mrdbourke/detr_finetuned_trashify_box_detector_synthetic_data_only

trashify-box-detection-model-synthetic-data-only

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

• Loss: 0.7824

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
67.3985	1.0	118	4.7679
3.2012	2.0	236	2.1811
2.1642	3.0	354	1.6674
1.7756	4.0	472	1.3667
1.5894	5.0	590	1.3178
1.497	6.0	708	1.1129
1.4228	7.0	826	1.1679
1.4065	8.0	944	1.1444
1.4096	9.0	1062	1.1029
1.3762	10.0	1180	1.0651
1.33	11.0	1298	1.0473
1.3183	12.0	1416	1.0753
1.3158	13.0	1534	1.0515
1.2898	14.0	1652	0.9913
1.292	15.0	1770	0.9918
1.2323	16.0	1888	0.9896
1.2356	17.0	2006	1.0034
1.2407	18.0	2124	0.9813
1.2384	19.0	2242	0.9685
1.2229	20.0	2360	0.9606
1.2011	21.0	2478	0.9410
1.1809	22.0	2596	0.9628
1.2145	23.0	2714	0.9707
1.208	24.0	2832	0.9811
1.2109	25.0	2950	0.9749
1.2073	26.0	3068	0.9966
1.1875	27.0	3186	0.9225
1.1786	28.0	3304	0.9692
1.1782	29.0	3422	0.9331
1.1617	30.0	3540	0.9494
1.1636	31.0	3658	0.9176
1.1391	32.0	3776	0.9059
1.1643	33.0	3894	0.9177
1.1486	34.0	4012	0.9351
1.1229	35.0	4130	0.9027
1.11	36.0	4248	0.9060
1.112	37.0	4366	0.8910
1.1112	38.0	4484	0.8954
1.0994	39.0	4602	0.8769
1.0821	40.0	4720	0.9111
1.088	41.0	4838	0.8496
1.049	42.0	4956	0.8694
1.0804	43.0	5074	0.8783
1.0643	44.0	5192	0.8671
1.06	45.0	5310	0.8707
1.0558	46.0	5428	0.8754
1.0448	47.0	5546	0.8606
1.0341	48.0	5664	0.8649
1.0433	49.0	5782	0.8479
1.0211	50.0	5900	0.8530
1.0186	51.0	6018	0.8419
0.9921	52.0	6136	0.8340
1.0261	53.0	6254	0.8279
1.0006	54.0	6372	0.8395
0.9926	55.0	6490	0.8368
0.9965	56.0	6608	0.8197
0.997	57.0	6726	0.8420
0.9638	58.0	6844	0.8309
0.9667	59.0	6962	0.8085
0.9561	60.0	7080	0.8440
0.9654	61.0	7198	0.8254
0.969	62.0	7316	0.8226
0.9621	63.0	7434	0.8033
0.9466	64.0	7552	0.8182
0.9406	65.0	7670	0.8176
0.9365	66.0	7788	0.8037
0.9163	67.0	7906	0.8124
0.9133	68.0	8024	0.8080
0.8986	69.0	8142	0.7981
0.9227	70.0	8260	0.8112
0.9212	71.0	8378	0.8031
0.8991	72.0	8496	0.7998
0.9026	73.0	8614	0.7971
0.8794	74.0	8732	0.7959
0.8874	75.0	8850	0.7961
0.8922	76.0	8968	0.8009
0.8905	77.0	9086	0.7930
0.8785	78.0	9204	0.7991
0.8723	79.0	9322	0.7887
0.8634	80.0	9440	0.7800
0.847	81.0	9558	0.7903
0.8418	82.0	9676	0.7874
0.852	83.0	9794	0.7950
0.8322	84.0	9912	0.7967
0.8489	85.0	10030	0.7806
0.8359	86.0	10148	0.7919
0.8279	87.0	10266	0.7896
0.8453	88.0	10384	0.7926
0.8029	89.0	10502	0.7913
0.8088	90.0	10620	0.7843
0.8269	91.0	10738	0.7820
0.8201	92.0	10856	0.7880
0.8177	93.0	10974	0.7847
0.8224	94.0	11092	0.7869
0.8134	95.0	11210	0.7869
0.799	96.0	11328	0.7866
0.8229	97.0	11446	0.7864
0.7831	98.0	11564	0.7857
0.7936	99.0	11682	0.7835
0.7706	100.0	11800	0.7824

Framework versions

• Transformers 4.44.0
• Pytorch 2.2.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1