rtdetr-r50-ball-handler-finetune

Gelmi/rtdetr-r50-ball-handler-finetune

rtdetr-r50-ball-handler-finetune is a object detection model from Gelmi. Use it when you need objects located in an image. It is set up for transformers. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

5% of all-time downloads

All-time downloads

316

Public

Parameters

42.9M

858 MB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors172 MB · 100%

At a glance

Task: Object Detection
Library: transformers
License: apache-2.0
Model type: rt_detr
Access: Public
Created: Feb 7, 2025
Updated: Feb 21, 2025
SHA: a4623d4c

Base models

PekingU/rtdetr_r50vd_coco_o365

Model card

From the Hugging Face model README

rtdetr-r50-ball-handler-finetune

This model is a fine-tuned version of PekingU/rtdetr_r50vd_coco_o365 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 12.9794
Map: 0.0003
Map 50: 0.0015
Map 75: 0.0
Map Small: 0.0
Map Medium: 0.0003
Map Large: 0.003
Mar 1: 0.0005
Mar 10: 0.0067
Mar 100: 0.0168
Mar Small: 0.0
Mar Medium: 0.0174
Mar Large: 0.0194
Map Player: 0.0006
Mar 100 Player: 0.0245
Map Ball-handler: 0.0
Mar 100 Ball-handler: 0.009

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 300
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map Small	Map Medium	Map Large	Mar 1	Mar 10	Mar 100	Mar Small	Mar Medium	Mar Large	Map Player	Mar 100 Player	Map Ball-handler	Mar 100 Ball-handler
No log	1.0	232	13.8912	0.0003	0.0013	-1.0	0.0005	0.0001	0.0004	0.0091	0.0195	-1.0	0.0304	0.0053	0.0004	0.0167	0.0002	0.0223
No log	2.0	464	11.6078	0.0002	0.001	-1.0	0.0004	0.0	0.0011	0.0105	0.0186	-1.0	0.029	0.0057	0.0003	0.0146	0.0001	0.0226
30.7716	3.0	696	11.0199	0.0003	0.0015	-1.0	0.0007	0.0	0.0029	0.0103	0.0204	-1.0	0.0332	0.0048	0.0003	0.0169	0.0003	0.0239
30.7716	4.0	928	10.8415	0.0002	0.0012	-1.0	0.0005	0.0	0.0012	0.0122	0.0188	-1.0	0.0304	0.005	0.0003	0.0146	0.0002	0.023
12.4494	5.0	1160	10.9467	0.0002	0.001	-1.0	0.0004	0.0	0.0011	0.0106	0.0168	-1.0	0.0282	0.0025	0.0003	0.0158	0.0001	0.0179
12.4494	6.0	1392	11.2840	0.0003	0.0013	-1.0	0.0006	0.0	0.002	0.0114	0.0207	-1.0	0.0348	0.0041	0.0003	0.0159	0.0002	0.0255
10.9839	7.0	1624	11.3669	0.0002	0.0011	-1.0	0.0005	0.0	0.0017	0.012	0.02	-1.0	0.0333	0.0035	0.0003	0.0171	0.0002	0.023
10.9839	8.0	1856	11.4150	0.0003	0.0012	-1.0	0.0006	0.0	0.0007	0.0111	0.0194	-1.0	0.0325	0.0031	0.0003	0.0164	0.0002	0.0223
10.0392	9.0	2088	11.2706	0.0002	0.0011	-1.0	0.0005	0.0	0.0019	0.0118	0.0198	-1.0	0.0337	0.0033	0.0003	0.0163	0.0002	0.0233
10.0392	10.0	2320	11.3800	0.0002	0.001	-1.0	0.0005	0.0001	0.0019	0.0112	0.0189	-1.0	0.0322	0.0029	0.0003	0.0155	0.0002	0.0223

Framework versions

Transformers 4.50.0.dev0
Pytorch 2.5.1+cu124
Datasets 3.3.2
Tokenizers 0.21.0

Task: Object Detection
Library: transformers
Type: rt_detr
License: apache-2.0
Created: Feb 7, 2025
Updated: Feb 21, 2025