daigram_detr_r50_albumentations

This model is a fine-tuned version of facebook/detr-resnet-50 on the bpmn-shapes dataset. It achieves the following results on the evaluation set:

• Loss: 1.0088

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 500

Training results

Training Loss	Epoch	Step	Validation Loss
3.8163	2.63	50	3.0660
2.9036	5.26	100	2.8878
2.7516	7.89	150	2.8043
2.6278	10.53	200	2.6820
2.4806	13.16	250	2.5676
2.3781	15.79	300	2.4282
2.253	18.42	350	2.3161
2.1405	21.05	400	2.1735
2.0263	23.68	450	2.0909
1.9732	26.32	500	2.0120
1.8647	28.95	550	1.9260
1.7793	31.58	600	1.8655
1.7706	34.21	650	1.8166
1.6792	36.84	700	1.7325
1.5654	39.47	750	1.7061
1.5802	42.11	800	1.6463
1.5053	44.74	850	1.5985
1.4858	47.37	900	1.6060
1.4186	50.0	950	1.5563
1.4391	52.63	1000	1.5219
1.3938	55.26	1050	1.4995
1.3734	57.89	1100	1.4661
1.3379	60.53	1150	1.4451
1.341	63.16	1200	1.4854
1.3647	65.79	1250	1.4509
1.3198	68.42	1300	1.4116
1.3054	71.05	1350	1.3821
1.2945	73.68	1400	1.3952
1.2899	76.32	1450	1.3868
1.2533	78.95	1500	1.3580
1.2655	81.58	1550	1.3374
1.2649	84.21	1600	1.3451
1.2286	86.84	1650	1.2973
1.2497	89.47	1700	1.3322
1.2456	92.11	1750	1.3289
1.2234	94.74	1800	1.3080
1.1695	97.37	1850	1.3218
1.2265	100.0	1900	1.3280
1.1899	102.63	1950	1.2834
1.1914	105.26	2000	1.2931
1.1698	107.89	2050	1.3176
1.177	110.53	2100	1.2896
1.1625	113.16	2150	1.2936
1.1626	115.79	2200	1.2614
1.1698	118.42	2250	1.2545
1.1703	121.05	2300	1.2398
1.1659	123.68	2350	1.2254
1.1734	126.32	2400	1.2489
1.1234	128.95	2450	1.2072
1.1464	131.58	2500	1.1707
1.1268	134.21	2550	1.1971
1.1511	136.84	2600	1.2247
1.1234	139.47	2650	1.1921
1.0923	142.11	2700	1.1751
1.1267	144.74	2750	1.1905
1.1021	147.37	2800	1.1885
1.1075	150.0	2850	1.1780
1.1116	152.63	2900	1.1666
1.0987	155.26	2950	1.1694
1.0974	157.89	3000	1.1931
1.0867	160.53	3050	1.1461
1.1076	163.16	3100	1.1501
1.0912	165.79	3150	1.1611
1.0671	168.42	3200	1.1718
1.0981	171.05	3250	1.1961
1.0602	173.68	3300	1.1786
1.0305	176.32	3350	1.1640
1.0647	178.95	3400	1.1416
1.0628	181.58	3450	1.1296
1.0856	184.21	3500	1.1140
1.0626	186.84	3550	1.1214
1.0782	189.47	3600	1.1449
1.0601	192.11	3650	1.1441
1.0906	194.74	3700	1.1396
1.0376	197.37	3750	1.1271
1.0625	200.0	3800	1.1397
1.057	202.63	3850	1.1121
1.0448	205.26	3900	1.1376
1.0747	207.89	3950	1.1475
1.0605	210.53	4000	1.0916
1.0344	213.16	4050	1.1001
1.0443	215.79	4100	1.0976
1.0202	218.42	4150	1.1240
1.078	221.05	4200	1.1024
1.0251	223.68	4250	1.0793
1.0353	226.32	4300	1.1153
1.0047	228.95	4350	1.0972
1.0143	231.58	4400	1.0948
1.0172	234.21	4450	1.1265
1.0299	236.84	4500	1.1038
0.9968	239.47	4550	1.0901
1.0233	242.11	4600	1.0945
