_probs_dropout_prob": 0.1,
  "dtype": "float32",
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.1,
  "hidden_size": 32,
  "image_size": [
    128,
    96
  ],
  "initializer_range": 0.02,
  "intermediate_size": 37,
  "layer_norm_eps": 1e-12,
  "mlp_ratio": 4,
  "model_type": "vitpose_backbone",
  "num_attention_heads": 4,
  "num_channels": 3,
  "num_experts": 1,
  "num_hidden_layers": 2,
  "out_features": [
    "stage2"
  ],
  "out_indices": [
    2
  ],
  "part_features": 256,
  "patch_size": [
    8,
    8
  ],
  "qkv_bias": true,
  "stage_names": [
    "stem",
    "stage1",
    "stage2"
  ],
  "transformers_version": "5.16.0.dev0"
}