pranaydeeps/lettuce_pos_en_mono

pos_final_mono_en

This model is a fine-tuned version of FacebookAI/roberta-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0681
• Precision: 0.9696
• Recall: 0.9714
• F1: 0.9705
• Accuracy: 0.9796

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: 5e-05
• train_batch_size: 256
• eval_batch_size: 256
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 1024
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 40.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.99	60	2.7933	0.3216	0.0997	0.1522	0.2833
No log	1.99	120	0.3818	0.9075	0.8989	0.9032	0.9224
No log	2.99	180	0.1156	0.9602	0.9607	0.9605	0.9721
No log	3.99	240	0.0911	0.9634	0.9650	0.9642	0.9748
No log	4.99	300	0.0794	0.9664	0.9679	0.9671	0.9772
No log	5.99	360	0.0741	0.9670	0.9697	0.9683	0.9781
No log	6.99	420	0.0695	0.9683	0.9702	0.9693	0.9787
No log	7.99	480	0.0688	0.9686	0.9700	0.9693	0.9789
0.7281	8.99	540	0.0675	0.9688	0.9703	0.9695	0.9789
0.7281	9.99	600	0.0670	0.9687	0.9705	0.9696	0.9791
0.7281	10.99	660	0.0658	0.9696	0.9702	0.9699	0.9792
0.7281	11.99	720	0.0670	0.9684	0.9715	0.9700	0.9793
0.7281	12.99	780	0.0672	0.9689	0.9711	0.9700	0.9792
0.7281	13.99	840	0.0678	0.9698	0.9708	0.9703	0.9796
0.7281	14.99	900	0.0681	0.9696	0.9714	0.9705	0.9796
0.7281	15.99	960	0.0706	0.9696	0.9711	0.9703	0.9795
0.0484	16.99	1020	0.0725	0.9694	0.9705	0.9699	0.9793
0.0484	17.99	1080	0.0735	0.9689	0.9705	0.9697	0.9791
0.0484	18.99	1140	0.0745	0.9690	0.9705	0.9698	0.9792
0.0484	19.99	1200	0.0769	0.9690	0.9706	0.9698	0.9791
0.0484	20.99	1260	0.0797	0.9691	0.9703	0.9697	0.9791
0.0484	21.99	1320	0.0808	0.9689	0.9705	0.9697	0.9791
0.0484	22.99	1380	0.0838	0.9691	0.9702	0.9697	0.9791
0.0484	23.99	1440	0.0861	0.9685	0.9704	0.9695	0.9789
0.0289	24.99	1500	0.0879	0.9684	0.9698	0.9691	0.9787
0.0289	25.99	1560	0.0887	0.9684	0.9703	0.9694	0.9789
0.0289	26.99	1620	0.0910	0.9684	0.9698	0.9691	0.9787
0.0289	27.99	1680	0.0924	0.9684	0.9697	0.9691	0.9787
0.0289	28.99	1740	0.0950	0.9693	0.9692	0.9693	0.9788
0.0289	29.99	1800	0.0962	0.9692	0.9697	0.9694	0.9789
0.0289	30.99	1860	0.0977	0.9687	0.9699	0.9693	0.9787
0.0289	31.99	1920	0.0979	0.9688	0.9699	0.9694	0.9788
0.0289	32.99	1980	0.1000	0.9687	0.9698	0.9692	0.9788
0.018	33.99	2040	0.1021	0.9688	0.9698	0.9693	0.9788
0.018	34.99	2100	0.1037	0.9687	0.9701	0.9694	0.9788
0.018	35.99	2160	0.1035	0.9688	0.9703	0.9696	0.9790
0.018	36.99	2220	0.1042	0.9688	0.9700	0.9694	0.9789
0.018	37.99	2280	0.1053	0.9685	0.9699	0.9692	0.9787
0.018	38.99	2340	0.1052	0.9689	0.9700	0.9695	0.9789
0.018	39.99	2400	0.1054	0.9688	0.9700	0.9694	0.9788

Framework versions

• Transformers 4.25.1
• Pytorch 1.12.0
• Datasets 2.18.0
• Tokenizers 0.13.2