metadata

license: apache-2.0
base_model: google-t5/t5-small
tags:
  - generated_from_trainer
metrics:
  - rouge
model-index:
  - name: flan-t5-small-summarization
    results: []

flan-t5-small-summarization

This model is a fine-tuned version of google-t5/t5-small on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.9716
Rouge1: 14.8237
Rouge2: 5.3275
Rougel: 12.6729
Rougelsum: 13.6266
Gen Len: 18.968

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: 6
eval_batch_size: 6
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 24
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 5
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	0.12	100	2.0773	15.1231	5.4025	12.9496	13.9319	18.94
No log	0.24	200	2.0736	14.7565	5.2799	12.6268	13.5578	18.94
No log	0.36	300	2.0632	14.8383	5.2319	12.6555	13.6597	18.968
No log	0.48	400	2.0629	14.8558	5.2815	12.6581	13.6503	18.968
2.2157	0.6	500	2.0583	14.8736	5.3228	12.649	13.6717	18.968
2.2157	0.72	600	2.0520	14.8178	5.3112	12.586	13.6262	18.968
2.2157	0.84	700	2.0467	14.9042	5.3468	12.6543	13.6596	18.968
2.2157	0.96	800	2.0435	14.8682	5.3287	12.661	13.6869	18.968
2.2157	1.08	900	2.0375	14.9469	5.362	12.7083	13.7525	18.968
2.1846	1.2	1000	2.0324	14.8316	5.3471	12.6593	13.6452	18.968
2.1846	1.32	1100	2.0309	14.6717	5.2555	12.5319	13.4962	18.968
2.1846	1.44	1200	2.0189	14.8455	5.3386	12.6002	13.6588	18.968
2.1846	1.56	1300	2.0182	14.9323	5.3902	12.7187	13.7579	18.968
2.1846	1.68	1400	2.0172	14.969	5.4698	12.8021	13.8116	18.968
2.1596	1.8	1500	2.0105	15.0152	5.5355	12.8098	13.8475	18.968
2.1596	1.92	1600	2.0100	15.0009	5.3835	12.764	13.785	18.968
2.1596	2.04	1700	2.0083	14.8145	5.2912	12.6179	13.6279	18.968
2.1596	2.16	1800	2.0035	14.8232	5.2131	12.6386	13.6297	18.968
2.1596	2.28	1900	2.0006	14.8076	5.2617	12.6578	13.6631	18.968
2.1405	2.4	2000	1.9983	14.6508	5.0855	12.4956	13.4989	18.968
2.1405	2.52	2100	1.9965	14.9548	5.2857	12.6947	13.7664	18.968
2.1405	2.64	2200	1.9917	14.8786	5.2212	12.6813	13.6609	18.968
2.1405	2.76	2300	1.9904	15.0902	5.4835	12.8911	13.9191	18.968
2.1405	2.88	2400	1.9880	14.8188	5.2057	12.6325	13.6335	18.968
2.1287	3.0	2500	1.9844	14.7362	5.2487	12.6559	13.64	18.968
2.1287	3.12	2600	1.9834	14.9356	5.3404	12.7325	13.7185	18.968
2.1287	3.24	2700	1.9839	14.9543	5.4587	12.757	13.767	18.968
2.1287	3.36	2800	1.9821	14.8174	5.2522	12.6935	13.6292	18.968
2.1287	3.48	2900	1.9816	14.8201	5.2606	12.6679	13.6275	18.968
2.1149	3.6	3000	1.9795	14.8112	5.253	12.5789	13.5714	18.968
2.1149	3.72	3100	1.9788	14.7946	5.3272	12.6237	13.614	18.968
2.1149	3.84	3200	1.9761	14.8197	5.295	12.6209	13.6327	18.968
2.1149	3.96	3300	1.9761	14.7752	5.2759	12.6239	13.6167	18.968
2.1149	4.08	3400	1.9714	14.7938	5.2988	12.7085	13.6708	18.968
2.1138	4.2	3500	1.9729	14.8006	5.2526	12.6427	13.6018	18.968
2.1138	4.32	3600	1.9751	14.7531	5.2913	12.6372	13.5782	18.968
2.1138	4.44	3700	1.9743	14.7556	5.2694	12.6372	13.5786	18.968
2.1138	4.56	3800	1.9710	14.8124	5.2887	12.7095	13.6666	18.968
2.1138	4.68	3900	1.9725	14.7104	5.2357	12.5839	13.5364	18.968
2.1033	4.8	4000	1.9726	14.7673	5.2771	12.6343	13.5731	18.968
2.1033	4.92	4100	1.9716	14.8237	5.3275	12.6729	13.6266	18.968

Framework versions

Transformers 4.38.2
Pytorch 2.2.1+cu121
Datasets 2.18.0
Tokenizers 0.15.2