vit-base_rvl-cdip-tiny_rvl_cdip-NK1000_simkd

jordyvl/vit-base_rvl-cdip-tiny_rvl_cdip-NK1000_simkd

vit-base_rvl-cdip-tiny_rvl_cdip-NK1000_simkd is a image classification model from jordyvl. 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

2% of all-time downloads

All-time downloads

387

Public

Repo size

47.8 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin22.6 MB · 90%

At a glance

Task: Image Classification
Library: transformers
License: apache-2.0
Model type: vit
Access: Public
Created: Jul 27, 2023
Updated: Jul 29, 2023
SHA: db4e087c

Model card

From the Hugging Face model README

vit-base_rvl-cdip-tiny_rvl_cdip-NK1000_simkd

This model is a fine-tuned version of WinKawaks/vit-tiny-patch16-224 on the None dataset. It achieves the following results on the evaluation set:

Loss: 37.0129
Accuracy: 0.8277
Brier Loss: 0.3307
Nll: 1.8775
F1 Micro: 0.8277
F1 Macro: 0.8289
Ece: 0.1649
Aurc: 0.0944

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: 0.0001
train_batch_size: 64
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Brier Loss	Nll	F1 Micro	F1 Macro	Ece	Aurc
No log	1.0	250	56.2115	0.3142	0.8385	3.5992	0.3142	0.2499	0.1012	0.5692
56.615	2.0	500	54.0327	0.4025	0.9176	3.1629	0.4025	0.3116	0.4002	0.3781
56.615	3.0	750	49.9569	0.4728	0.8906	2.8997	0.4728	0.4076	0.4129	0.2864
50.7474	4.0	1000	47.4945	0.5685	0.7670	2.6755	0.5685	0.5350	0.3561	0.2844
50.7474	5.0	1250	45.5054	0.6378	0.6629	2.5408	0.6378	0.6030	0.3212	0.1851
45.4907	6.0	1500	43.9471	0.679	0.5949	2.6322	0.679	0.6636	0.2925	0.1474
45.4907	7.0	1750	42.9273	0.7342	0.4843	2.4382	0.7342	0.7365	0.2245	0.1436
42.5191	8.0	2000	41.9715	0.7548	0.4560	2.3596	0.7548	0.7533	0.2231	0.1400
42.5191	9.0	2250	41.4349	0.7722	0.4310	2.3144	0.7722	0.7718	0.2103	0.1304
40.8849	10.0	2500	41.0961	0.7805	0.4187	2.2268	0.7805	0.7826	0.2047	0.1305
40.8849	11.0	2750	40.5831	0.7893	0.4030	2.1663	0.7893	0.7930	0.2001	0.1246
39.8394	12.0	3000	40.1596	0.7987	0.3877	2.1719	0.7987	0.8015	0.1929	0.1162
39.8394	13.0	3250	39.8469	0.8033	0.3821	2.1455	0.8033	0.8077	0.1889	0.1183
38.9442	14.0	3500	39.5865	0.8055	0.3761	2.1121	0.8055	0.8096	0.1864	0.1110
38.9442	15.0	3750	39.4686	0.81	0.3693	2.0948	0.81	0.8125	0.1831	0.1114
38.3612	16.0	4000	39.1387	0.8207	0.3446	1.9957	0.8207	0.8219	0.1716	0.1038
38.3612	17.0	4250	38.8950	0.8143	0.3575	2.0339	0.8143	0.8152	0.1781	0.1034
37.7855	18.0	4500	38.6442	0.8215	0.3442	1.9658	0.8215	0.8236	0.1718	0.1036
37.7855	19.0	4750	38.5218	0.8197	0.3477	1.9627	0.8197	0.8220	0.1735	0.1070
