flan-t5-base-extraction-cnndm_8000-all-loss-ep50

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

• Loss: 1.7340
• Hint Hit Num: 2.3422
• Hint Precision: 0.4261
• Num: 5.4786
• Gen Len: 18.9996

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
• train_batch_size: 60
• eval_batch_size: 400
• seed: 1799
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Hint Hit Num	Hint Precision	Num	Gen Len
2.4434	0.75	100	1.9136	2.1212	0.4054	5.2222	18.9802
2.1367	1.49	200	1.8314	2.2656	0.4192	5.406	18.9969
2.0682	2.24	300	1.8232	2.2491	0.422	5.3241	18.9982
2.0331	2.99	400	1.7910	2.3106	0.4271	5.4092	18.999
1.9952	3.73	500	1.7794	2.2826	0.4217	5.4027	18.9992
1.9603	4.48	600	1.7714	2.3004	0.4245	5.4066	18.9994
1.9473	5.22	700	1.7633	2.3693	0.4321	5.4749	18.9996
1.9293	5.97	800	1.7619	2.2834	0.4222	5.3942	18.9997
1.9092	6.72	900	1.7556	2.3342	0.4263	5.4603	18.9999
1.8954	7.46	1000	1.7560	2.3196	0.4262	5.4271	18.9994
1.886	8.21	1100	1.7389	2.3588	0.4307	5.4604	18.9996
1.8532	8.96	1200	1.7499	2.3351	0.4279	5.4403	18.9998
1.8496	9.7	1300	1.7390	2.3407	0.4267	5.4666	18.9996
1.8309	10.45	1400	1.7436	2.337	0.4281	5.4377	18.9992
1.8208	11.19	1500	1.7375	2.3605	0.4292	5.4822	18.9997
1.8082	11.94	1600	1.7390	2.3582	0.4303	5.46	18.9993
1.805	12.69	1700	1.7340	2.3422	0.4261	5.4786	18.9996
1.7704	13.43	1800	1.7448	2.351	0.428	5.4723	18.9996
1.7868	14.18	1900	1.7367	2.3787	0.4303	5.5085	18.9996
1.7693	14.93	2000	1.7445	2.3526	0.4292	5.4647	18.9992
1.7541	15.67	2100	1.7391	2.3799	0.4314	5.5012	18.9996
1.7465	16.42	2200	1.7475	2.3556	0.4281	5.4835	18.9991

Framework versions

• Transformers 4.18.0
• Pytorch 1.10.0+cu111
• Datasets 2.5.1
• Tokenizers 0.12.1