vit-base-codenames

freddie2004/vit-base-codenames

vit-base-codenames is a image classification model from freddie2004. Use it when you need a label for 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

13% of all-time downloads

All-time downloads

Public

Parameters

85.8M

17.5 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors343 MB · 100%

At a glance

Task: Image Classification
Library: transformers
License: apache-2.0
Model type: vit
Access: Public
Created: Apr 11, 2024
Updated: Apr 11, 2024
SHA: 2014e0d6

Base models

google/vit-base-patch16-224-in21k

Datasets

imagefolder

Model card

From the Hugging Face model README

vit-base-codenames

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the codenames-pictures dataset. It achieves the following results on the evaluation set:

Loss: 0.7722
Accuracy: 0.4643

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: 0.0002
train_batch_size: 16
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.5428	0.16	100	2.4585	0.4106
1.498	0.31	200	1.5259	0.4449
1.2131	0.47	300	1.1431	0.4652
1.0505	0.63	400	1.0880	0.4485
0.9738	0.78	500	1.0141	0.4599
1.0137	0.94	600	0.9250	0.4670
0.932	1.1	700	0.9285	0.4731
0.9456	1.25	800	0.8803	0.4678
0.7922	1.41	900	0.8776	0.5084
0.8335	1.56	1000	0.8627	0.4811
0.7652	1.72	1100	0.8779	0.4952
0.8312	1.88	1200	0.8086	0.4872
0.7136	2.03	1300	0.8158	0.4590
0.7662	2.19	1400	0.8138	0.4661
0.7175	2.35	1500	0.8079	0.4775
0.7614	2.5	1600	0.8031	0.4802
0.7665	2.66	1700	0.8083	0.4467
0.7557	2.82	1800	0.8003	0.4819
0.7678	2.97	1900	0.8159	0.5013
0.697	3.13	2000	0.7845	0.4855
0.7211	3.29	2100	0.7942	0.4670
0.7307	3.44	2200	0.7908	0.4740
0.7122	3.6	2300	0.7920	0.4502
0.7597	3.76	2400	0.7722	0.4643
0.7165	3.91	2500	0.7784	0.4819
0.7238	4.07	2600	0.7896	0.4687
0.7047	4.23	2700	0.7849	0.4617
0.7122	4.38	2800	0.7963	0.4881
0.7574	4.54	2900	0.8179	0.4934
0.7987	4.69	3000	0.8060	0.4344
0.7517	4.85	3100	0.8018	0.4537
0.7402	5.01	3200	0.8076	0.4784
0.7223	5.16	3300	0.8026	0.4405
0.7417	5.32	3400	0.8005	0.4185
0.7236	5.48	3500	0.7964	0.4238
0.7114	5.63	3600	0.7989	0.3991
0.7237	5.79	3700	0.8229	0.4070
0.7203	5.95	3800	0.7989	0.4740
0.7081	6.1	3900	0.8194	0.4211
0.6839	6.26	4000	0.8013	0.4300
0.6832	6.42	4100	0.8065	0.3789
0.7093	6.57	4200	0.8014	0.3930
0.7037	6.73	4300	0.8170	0.4185
0.6999	6.89	4400	0.8318	0.4123
0.7166	7.04	4500	0.8254	0.4256
0.6838	7.2	4600	0.8204	0.3903
0.7135	7.36	4700	0.8385	0.3612
0.6999	7.51	4800	0.8406	0.3568
0.7007	7.67	4900	0.8353	0.3674
0.6892	7.82	5000	0.8318	0.3339
0.7036	7.98	5100	0.8310	0.3198

Framework versions

Transformers 4.37.2
Pytorch 2.2.1
Datasets 2.12.0
Tokenizers 0.15.1

Task: Image Classification
Library: transformers
Type: vit
License: apache-2.0
Created: Apr 11, 2024
Updated: Apr 11, 2024