sewd-classifier-aug-large

This model is a fine-tuned version of asapp/sew-d-tiny-100k-ft-ls100h on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.2094
• Accuracy: 0.1712
• Precision: 0.0867
• Recall: 0.1712
• F1: 0.0887
• Binary: 0.4166

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: 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
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Binary
No log	0.1	50	4.4166	0.0108	0.0010	0.0108	0.0013	0.1291
No log	0.2	100	4.3510	0.0135	0.0069	0.0135	0.0026	0.1367
No log	0.29	150	4.2531	0.0323	0.0016	0.0323	0.0030	0.2372
No log	0.39	200	4.1748	0.0350	0.0022	0.0350	0.0039	0.2779
No log	0.49	250	4.0970	0.0431	0.0076	0.0431	0.0094	0.2969
No log	0.59	300	4.0268	0.0499	0.0179	0.0499	0.0116	0.3198
No log	0.69	350	3.9648	0.0472	0.0029	0.0472	0.0054	0.3175
No log	0.78	400	3.8977	0.0553	0.0279	0.0553	0.0164	0.3283
No log	0.88	450	3.8290	0.0768	0.0239	0.0768	0.0230	0.3437
No log	0.98	500	3.7650	0.0701	0.0313	0.0701	0.0252	0.3418
4.166	1.08	550	3.7060	0.0997	0.0444	0.0997	0.0418	0.3625
4.166	1.18	600	3.6526	0.1132	0.0497	0.1132	0.0501	0.3733
4.166	1.27	650	3.5931	0.1132	0.0433	0.1132	0.0509	0.3732
4.166	1.37	700	3.5467	0.1240	0.0664	0.1240	0.0558	0.3805
4.166	1.47	750	3.4981	0.1213	0.0589	0.1213	0.0553	0.3792
4.166	1.57	800	3.4544	0.1361	0.0606	0.1361	0.0641	0.3904
4.166	1.67	850	3.4081	0.1456	0.0557	0.1456	0.0655	0.3980
4.166	1.76	900	3.3696	0.1536	0.0812	0.1536	0.0786	0.4040
4.166	1.86	950	3.3213	0.1523	0.0609	0.1523	0.0745	0.4038
4.166	1.96	1000	3.2814	0.1779	0.0991	0.1779	0.0922	0.4217
3.6654	2.06	1050	3.2435	0.1550	0.0883	0.1550	0.0737	0.4061
3.6654	2.16	1100	3.2094	0.1712	0.0867	0.1712	0.0887	0.4166

Framework versions

• Transformers 4.38.2
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.15.1