Modelpublic

Mauregato/qqq-finetuned-on-calls

sourceHugging Facemitupdated 3y agoView on Hugging Face

0likes7downloads

Model Card

qqq-finetuned-on-calls

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

—Loss: 0.0021
—Rouge-1: 1.0
—Rouge-2: 1.0
—Rouge-l: 1.0
—Gen Len: 11.0
—Avg Rouge F: 1.0

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: 2e-05
—trainbatchsize: 3
—evalbatchsize: 1
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—lrschedulerwarmup_steps: 50
—num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge-1	Rouge-2	Rouge-l	Gen Len	Avg Rouge F
2.1855	3.12	25	1.4282	0.0	0.0	0.0	15.0	0.0
1.5665	6.25	50	0.6420	0.1818	0.0	0.1818	12.0	0.1212
1.1046	9.38	75	0.2184	1.0	1.0	1.0	11.0	1.0
0.8218	12.5	100	0.1098	1.0	1.0	1.0	11.0	1.0
0.606	15.62	125	0.0749	1.0	1.0	1.0	11.0	1.0
0.5488	18.75	150	0.0577	1.0	1.0	1.0	11.0	1.0
0.4161	21.88	175	0.0684	1.0	1.0	1.0	11.0	1.0
0.3196	25.0	200	0.0570	1.0	1.0	1.0	11.0	1.0
0.2929	28.12	225	0.0416	1.0	1.0	1.0	11.0	1.0
0.2519	31.25	250	0.0247	1.0	1.0	1.0	11.0	1.0
0.178	34.38	275	0.0118	1.0	1.0	1.0	11.0	1.0
0.1603	37.5	300	0.0064	1.0	1.0	1.0	11.0	1.0
0.1684	40.62	325	0.0051	1.0	1.0	1.0	11.0	1.0
0.1326	43.75	350	0.0051	1.0	1.0	1.0	11.0	1.0
0.1349	46.88	375	0.0064	1.0	1.0	1.0	11.0	1.0
0.1105	50.0	400	0.0061	1.0	1.0	1.0	11.0	1.0
0.1026	53.12	425	0.0049	1.0	1.0	1.0	11.0	1.0
0.0936	56.25	450	0.0030	1.0	1.0	1.0	11.0	1.0
0.0704	59.38	475	0.0025	1.0	1.0	1.0	11.0	1.0
0.0699	62.5	500	0.0021	1.0	1.0	1.0	11.0	1.0
0.0863	65.62	525	0.0020	1.0	1.0	1.0	11.0	1.0
0.0595	68.75	550	0.0024	1.0	1.0	1.0	11.0	1.0
0.0594	71.88	575	0.0028	1.0	1.0	1.0	11.0	1.0
0.0683	75.0	600	0.0026	1.0	1.0	1.0	11.0	1.0
0.074	78.12	625	0.0025	1.0	1.0	1.0	11.0	1.0

Framework versions

—Transformers 4.29.0
—Pytorch 2.0.0+cu118
—Datasets 2.12.0
—Tokenizers 0.13.3