vit-msn-small-lateral_flow_ivalidation_green_channel

This model is a fine-tuned version of facebook/vit-msn-small on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.5924
• Accuracy: 0.7454

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-07
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: reduce_lr_on_plateau
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 100
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.9312	0.9231	6	0.9478	0.2788
0.7936	2.0	13	0.8935	0.2491
0.718	2.9231	19	0.8518	0.2249
0.6787	4.0	26	0.8097	0.2119
0.6538	4.9231	32	0.7776	0.2770
0.6265	6.0	39	0.7428	0.3662
0.5812	6.9231	45	0.7191	0.4424
0.6138	8.0	52	0.6996	0.4963
0.57	8.9231	58	0.6900	0.5409
0.5595	10.0	65	0.6804	0.5762
0.5288	10.9231	71	0.6727	0.6004
0.5094	12.0	78	0.6613	0.6320
0.5073	12.9231	84	0.6490	0.6636
0.4504	14.0	91	0.6386	0.6970
0.4868	14.9231	97	0.6305	0.7156
0.4799	16.0	104	0.6264	0.7175
0.4861	16.9231	110	0.6235	0.7175
0.4975	18.0	117	0.6163	0.7286
0.4712	18.9231	123	0.6127	0.7398
0.468	20.0	130	0.6107	0.7416
0.4562	20.9231	136	0.6070	0.7454
0.5195	22.0	143	0.6056	0.7454
0.4385	22.9231	149	0.6033	0.7416
0.4211	24.0	156	0.6050	0.7342
0.4364	24.9231	162	0.6023	0.7361
0.4327	26.0	169	0.5980	0.7416
0.4757	26.9231	175	0.6000	0.7361
0.4287	28.0	182	0.5924	0.7454
0.4313	28.9231	188	0.5970	0.7361
0.4483	30.0	195	0.5962	0.7398
0.3956	30.9231	201	0.5976	0.7305
0.41	32.0	208	0.6060	0.7212
0.4371	32.9231	214	0.6050	0.7193
0.4169	34.0	221	0.6045	0.7212
0.3882	34.9231	227	0.6020	0.7230
0.5097	36.0	234	0.6011	0.7286
0.476	36.9231	240	0.6027	0.7268
0.387	38.0	247	0.6012	0.7249
0.4744	38.9231	253	0.6017	0.7230
0.4712	40.0	260	0.6025	0.7230
0.4242	40.9231	266	0.6022	0.7230
0.4087	42.0	273	0.6021	0.7230
0.4009	42.9231	279	0.6026	0.7230
0.4219	44.0	286	0.6026	0.7230
0.4208	44.9231	292	0.6024	0.7230
0.3644	46.0	299	0.6013	0.7230
0.4458	46.9231	305	0.5997	0.7286
0.425	48.0	312	0.5991	0.7286
0.3982	48.9231	318	0.5995	0.7286
0.4167	50.0	325	0.5992	0.7286
0.4112	50.9231	331	0.5992	0.7286
0.4073	52.0	338	0.5992	0.7286
0.4413	52.9231	344	0.5991	0.7286
0.4326	54.0	351	0.5991	0.7286
0.4206	54.9231	357	0.5992	0.7286
0.3776	56.0	364	0.5993	0.7286
0.3792	56.9231	370	0.5994	0.7286
0.4075	58.0	377	0.5995	0.7286
0.4412	58.9231	383	0.5995	0.7286
0.4137	60.0	390	0.5995	0.7286
0.424	60.9231	396	0.5995	0.7286
0.3988	62.0	403	0.5997	0.7286
0.4167	62.9231	409	0.5996	0.7286
0.41	64.0	416	0.5997	0.7286
0.4235	64.9231	422	0.5997	0.7286
0.4544	66.0	429	0.5998	0.7286
0.4495	66.9231	435	0.5997	0.7286
0.424	68.0	442	0.5997	0.7286
0.4053	68.9231	448	0.5997	0.7286
0.426	70.0	455	0.5999	0.7286
0.3865	70.9231	461	0.6000	0.7286
0.3732	72.0	468	0.6001	0.7286
0.4289	72.9231	474	0.6002	0.7286
0.4524	74.0	481	0.6002	0.7286
0.4081	74.9231	487	0.6002	0.7286
0.384	76.0	494	0.6001	0.7286
0.4177	76.9231	500	0.6000	0.7286
0.3777	78.0	507	0.6000	0.7286
0.4226	78.9231	513	0.6000	0.7286
0.419	80.0	520	0.6000	0.7286
0.3956	80.9231	526	0.6000	0.7286
0.3669	82.0	533	0.6000	0.7286
0.3902	82.9231	539	0.6000	0.7286
0.4193	84.0	546	0.6001	0.7286
0.4115	84.9231	552	0.6001	0.7286
0.3923	86.0	559	0.6001	0.7286
0.4011	86.9231	565	0.6001	0.7286
0.4765	88.0	572	0.6000	0.7286
0.4034	88.9231	578	0.5999	0.7286
0.3867	90.0	585	0.5998	0.7286
0.4201	90.9231	591	0.5998	0.7286
0.4346	92.0	598	0.5998	0.7286
0.4171	92.3077	600	0.5998	0.7286

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.19.1