groma-det-pretrain / config.json
machuofan
init
992f71f
raw
history blame
2.25 kB
{
"architectures": [
"CustomDDETRModel"
],
"ddetr_cfg": {
"activation_dropout": 0.0,
"activation_function": "relu",
"attention_dropout": 0.0,
"auxiliary_loss": true,
"backbone": "resnet50",
"backbone_config": null,
"bbox_cost": 5,
"bbox_loss_coefficient": 5,
"class_cost": 2,
"cls_loss_coefficient": 2,
"d_model": 256,
"decoder_attention_heads": 8,
"decoder_ffn_dim": 1024,
"decoder_layers": 6,
"decoder_n_points": 4,
"dice_loss_coefficient": 1,
"dilation": false,
"disable_custom_kernels": false,
"dropout": 0.1,
"encoder_attention_heads": 8,
"encoder_ffn_dim": 1024,
"encoder_layerdrop": 0.0,
"encoder_layers": 6,
"encoder_n_points": 4,
"eos_coefficient": 0.1,
"focal_alpha": 0.25,
"giou_cost": 2,
"giou_loss_coefficient": 2,
"id2label": {
"0": "LABEL_0"
},
"init_std": 0.02,
"init_xavier_std": 1.0,
"is_encoder_decoder": true,
"label2id": {
"LABEL_0": 0
},
"mask_loss_coefficient": 1,
"max_position_embeddings": 1024,
"model_type": "deformable_detr",
"num_channels": 3,
"num_feature_levels": 1,
"num_queries": 300,
"position_embedding_type": "sine",
"transformers_version": "4.32.0",
"two_stage": true,
"two_stage_num_proposals": 300,
"use_pretrained_backbone": true,
"use_timm_backbone": true,
"with_box_refine": true
},
"model_type": "ddetr",
"torch_dtype": "float32",
"transformers_version": "4.32.0",
"vis_encoder_cfg": {
"architectures": [
"Dinov2Model"
],
"attention_probs_dropout_prob": 0.0,
"drop_path_rate": 0.0,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.0,
"hidden_size": 1024,
"image_size": 518,
"initializer_range": 0.02,
"layer_norm_eps": 1e-06,
"layerscale_value": 1.0,
"mlp_ratio": 4,
"model_type": "dinov2",
"num_attention_heads": 16,
"num_channels": 3,
"num_hidden_layers": 24,
"patch_size": 14,
"qkv_bias": true,
"torch_dtype": "float32",
"transformers_version": "4.32.0",
"use_swiglu_ffn": false
},
"vis_encoder_path": "checkpoints/dinov2-large",
"vis_output_layer": -2,
"zs_weight_path": null
}