dfine_coco2017-1k

This model is a fine-tuned version of ustc-community/dfine-large-obj365 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.0118
• Map: 0.0299
• Map 50: 0.0372
• Map 75: 0.0329
• Map Airplane: 0.0221
• Map Apple: 0.0
• Map Backpack: 0.0021
• Map Banana: 0.0063
• Map Baseball bat: 0.0
• Map Baseball glove: 0.1984
• Map Bear: 0.0
• Map Bed: 0.0
• Map Bench: 0.0009
• Map Bicycle: 0.0
• Map Bird: 0.0016
• Map Boat: 0.0123
• Map Book: 0.0002
• Map Bottle: 0.046
• Map Bowl: 0.0588
• Map Broccoli: 0.0007
• Map Bus: 0.3589
• Map Cake: 0.0142
• Map Car: 0.1625
• Map Carrot: 0.0
• Map Cat: 0.0047
• Map Cell phone: 0.08
• Map Chair: 0.0741
• Map Clock: 0.001
• Map Couch: 0.0125
• Map Cow: 0.1326
• Map Cup: 0.0592
• Map Dining table: 0.0596
• Map Dog: 0.0
• Map Donut: -1.0
• Map Elephant: 0.0034
• Map Fire hydrant: -1.0
• Map Fork: 0.0027
• Map Frisbee: 0.0008
• Map Giraffe: 0.0976
• Map Handbag: 0.0022
• Map Horse: 0.0022
• Map Hot dog: 0.0
• Map Keyboard: 0.0
• Map Kite: 0.008
• Map Knife: 0.001
• Map Laptop: 0.0009
• Map Large: 0.0435
• Map Medium: 0.0369
• Map Microwave: 0.0
• Map Motorcycle: 0.0006
• Map Mouse: 0.0
• Map Orange: 0.0
• Map Oven: 0.0
• Map Parking meter: 0.0
• Map Person: 0.1633
• Map Pizza: 0.0451
• Map Potted plant: 0.0026
• Map Refrigerator: 0.0013
• Map Remote: 0.0053
• Map Sandwich: 0.0
• Map Scissors: 0.0
• Map Sheep: 0.0
• Map Sink: 0.0001
• Map Skateboard: 0.0
• Map Skis: 0.0049
• Map Small: 0.0085
• Map Snowboard: 0.0
• Map Spoon: 0.0
• Map Sports ball: 0.0
• Map Stop sign: 0.006
• Map Suitcase: 0.0009
• Map Surfboard: 0.0027
• Map Teddy bear: 0.0
• Map Tennis racket: 0.1978
• Map Tie: 0.0
• Map Toaster: -1.0
• Map Toilet: 0.0101
• Map Toothbrush: 0.0
• Map Traffic light: 0.0388
• Map Train: 0.0096
• Map Truck: 0.108
• Map Tv: 0.2457
• Map Umbrella: 0.0001
• Map Vase: 0.0002
• Map Wine glass: 0.0
• Map Zebra: 0.0015
• Mar 1: 0.1037
• Mar 10: 0.1704
• Mar 100: 0.1887
• Mar 100 Airplane: 0.6857
• Mar 100 Apple: 0.0
• Mar 100 Backpack: 0.0333
• Mar 100 Banana: 0.0545
• Mar 100 Baseball bat: 0.0
• Mar 100 Baseball glove: 0.3438
• Mar 100 Bear: 0.0
• Mar 100 Bed: 0.0
• Mar 100 Bench: 0.275
• Mar 100 Bicycle: 0.0
• Mar 100 Bird: 0.5
• Mar 100 Boat: 0.2408
• Mar 100 Book: 0.0344
• Mar 100 Bottle: 0.3444
• Mar 100 Bowl: 0.224
• Mar 100 Broccoli: 0.1143
• Mar 100 Bus: 0.7917
• Mar 100 Cake: 0.25
• Mar 100 Car: 0.3451
• Mar 100 Carrot: 0.0
• Mar 100 Cat: 0.3692
• Mar 100 Cell phone: 0.1474
