fb-detr-aug-table_detection_v1.0

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

• Loss: 0.3284

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
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 300

Training results

Training Loss	Epoch	Step	Validation Loss
2.1127	1.21	20	1.6338
1.8818	2.42	40	1.0008
1.6752	3.64	60	1.4500
1.516	4.85	80	0.9846
1.328	6.06	100	1.0746
1.2713	7.27	120	1.1575
1.1762	8.48	140	0.7001
1.1547	9.7	160	1.0982
1.2178	10.91	180	1.2437
1.0999	12.12	200	0.9853
1.1452	13.33	220	0.8249
1.0528	14.55	240	0.7035
1.0157	15.76	260	0.7584
0.9898	16.97	280	0.7169
0.9011	18.18	300	0.9833
0.9248	19.39	320	0.5799
0.9295	20.61	340	0.7567
0.8687	21.82	360	0.8273
0.9934	23.03	380	0.7053
0.9039	24.24	400	0.7121
0.9244	25.45	420	0.7668
0.8525	26.67	440	0.8034
0.8996	27.88	460	0.7558
0.9486	29.09	480	0.6570
0.9838	30.3	500	0.6775
1.0131	31.52	520	0.6643
0.911	32.73	540	0.6673
0.9749	33.94	560	0.7285
0.9277	35.15	580	0.5660
0.885	36.36	600	0.6928
0.8128	37.58	620	0.6517
0.8082	38.79	640	0.6254
0.8702	40.0	660	0.7354
0.8563	41.21	680	0.6653
0.8147	42.42	700	0.7279
0.7741	43.64	720	0.8649
0.7128	44.85	740	0.6545
0.7806	46.06	760	0.6264
0.7497	47.27	780	0.6577
0.687	48.48	800	0.6218
0.761	49.7	820	0.8314
0.7987	50.91	840	0.6444
0.7357	52.12	860	0.6575
0.7023	53.33	880	0.5817
0.6802	54.55	900	0.6244
0.7285	55.76	920	0.5916
0.6959	56.97	940	0.5081
0.6638	58.18	960	0.5037
0.6957	59.39	980	0.5085
0.6571	60.61	1000	0.4837
0.6837	61.82	1020	0.6387
0.7012	63.03	1040	0.4773
0.7139	64.24	1060	0.5028
0.7234	65.45	1080	0.5678
0.7228	66.67	1100	0.6430
0.6973	67.88	1120	0.6091
0.7096	69.09	1140	0.4702
0.6688	70.3	1160	0.5281
0.6378	71.52	1180	0.5869
0.6533	72.73	1200	0.5513
0.5966	73.94	1220	0.5030
0.6459	75.15	1240	0.5056
0.6496	76.36	1260	0.5982
0.7562	77.58	1280	0.4316
0.6744	78.79	1300	0.5127
0.725	80.0	1320	0.4750
0.6317	81.21	1340	0.5916
0.6138	82.42	1360	0.5602
0.5979	83.64	1380	0.5578
0.6455	84.85	1400	0.5035
0.6428	86.06	1420	0.4647
0.6101	87.27	1440	0.5262
0.6003	88.48	1460	0.4931
0.6019	89.7	1480	0.4655
0.609	90.91	1500	0.5081
0.6059	92.12	1520	0.4959
0.5952	93.33	1540	0.4069
0.6115	94.55	1560	0.5783
0.6277	95.76	1580	0.5889
0.6392	96.97	1600	0.5349
0.6003	98.18	1620	0.4729
0.6195	99.39	1640	0.4943
0.6209	100.61	1660	0.5134
0.6042	101.82	1680	0.5111
0.5964	103.03	1700	0.4301
0.5716	104.24	1720	0.4129
0.5466	105.45	1740	0.5458
0.5679	106.67	1760	0.5224
0.5754	107.88	1780	0.4612
0.543	109.09	1800	0.4411
0.5434	110.3	1820	0.3614
0.5682	111.52	1840	0.4925
0.6027	112.73	1860	0.4388
0.5683	113.94	1880	0.4456
0.5566	115.15	1900	0.4899
0.5738	116.36	1920	0.4500
0.5494	117.58	1940	0.4949
0.5848	118.79	1960	0.4078
0.6483	120.0	1980	0.4234
0.5738	121.21	2000	0.6240
0.5656	122.42	2020	0.6076
0.52	123.64	2040	0.4267
0.5692	124.85	2060	0.4629
0.5728	126.06	2080	0.4723
0.6444	127.27	2100	0.4098
0.565	128.48	2120	0.4331
0.5484	129.7	2140	0.4324
0.5164	130.91	2160	0.4289
0.5354	132.12	2180	0.3927
0.5332	133.33	2200	0.3951
0.4956	134.55	2220	0.4877
0.5107	135.76	2240	0.5421
0.5192	136.97	2260	0.4340
0.4702	138.18	2280	0.5052
0.4863	139.39	2300	0.4147
0.4977	140.61	2320	0.4434
0.5222	141.82	2340	0.4550
0.5292	143.03	2360	0.4839
0.5376	144.24	2380	0.3728
