10-classifier-finetuned-padchest

This model is a fine-tuned version of microsoft/resnet-50 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2197
• F1: 0.9062

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-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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	F1
0.7175	1.0	18	0.6978	0.5508
0.6996	2.0	36	0.6699	0.7605
0.659	3.0	54	0.6291	0.8044
0.5937	4.0	72	0.5778	0.7920
0.5124	5.0	90	0.5113	0.7934
0.4668	6.0	108	0.4066	0.7934
0.4079	7.0	126	0.4105	0.7934
0.363	8.0	144	0.3652	0.7934
0.337	9.0	162	0.3410	0.7934
0.3172	10.0	180	0.3272	0.7934
0.3082	11.0	198	0.2930	0.7934
0.2967	12.0	216	0.2814	0.7934
0.2889	13.0	234	0.2665	0.7934
0.2636	14.0	252	0.2846	0.7934
0.2694	15.0	270	0.2610	0.7934
0.2663	16.0	288	0.2828	0.7934
0.2573	17.0	306	0.2615	0.7934
0.2558	18.0	324	0.2606	0.7934
0.2492	19.0	342	0.2532	0.7934
0.2513	20.0	360	0.2559	0.7934
0.2429	21.0	378	0.2497	0.7934
0.2361	22.0	396	0.2412	0.7934
0.2423	23.0	414	0.2494	0.8235
0.2479	24.0	432	0.2446	0.8290
0.2237	25.0	450	0.2425	0.8428
0.2282	26.0	468	0.2446	0.8573
0.2343	27.0	486	0.2348	0.8344
0.2169	28.0	504	0.2358	0.8547
0.2169	29.0	522	0.2400	0.8622
0.2341	30.0	540	0.2342	0.8579
0.2241	31.0	558	0.2266	0.8511
0.2132	32.0	576	0.2250	0.8662
0.2155	33.0	594	0.2222	0.8485
0.2014	34.0	612	0.2279	0.8659
0.2033	35.0	630	0.2296	0.8886
0.1993	36.0	648	0.2252	0.8909
0.228	37.0	666	0.2226	0.8742
0.2292	38.0	684	0.2274	0.9030
0.202	39.0	702	0.2307	0.8997
0.2133	40.0	720	0.2244	0.8977
0.214	41.0	738	0.2281	0.9053
0.2203	42.0	756	0.2251	0.9020
0.2071	43.0	774	0.2214	0.8848
0.2125	44.0	792	0.2196	0.8932
0.2137	45.0	810	0.2187	0.8811
0.2073	46.0	828	0.2183	0.9020
0.2119	47.0	846	0.2185	0.9109
0.2018	48.0	864	0.2199	0.8943
0.1971	49.0	882	0.2211	0.9053
0.2079	50.0	900	0.2197	0.9062

Framework versions

• Transformers 4.28.0.dev0
• Pytorch 2.0.0+cu117
• Datasets 2.18.0
• Tokenizers 0.13.3