• Mar 100 Chair: 0.2458
• Mar 100 Clock: 0.1412
• Mar 100 Couch: 0.3455
• Mar 100 Cow: 0.4625
• Mar 100 Cup: 0.2686
• Mar 100 Dining table: 0.45
• Mar 100 Dog: 0.0
• Mar 100 Donut: -1.0
• Mar 100 Elephant: 0.5111
• Mar 100 Fire hydrant: -1.0
• Mar 100 Fork: 0.0643
• Mar 100 Frisbee: 0.6333
• Mar 100 Giraffe: 0.9556
• Mar 100 Handbag: 0.136
• Mar 100 Horse: 0.6
• Mar 100 Hot dog: 0.0
• Mar 100 Keyboard: 0.0
• Mar 100 Kite: 0.1611
• Mar 100 Knife: 0.0333
• Mar 100 Laptop: 0.1222
• Mar 100 Microwave: 0.0
• Mar 100 Motorcycle: 0.1437
• Mar 100 Mouse: 0.0
• Mar 100 Orange: 0.0
• Mar 100 Oven: 0.0
• Mar 100 Parking meter: 0.0
• Mar 100 Person: 0.4451
• Mar 100 Pizza: 0.16
• Mar 100 Potted plant: 0.205
• Mar 100 Refrigerator: 0.1286
• Mar 100 Remote: 0.0905
• Mar 100 Sandwich: 0.0
• Mar 100 Scissors: 0.0
• Mar 100 Sheep: 0.0
• Mar 100 Sink: 0.05
• Mar 100 Skateboard: 0.0
• Mar 100 Skis: 0.0875
• Mar 100 Snowboard: 0.0
• Mar 100 Spoon: 0.0
• Mar 100 Sports ball: 0.0
• Mar 100 Stop sign: 0.225
• Mar 100 Suitcase: 0.225
• Mar 100 Surfboard: 0.3818
• Mar 100 Teddy bear: 0.0
• Mar 100 Tennis racket: 0.5
• Mar 100 Tie: 0.0
• Mar 100 Toaster: -1.0
• Mar 100 Toilet: 0.2529
• Mar 100 Toothbrush: 0.0
• Mar 100 Traffic light: 0.1083
• Mar 100 Train: 0.1111
• Mar 100 Truck: 0.569
• Mar 100 Tv: 0.5615
• Mar 100 Umbrella: 0.0353
• Mar 100 Vase: 0.0833
• Mar 100 Wine glass: 0.0
• Mar 100 Zebra: 0.0909
• Mar Large: 0.2753
• Mar Medium: 0.1135
• Mar Small: 0.0232

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: 1
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Airplane	Map Apple	Map Backpack	Map Banana	Map Baseball bat	Map Baseball glove	Map Bear	Map Bed	Map Bench	Map Bicycle	Map Bird	Map Boat	Map Book	Map Bottle	Map Bowl	Map Broccoli	Map Bus	Map Cake	Map Car	Map Carrot	Map Cat	Map Cell phone	Map Chair	Map Clock	Map Couch	Map Cow	Map Cup	Map Dining table	Map Dog	Map Donut	Map Elephant	Map Fire hydrant	Map Fork	Map Frisbee	Map Giraffe	Map Handbag	Map Horse	Map Hot dog	Map Keyboard	Map Kite	Map Knife	Map Laptop	Map Large	Map Medium	Map Microwave	Map Motorcycle	Map Mouse	Map Orange	Map Oven	Map Parking meter	Map Person	Map Pizza	Map Potted plant	Map Refrigerator	Map Remote	Map Sandwich	Map Scissors	Map Sheep	Map Sink	Map Skateboard	Map Skis	Map Small	Map Snowboard	Map Spoon	Map Sports ball	Map Stop sign	Map Suitcase	Map Surfboard	Map Teddy bear	Map Tennis racket	Map Tie	Map Toaster	Map Toilet	Map Toothbrush	Map Traffic light	Map Train	Map Truck	