0.4915	145.45	2400	0.4733
0.4641	146.67	2420	0.3470
0.5144	147.88	2440	0.3606
0.4891	149.09	2460	0.4212
0.4758	150.3	2480	0.6014
0.4901	151.52	2500	0.3525
0.4809	152.73	2520	0.4205
0.486	153.94	2540	0.3663
0.4943	155.15	2560	0.5401
0.4857	156.36	2580	0.4914
0.4898	157.58	2600	0.4820
0.4783	158.79	2620	0.4178
0.4941	160.0	2640	0.4133
0.4607	161.21	2660	0.3855
0.4797	162.42	2680	0.3911
0.4874	163.64	2700	0.3821
0.4799	164.85	2720	0.4532
0.4683	166.06	2740	0.4442
0.4843	167.27	2760	0.3532
0.4781	168.48	2780	0.5200
0.4561	169.7	2800	0.4211
0.4745	170.91	2820	0.4610
0.4872	172.12	2840	0.3453
0.4299	173.33	2860	0.4454
0.4609	174.55	2880	0.3775
0.4318	175.76	2900	0.4044
0.4429	176.97	2920	0.5326
0.4521	178.18	2940	0.3521
0.46	179.39	2960	0.4162
0.4858	180.61	2980	0.4760
0.4483	181.82	3000	0.3208
0.4553	183.03	3020	0.3736
0.4497	184.24	3040	0.3852
0.4487	185.45	3060	0.4270
0.4646	186.67	3080	0.4376
0.4538	187.88	3100	0.4299
0.4915	189.09	3120	0.2842
0.4194	190.3	3140	0.4162
0.4571	191.52	3160	0.4434
0.4228	192.73	3180	0.6554
0.4345	193.94	3200	0.2984
0.4424	195.15	3220	0.3035
0.4259	196.36	3240	0.4230
0.4161	197.58	3260	0.2558
0.405	198.79	3280	0.3711
0.4385	200.0	3300	0.2988
0.4034	201.21	3320	0.4759
0.4203	202.42	3340	0.3641
0.4559	203.64	3360	0.3186
0.4457	204.85	3380	0.3593
0.4072	206.06	3400	0.3301
0.4254	207.27	3420	0.2779
0.4153	208.48	3440	0.3963
0.4259	209.7	3460	0.3817
0.4273	210.91	3480	0.3069
0.3945	212.12	3500	0.3477
0.3849	213.33	3520	0.3495
0.3944	214.55	3540	0.4825
0.3881	215.76	3560	0.3790
0.3856	216.97	3580	0.2898
0.4108	218.18	3600	0.3521
0.4194	219.39	3620	0.2938
0.3683	220.61	3640	0.2290
0.4111	221.82	3660	0.3704
0.4078	223.03	3680	0.3231
0.3852	224.24	3700	0.2568
0.407	225.45	3720	0.4309
0.3753	226.67	3740	0.3829
0.3963	227.88	3760	0.3988
0.3683	229.09	3780	0.3014
0.3786	230.3	3800	0.2988
0.3705	231.52	3820	0.3167
0.3822	232.73	3840	0.3800
0.3496	233.94	3860	0.3660
0.407	235.15	3880	0.3476
0.3938	236.36	3900	0.3337
0.3526	237.58	3920	0.3130
0.3815	238.79	3940	0.2702
0.3677	240.0	3960	0.3134
0.4319	241.21	3980	0.3871
0.401	242.42	4000	0.4471
0.3538	243.64	4020	0.3134
0.3605	244.85	4040	0.2553
0.3585	246.06	4060	0.2506
0.3879	247.27	4080	0.3194
0.3638	248.48	4100	0.4381
0.3649	249.7	4120	0.3818
0.3529	250.91	4140	0.2432
0.3841	252.12	4160	0.2769
0.3755	253.33	4180	0.3376
0.3504	254.55	4200	0.2689
0.3653	255.76	4220	0.2874
0.3614	256.97	4240	0.4095
0.3909	258.18	4260	0.2556
0.3547	259.39	4280	0.4043
0.3613	260.61	4300	0.2781
0.3268	261.82	4320	0.2558
0.367	263.03	4340	0.3386
0.3317	264.24	4360	0.2605
0.3733	265.45	4380	0.2535
0.3878	266.67	4400	0.2325
0.3596	267.88	4420	0.2849
0.3482	269.09	4440	0.2811
0.3609	270.3	4460	0.3282
0.373	271.52	4480	0.4058
0.3792	272.73	4500	0.2404
0.3563	273.94	4520	0.3351
0.3215	275.15	4540	0.4536
0.3389	276.36	4560	0.4224
0.354	277.58	4580	0.3298
0.3616	278.79	4600	0.3443
0.3629	280.0	4620	0.3889
0.3443	281.21	4640	0.3653
0.3407	282.42	4660	0.2257
0.3178	283.64	4680	0.3924
0.3364	284.85	4700	0.3184
0.3356	286.06	4720	0.3177
0.3711	287.27	4740	0.3729
0.3422	288.48	4760	0.2495
0.3375	289.7	4780	0.2142
0.3271	290.91	4800	0.3284

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.0
• Tokenizers 0.11.0