efficientnet-b0-accidents

This model is a fine-tuned version of google/efficientnet-b0 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.3330
• Accuracy: 0.8367

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: 256
• eval_batch_size: 256
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 1024
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	1	0.4780	0.7959
No log	2.0	2	0.4610	0.7959
No log	3.0	3	0.4518	0.7755
No log	4.0	4	0.4450	0.8163
No log	5.0	5	0.4424	0.8265
No log	6.0	6	0.4483	0.7959
No log	7.0	7	0.4533	0.7959
No log	8.0	8	0.4557	0.7959
No log	9.0	9	0.4556	0.8265
0.4528	10.0	10	0.4453	0.8163
0.4528	11.0	11	0.4558	0.7755
0.4528	12.0	12	0.4390	0.8265
0.4528	13.0	13	0.4322	0.7959
0.4528	14.0	14	0.4323	0.8163
0.4528	15.0	15	0.4127	0.8061
0.4528	16.0	16	0.4341	0.8061
0.4528	17.0	17	0.4144	0.8265
0.4528	18.0	18	0.4275	0.8265
0.4528	19.0	19	0.3988	0.8673
0.4233	20.0	20	0.4210	0.7959
0.4233	21.0	21	0.4223	0.7755
0.4233	22.0	22	0.4288	0.8265
0.4233	23.0	23	0.3851	0.8571
0.4233	24.0	24	0.3956	0.8061
0.4233	25.0	25	0.4159	0.8367
0.4233	26.0	26	0.4055	0.8163
0.4233	27.0	27	0.3861	0.8163
0.4233	28.0	28	0.3751	0.8469
0.4233	29.0	29	0.3915	0.8367
0.3846	30.0	30	0.3705	0.8571
0.3846	31.0	31	0.3868	0.8367
0.3846	32.0	32	0.3710	0.8469
0.3846	33.0	33	0.3770	0.8469
0.3846	34.0	34	0.3903	0.8265
0.3846	35.0	35	0.3864	0.8469
0.3846	36.0	36	0.3728	0.8265
0.3846	37.0	37	0.3772	0.8367
0.3846	38.0	38	0.3633	0.8163
0.3846	39.0	39	0.3824	0.8469
0.3714	40.0	40	0.3520	0.8571
0.3714	41.0	41	0.3844	0.8469
0.3714	42.0	42	0.3564	0.8469
0.3714	43.0	43	0.3747	0.8673
0.3714	44.0	44	0.3395	0.8571
0.3714	45.0	45	0.3871	0.8163
0.3714	46.0	46	0.3487	0.8367
0.3714	47.0	47	0.3798	0.8163
0.3714	48.0	48	0.3848	0.8367
0.3714	49.0	49	0.3978	0.8265
0.3618	50.0	50	0.3384	0.8571
0.3618	51.0	51	0.3647	0.8265
0.3618	52.0	52	0.3544	0.8571
0.3618	53.0	53	0.4289	0.8163
0.3618	54.0	54	0.3568	0.8673
0.3618	55.0	55	0.3727	0.8673
0.3618	56.0	56	0.3796	0.8265
0.3618	57.0	57	0.3678	0.8571
0.3618	58.0	58	0.3719	0.8469
0.3618	59.0	59	0.3808	0.8878
0.327	60.0	60	0.3783	0.8163
0.327	61.0	61	0.3637	0.8367
0.327	62.0	62	0.3743	0.8367
0.327	63.0	63	0.3554	0.8571
0.327	64.0	64	0.3544	0.8265
0.327	65.0	65	0.3615	0.8469
0.327	66.0	66	0.3503	0.8673
0.327	67.0	67	0.3914	0.7959
0.327	68.0	68	0.3687	0.8367
0.327	69.0	69	0.3296	0.8878
0.3136	70.0	70	0.3548	0.8571
0.3136	71.0	71	0.3810	0.8265
0.3136	72.0	72	0.3522	0.8469
0.3136	73.0	73	0.3852	0.8367
0.3136	74.0	74	0.3434	0.8571
0.3136	75.0	75	0.3596	0.8571
0.3136	76.0	76	0.3551	0.8367
0.3136	77.0	77	0.4257	0.8163
0.3136	78.0	78	0.3554	0.8367
0.3136	79.0	79	0.3352	0.8265
0.316	80.0	80	0.3773	0.8367
0.316	81.0	81	0.3305	0.8469
0.316	82.0	82	0.3614	0.8571
0.316	83.0	83	0.3491	0.8265
0.316	84.0	84	0.3479	0.8571
0.316	85.0	85	0.3684	0.8367
0.316	86.0	86	0.3511	0.8571
0.316	87.0	87	0.3658	0.8265
0.316	88.0	88	0.3333	0.8367
0.316	89.0	89	0.3584	0.8776
0.3089	90.0	90	0.3277	0.8571
0.3089	91.0	91	0.3875	0.8367
0.3089	92.0	92	0.3757	0.8367
0.3089	93.0	93	0.3488	0.8367
0.3089	94.0	94	0.3282	0.8571
0.3089	95.0	95	0.3613	0.8571
0.3089	96.0	96	0.3753	0.8469
0.3089	97.0	97	0.3625	0.8469
0.3089	98.0	98	0.3930	0.8265
0.3089	99.0	99	0.3338	0.8469
0.3131	100.0	100	0.3330	0.8367

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1