Map Tv	Map Umbrella	Map Vase	Map Wine glass	Map Zebra	Mar 1	Mar 10	Mar 100	Mar 100 Airplane	Mar 100 Apple	Mar 100 Backpack	Mar 100 Banana	Mar 100 Baseball bat	Mar 100 Baseball glove	Mar 100 Bear	Mar 100 Bed	Mar 100 Bench	Mar 100 Bicycle	Mar 100 Bird	Mar 100 Boat	Mar 100 Book	Mar 100 Bottle	Mar 100 Bowl	Mar 100 Broccoli	Mar 100 Bus	Mar 100 Cake	Mar 100 Car	Mar 100 Carrot	Mar 100 Cat	Mar 100 Cell phone	Mar 100 Chair	Mar 100 Clock	Mar 100 Couch	Mar 100 Cow	Mar 100 Cup	Mar 100 Dining table	Mar 100 Dog	Mar 100 Donut	Mar 100 Elephant	Mar 100 Fire hydrant	Mar 100 Fork	Mar 100 Frisbee	Mar 100 Giraffe	Mar 100 Handbag	Mar 100 Horse	Mar 100 Hot dog	Mar 100 Keyboard	Mar 100 Kite	Mar 100 Knife	Mar 100 Laptop	Mar 100 Microwave	Mar 100 Motorcycle	Mar 100 Mouse	Mar 100 Orange	Mar 100 Oven	Mar 100 Parking meter	Mar 100 Person	Mar 100 Pizza	Mar 100 Potted plant	Mar 100 Refrigerator	Mar 100 Remote	Mar 100 Sandwich	Mar 100 Scissors	Mar 100 Sheep	Mar 100 Sink	Mar 100 Skateboard	Mar 100 Skis	Mar 100 Snowboard	Mar 100 Spoon	Mar 100 Sports ball	Mar 100 Stop sign	Mar 100 Suitcase	Mar 100 Surfboard	Mar 100 Teddy bear	Mar 100 Tennis racket	Mar 100 Tie	Mar 100 Toaster	Mar 100 Toilet	Mar 100 Toothbrush	Mar 100 Traffic light	Mar 100 Train	Mar 100 Truck	Mar 100 Tv	Mar 100 Umbrella	Mar 100 Vase	Mar 100 Wine glass	Mar 100 Zebra	Mar Large	Mar Medium	Mar Small
76.1385	1.0	740	3.0118	0.0299	0.0372	0.0329	0.0221	0.0	0.0021	0.0063	0.0	0.1984	0.0	0.0	0.0009	0.0	0.0016	0.0123	0.0002	0.046	0.0588	0.0007	0.3589	0.0142	0.1625	0.0	0.0047	0.08	0.0741	0.001	0.0125	0.1326	0.0592	0.0596	0.0	-1.0	0.0034	-1.0	0.0027	0.0008	0.0976	0.0022	0.0022	0.0	0.0	0.008	0.001	0.0009	0.0435	0.0369	0.0	0.0006	0.0	0.0	0.0	0.0	0.1633	0.0451	0.0026	0.0013	0.0053	0.0	0.0	0.0	0.0001	0.0	0.0049	0.0085	0.0	0.0	0.0	0.006	0.0009	0.0027	0.0	0.1978	0.0	-1.0	0.0101	0.0	0.0388	0.0096	0.108	0.2457	0.0001	0.0002	0.0	0.0015	0.1037	0.1704	0.1887	0.6857	0.0	0.0333	0.0545	0.0	0.3438	0.0	0.0	0.275	0.0	0.5	0.2408	0.0344	0.3444	0.224	0.1143	0.7917	0.25	0.3451	0.0	0.3692	0.1474	0.2458	0.1412	0.3455	0.4625	0.2686	0.45	0.0	-1.0	0.5111	-1.0	0.0643	0.6333	0.9556	0.136	0.6	0.0	0.0	0.1611	0.0333	0.1222	0.0	0.1437	0.0	0.0	0.0	0.0	0.4451	0.16	0.205	0.1286	0.0905	0.0	0.0	0.0	0.05	0.0	0.0875	0.0	0.0	0.0	0.225	0.225	0.3818	0.0	0.5	0.0	-1.0	0.2529	0.0	0.1083	0.1111	0.569	0.5615	0.0353	0.0833	0.0	0.0909	0.2753	0.1135	0.0232

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.0+cu126
• Datasets 4.4.2
• Tokenizers 0.22.2