samolet_encoder_finetuned

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

• Loss: 0.1165

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 1337
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.3356	1.0	87	0.3257
0.2405	2.0	174	0.2403
0.1623	3.0	261	0.1606
0.1452	4.0	348	0.1454
0.1373	5.0	435	0.1413
0.1304	6.0	522	0.1369
0.129	7.0	609	0.1346
0.1291	8.0	696	0.1299
0.1277	9.0	783	0.1294
0.1244	10.0	870	0.1284
0.1275	11.0	957	0.1285
0.1196	12.0	1044	0.1264
0.1219	13.0	1131	0.1263
0.1195	14.0	1218	0.1265
0.1231	15.0	1305	0.1239
0.1208	16.0	1392	0.1216
0.118	17.0	1479	0.1223
0.1143	18.0	1566	0.1201
0.1177	19.0	1653	0.1198
0.1139	20.0	1740	0.1194
0.1152	21.0	1827	0.1193
0.1162	22.0	1914	0.1199
0.1113	23.0	2001	0.1183
0.1134	24.0	2088	0.1183
0.1136	25.0	2175	0.1184
0.1132	26.0	2262	0.1196
0.1156	27.0	2349	0.1185
0.1153	28.0	2436	0.1166
0.1139	29.0	2523	0.1153
0.1103	30.0	2610	0.1164

Framework versions

• Transformers 4.30.0.dev0
• Pytorch 2.0.1
• Datasets 2.12.0
• Tokenizers 0.13.3