0.9943	244.74	4650	1.0918
1.0321	247.37	4700	1.1270
1.0113	250.0	4750	1.1060
1.0229	252.63	4800	1.0859
0.9945	255.26	4850	1.0875
1.0073	257.89	4900	1.0976
1.0096	260.53	4950	1.0933
1.0	263.16	5000	1.0821
1.0326	265.79	5050	1.0747
0.997	268.42	5100	1.0931
1.0056	271.05	5150	1.0853
0.9858	273.68	5200	1.0945
1.0005	276.32	5250	1.0669
1.0217	278.95	5300	1.0497
0.9777	281.58	5350	1.0672
0.9888	284.21	5400	1.0844
0.9662	286.84	5450	1.0524
1.0029	289.47	5500	1.0519
0.984	292.11	5550	1.0538
0.9724	294.74	5600	1.0524
0.991	297.37	5650	1.0553
0.9936	300.0	5700	1.0601
0.9817	302.63	5750	1.0524
0.9868	305.26	5800	1.0644
0.9982	307.89	5850	1.0523
0.9814	310.53	5900	1.0611
0.9761	313.16	5950	1.0505
0.9507	315.79	6000	1.0361
0.9786	318.42	6050	1.0275
0.9684	321.05	6100	1.0292
0.9759	323.68	6150	1.0529
0.9442	326.32	6200	1.0689
0.9653	328.95	6250	1.0696
0.9579	331.58	6300	1.0572
1.0016	334.21	6350	1.0660
0.9462	336.84	6400	1.0525
0.9596	339.47	6450	1.0505
0.9655	342.11	6500	1.0514
0.9713	344.74	6550	1.0616
0.952	347.37	6600	1.0497
0.9433	350.0	6650	1.0389
0.9619	352.63	6700	1.0404
0.9594	355.26	6750	1.0332
0.9586	357.89	6800	1.0323
0.9582	360.53	6850	1.0294
0.9437	363.16	6900	1.0329
0.9585	365.79	6950	1.0361
0.9661	368.42	7000	1.0428
0.9603	371.05	7050	1.0299
0.9619	373.68	7100	1.0416
0.9766	376.32	7150	1.0471
0.9547	378.95	7200	1.0498
0.967	381.58	7250	1.0318
0.9463	384.21	7300	1.0238
0.9531	386.84	7350	1.0329
0.9342	389.47	7400	1.0354
0.939	392.11	7450	1.0312
0.9635	394.74	7500	1.0325
0.9261	397.37	7550	1.0245
0.962	400.0	7600	1.0381
0.9385	402.63	7650	1.0243
0.9422	405.26	7700	1.0235
0.9285	407.89	7750	1.0286
0.9598	410.53	7800	1.0353
0.9529	413.16	7850	1.0361
0.928	415.79	7900	1.0316
0.935	418.42	7950	1.0263
0.9456	421.05	8000	1.0368
0.9387	423.68	8050	1.0440
0.9321	426.32	8100	1.0440
0.9236	428.95	8150	1.0394
0.9448	431.58	8200	1.0467
0.9151	434.21	8250	1.0516
0.9373	436.84	8300	1.0383
0.9577	439.47	8350	1.0190
0.9199	442.11	8400	1.0215
0.9321	444.74	8450	1.0184
0.9387	447.37	8500	1.0236
0.9382	450.0	8550	1.0259
0.9391	452.63	8600	1.0282
0.9392	455.26	8650	1.0193
0.9438	457.89	8700	1.0124
0.9398	460.53	8750	1.0060
0.9246	463.16	8800	1.0140
0.9383	465.79	8850	1.0145
0.9267	468.42	8900	1.0122
0.9253	471.05	8950	1.0144
0.9238	473.68	9000	1.0065
0.9082	476.32	9050	1.0136
0.9287	478.95	9100	1.0120
0.9161	481.58	9150	1.0120
0.9093	484.21	9200	1.0128
0.9264	486.84	9250	1.0125
0.9487	489.47	9300	1.0131
0.9398	492.11	9350	1.0101
0.9039	494.74	9400	1.0090
0.908	497.37	9450	1.0097
0.944	500.0	9500	1.0088

Framework versions

• Transformers 4.34.1
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.14.1