joel4899/bart-lowlr

output

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

• Loss: 0.0322

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss
14.3684	1.0	27	15.3473
12.8484	2.0	54	12.2183
10.0848	3.0	81	7.2310
6.5549	4.0	108	2.0826
3.238	5.0	135	0.3298
1.6033	6.0	162	0.3247
0.9275	7.0	189	0.2843
0.402	8.0	216	0.2750
0.2528	9.0	243	0.2256
0.2027	10.0	270	0.1815
0.1671	11.0	297	0.1480
0.1396	12.0	324	0.1112
0.1088	13.0	351	0.0797
0.0871	14.0	378	0.0638
0.0725	15.0	405	0.0582
0.0643	16.0	432	0.0546
0.0585	17.0	459	0.0522
0.0529	18.0	486	0.0498
0.0497	19.0	513	0.0478
0.0472	20.0	540	0.0473
0.0449	21.0	567	0.0461
0.0427	22.0	594	0.0455
0.0424	23.0	621	0.0449
0.042	24.0	648	0.0442
0.0415	25.0	675	0.0437
0.0396	26.0	702	0.0436
0.0398	27.0	729	0.0434
0.0398	28.0	756	0.0432
0.039	29.0	783	0.0432
0.0387	30.0	810	0.0432

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0+cpu
• Datasets 3.0.1
• Tokenizers 0.20.0