segformer-v1-breastcancer

This model is a fine-tuned version of nvidia/segformer-b0-finetuned-cityscapes-1024-1024 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2084
• Mean Iou: 0.6074
• Mean Accuracy: 0.7133
• Overall Accuracy: 0.6718
• Per Category Iou: [0.6503515075769412, 0.5644565972298056]
• Per Category Accuracy: [0.7843872475128127, 0.6421245639664888]

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: 6e-05
• train_batch_size: 3
• eval_batch_size: 3
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Per Category Iou	Per Category Accuracy
1.0349	1.82	20	0.9385	0.1001	0.3453	0.5410	[0.00702490904002239, 0.19315512632820392]	[0.00945884835694905, 0.6811663337407948]
0.8631	3.64	40	0.8712	0.1270	0.3748	0.5931	[9.482867619168036e-05, 0.25396146547703435]	[0.00011305396442568585, 0.7494807350208202]
0.6657	5.45	60	0.6510	0.1313	0.2115	0.3347	[0.00014806040864672785, 0.26239433754030744]	[0.00015073861923424781, 0.42294505232402135]
0.6924	7.27	80	0.5721	0.1917	0.3061	0.4833	[0.002107933665379521, 0.38125252274350296]	[0.0021291829966837506, 0.6101487731433172]
0.5177	9.09	100	0.4836	0.1991	0.3081	0.4876	[0.0, 0.3981538560328774]	[0.0, 0.6162060363932699]
0.3851	10.91	120	0.4029	0.2127	0.2893	0.4440	[0.023690796530116853, 0.40166184061437743]	[0.02372249020198975, 0.5548632022499826]
0.3266	12.73	140	0.3811	0.2300	0.3350	0.5130	[0.028268806709322924, 0.43178856750464767]	[0.02946940006029545, 0.6405295012074774]
0.3397	14.55	160	0.3353	0.2616	0.3719	0.5640	[0.04190433583118965, 0.4812188969936601]	[0.042357552004823634, 0.7015294713932202]
0.3008	16.36	180	0.3363	0.3885	0.4376	0.4135	[0.4227194892852987, 0.35420100310527275]	[0.47910385890865237, 0.3961867565069616]
0.2558	18.18	200	0.3163	0.4200	0.4832	0.4322	[0.48242302607476784, 0.35761699452079404]	[0.570677570093458, 0.3956699760492134]
0.2686	20.0	220	0.2771	0.4777	0.5444	0.5868	[0.4603203796001692, 0.49515000498355427]	[0.4716046126017486, 0.6171352474086441]
0.1953	21.82	240	0.2811	0.4756	0.5676	0.5920	[0.46844517569632155, 0.4827354154204578]	[0.5257386192342478, 0.6095276427854467]
0.1623	23.64	260	0.2612	0.4833	0.5416	0.5447	[0.506478482184174, 0.46020570281796136]	[0.5361961109436237, 0.5469524860121443]
0.1851	25.45	280	0.2620	0.5107	0.5880	0.5313	[0.5881106780729983, 0.4333538137452822]	[0.6852389207114863, 0.49066316846049113]
0.1315	27.27	300	0.2230	0.6652	0.7361	0.6967	[0.7577948727059535, 0.5726185409040606]	[0.8037006331022007, 0.6684903053973743]
0.1294	29.09	320	0.2330	0.5189	0.6179	0.6328	[0.506419446816051, 0.5313992809888866]	[0.5923462466083811, 0.6434165151108594]
0.1532	30.91	340	0.2326	0.5319	0.6251	0.6503	[0.5461152173144251, 0.5176845532961513]	[0.581945281881218, 0.6683163888971706]
0.1074	32.73	360	0.2280	0.5790	0.6418	0.5960	[0.6624514966740577, 0.4955288623414331]	[0.7205682845945132, 0.5631018753167765]
0.1184	34.55	380	0.2168	0.6385	0.7453	0.7145	[0.7140882114917724, 0.5629577265658137]	[0.7980479348809165, 0.6925007205112151]
0.1411	36.36	400	0.2191	0.5935	0.6776	0.6459	[0.6633485862587079, 0.5236754959973609]	[0.7320432619837203, 0.6231328821442413]
0.1224	38.18	420	0.2068	0.6114	0.6869	0.6689	[0.6632029659025639, 0.5596692813228747]	[0.717949201085318, 0.6559037198254872]
0.0892	40.0	440	0.2096	0.5867	0.6817	0.6756	[0.6250170137471076, 0.548339821945447]	[0.692191739523666, 0.6711785575862378]
0.103	41.82	460	0.2117	0.5693	0.6553	0.6511	[0.6029494984137872, 0.5356447598629901]	[0.6625150738619234, 0.6480725082734564]
0.0996	43.64	480	0.2082	0.6011	0.7024	0.6743	[0.6408627400521119, 0.5614076241331366]	[0.7507725354235755, 0.6540800810947796]
0.1095	45.45	500	0.2065	0.6254	0.7302	0.6836	[0.6779631615467104, 0.5728211009174312]	[0.8100504974374435, 0.6502936704332012]
0.097	47.27	520	0.2083	0.6079	0.7042	0.6628	[0.6564823383005202, 0.5592888498683055]	[0.7753052457039493, 0.6330858749987578]
0.0866	49.09	540	0.2084	0.6074	0.7133	0.6718	[0.6503515075769412, 0.5644565972298056]	[0.7843872475128127, 0.6421245639664888]

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2