flanT5_large_FINAL_Task2_semantic_pred

This model is a fine-tuned version of google/flan-t5-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2525
• Accuracy: 0.8814
• Precision: 0.8988
• Recall: 0.8604
• F1 score: 0.8792

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: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1 score
0.8828	0.4257	2500	0.8908	0.8257	0.7870	0.8946	0.8373
0.7602	0.8515	5000	0.9174	0.8371	0.8395	0.8348	0.8371
0.6466	1.2772	7500	0.6703	0.8671	0.9216	0.8034	0.8584
0.5325	1.7030	10000	0.6555	0.8657	0.8560	0.8803	0.8680
0.5041	2.1287	12500	0.8110	0.8557	0.8592	0.8519	0.8555
0.3444	2.5545	15000	0.8804	0.8729	0.8701	0.8775	0.8738
0.3324	2.9802	17500	0.8064	0.8686	0.8866	0.8462	0.8659
0.2141	3.4060	20000	0.7677	0.8686	0.8529	0.8917	0.8719
0.2403	3.8317	22500	0.6505	0.8686	0.9218	0.8063	0.8602
0.1754	4.2575	25000	1.1009	0.8543	0.832	0.8889	0.8595
0.1455	4.6832	27500	0.9066	0.8557	0.8189	0.9145	0.8641
0.1004	5.1090	30000	1.0664	0.8686	0.8607	0.8803	0.8704
0.0522	5.5347	32500	1.2579	0.8729	0.8808	0.8632	0.8719
0.0742	5.9605	35000	1.0095	0.8786	0.8912	0.8632	0.8770
0.0435	6.3862	37500	1.3122	0.8614	0.8342	0.9031	0.8673
0.0263	6.8120	40000	1.1928	0.8743	0.8725	0.8775	0.875
0.0226	7.2377	42500	1.2052	0.8757	0.8882	0.8604	0.8741
0.0219	7.6635	45000	1.1019	0.8829	0.8854	0.8803	0.8829
0.0271	8.0892	47500	1.2264	0.88	0.8761	0.8860	0.8810
0.0073	8.5150	50000	1.2916	0.8786	0.8866	0.8689	0.8777
0.0197	8.9407	52500	1.1860	0.88	0.8915	0.8661	0.8786
0.0066	9.3665	55000	1.2615	0.88	0.8938	0.8632	0.8783
0.0098	9.7922	57500	1.2525	0.8814	0.8988	0.8604	0.8792

Framework versions

• Transformers 4.44.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1