Instructions to use PekingU/rtdetr_r18vd_coco_o365 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use PekingU/rtdetr_r18vd_coco_o365 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("object-detection", model="PekingU/rtdetr_r18vd_coco_o365")# Load model directly from transformers import AutoImageProcessor, AutoModelForObjectDetection processor = AutoImageProcessor.from_pretrained("PekingU/rtdetr_r18vd_coco_o365") model = AutoModelForObjectDetection.from_pretrained("PekingU/rtdetr_r18vd_coco_o365", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Add config from convert_rt_detr_original_pytorch_checkpoint_to_pytorch.py
Browse files- config.json +4 -0
config.json
CHANGED
|
@@ -10,6 +10,7 @@
|
|
| 10 |
],
|
| 11 |
"attention_dropout": 0.0,
|
| 12 |
"auxiliary_loss": true,
|
|
|
|
| 13 |
"backbone_config": {
|
| 14 |
"depths": [
|
| 15 |
2,
|
|
@@ -36,6 +37,7 @@
|
|
| 36 |
4
|
| 37 |
]
|
| 38 |
},
|
|
|
|
| 39 |
"batch_norm_eps": 1e-05,
|
| 40 |
"box_noise_scale": 1.0,
|
| 41 |
"d_model": 256,
|
|
@@ -257,6 +259,8 @@
|
|
| 257 |
"torch_dtype": "float32",
|
| 258 |
"transformers_version": "4.42.0.dev0",
|
| 259 |
"use_focal_loss": true,
|
|
|
|
|
|
|
| 260 |
"weight_loss_bbox": 5.0,
|
| 261 |
"weight_loss_giou": 2.0,
|
| 262 |
"weight_loss_vfl": 1.0,
|
|
|
|
| 10 |
],
|
| 11 |
"attention_dropout": 0.0,
|
| 12 |
"auxiliary_loss": true,
|
| 13 |
+
"backbone": null,
|
| 14 |
"backbone_config": {
|
| 15 |
"depths": [
|
| 16 |
2,
|
|
|
|
| 37 |
4
|
| 38 |
]
|
| 39 |
},
|
| 40 |
+
"backbone_kwargs": null,
|
| 41 |
"batch_norm_eps": 1e-05,
|
| 42 |
"box_noise_scale": 1.0,
|
| 43 |
"d_model": 256,
|
|
|
|
| 259 |
"torch_dtype": "float32",
|
| 260 |
"transformers_version": "4.42.0.dev0",
|
| 261 |
"use_focal_loss": true,
|
| 262 |
+
"use_pretrained_backbone": false,
|
| 263 |
+
"use_timm_backbone": false,
|
| 264 |
"weight_loss_bbox": 5.0,
|
| 265 |
"weight_loss_giou": 2.0,
|
| 266 |
"weight_loss_vfl": 1.0,
|