37.3649	20.0	5000	38.3474	0.8225	0.3413	1.9886	0.8225	0.8239	0.1710	0.1028
37.3649	21.0	5250	38.2377	0.8257	0.3358	1.9864	0.8257	0.8269	0.1674	0.0957
37.0326	22.0	5500	38.1089	0.824	0.3418	1.9404	0.824	0.8257	0.1678	0.0980
37.0326	23.0	5750	37.9861	0.8273	0.3339	1.9540	0.8273	0.8285	0.1664	0.0985
36.7372	24.0	6000	37.8397	0.8255	0.3376	1.9492	0.8255	0.8268	0.1685	0.0944
36.7372	25.0	6250	37.7772	0.8253	0.3370	1.9078	0.8253	0.8255	0.1669	0.0997
36.4341	26.0	6500	37.6550	0.828	0.3325	1.9388	0.828	0.8284	0.1647	0.0943
36.4341	27.0	6750	37.5873	0.8255	0.3364	1.9319	0.8255	0.8261	0.1680	0.0920
36.2152	28.0	7000	37.5052	0.825	0.3379	1.8945	0.825	0.8268	0.1681	0.0981
36.2152	29.0	7250	37.4586	0.8243	0.3361	1.9094	0.8243	0.8251	0.1692	0.0945
36.0128	30.0	7500	37.3730	0.8277	0.3304	1.9062	0.8277	0.8288	0.1657	0.0946
36.0128	31.0	7750	37.3309	0.8277	0.3309	1.9045	0.8277	0.8291	0.1660	0.0947
35.8486	32.0	8000	37.2620	0.8267	0.3323	1.8884	0.8267	0.8279	0.1652	0.0950
35.8486	33.0	8250	37.2147	0.8275	0.3308	1.9079	0.8275	0.8290	0.1654	0.0960
35.6854	34.0	8500	37.1911	0.831	0.3252	1.8935	0.831	0.8323	0.1613	0.0939
35.6854	35.0	8750	37.1523	0.8283	0.3301	1.8847	0.8283	0.8293	0.1644	0.0972
35.5758	36.0	9000	37.1315	0.8305	0.3252	1.8941	0.8305	0.8317	0.1627	0.0934
35.5758	37.0	9250	37.1184	0.8275	0.3320	1.8844	0.8275	0.8285	0.1654	0.0923
35.4911	38.0	9500	37.1149	0.827	0.3327	1.8885	0.827	0.8288	0.1668	0.0953
35.4911	39.0	9750	37.1067	0.8267	0.3323	1.8846	0.8267	0.8281	0.1659	0.0932
35.4248	40.0	10000	37.0792	0.8293	0.3294	1.8840	0.8293	0.8305	0.1633	0.0937
35.4248	41.0	10250	37.0798	0.8297	0.3288	1.8718	0.8297	0.8309	0.1639	0.0929
35.3648	42.0	10500	37.0635	0.8265	0.3351	1.8883	0.8265	0.8279	0.1680	0.0951
35.3648	43.0	10750	37.0470	0.828	0.3308	1.8746	0.828	0.8294	0.1656	0.0939
35.2961	44.0	11000	37.0305	0.8273	0.3321	1.8901	0.8273	0.8286	0.1657	0.0932
35.2961	45.0	11250	37.0261	0.8275	0.3315	1.8823	0.8275	0.8287	0.1650	0.0949
35.241	46.0	11500	37.0253	0.827	0.3311	1.8751	0.827	0.8283	0.1662	0.0940
35.241	47.0	11750	37.0200	0.8277	0.3321	1.8708	0.8277	0.8289	0.1653	0.0949
35.2059	48.0	12000	37.0165	0.8277	0.3305	1.8745	0.8277	0.8289	0.1650	0.0934
35.2059	49.0	12250	37.0130	0.8275	0.3312	1.8743	0.8275	0.8287	0.1655	0.0942
35.18	50.0	12500	37.0129	0.8277	0.3307	1.8775	0.8277	0.8289	0.1649	0.0944

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

Task: Image Classification
Library: transformers
Type: vit
License: apache-2.0
Created: Jul 27, 2023
Updated: Jul 29, 2023