flan-t5-base-extraction-cnndm_20000-all-hint_precision-ep50-nonstop

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: 1.6910
• Hint Hit Num: 2.3981
• Hint Precision: 0.431
• Num: 5.5422
• Gen Len: 18.9991

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: 16
• eval_batch_size: 200
• 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
2.0952	0.8	1000	1.7450	2.2694	0.4207	5.4015	18.9993
1.9364	1.6	2000	1.7131	2.2371	0.4159	5.3613	19.0
1.8853	2.4	3000	1.7026	2.2893	0.4213	5.4161	18.9999
1.8383	3.2	4000	1.6955	2.2796	0.4206	5.404	18.9999
1.8087	4.0	5000	1.6866	2.3175	0.4244	5.4464	18.9996
1.7778	4.8	6000	1.6829	2.3311	0.423	5.4935	18.9996
1.7412	5.6	7000	1.6863	2.3112	0.4205	5.4712	18.9998
1.732	6.4	8000	1.6903	2.3108	0.4209	5.4711	18.999
1.6993	7.2	9000	1.6840	2.3855	0.4289	5.5382	18.9998
1.688	8.0	10000	1.6870	2.3089	0.4191	5.4856	18.9995
1.6609	8.8	11000	1.6910	2.3981	0.431	5.5422	18.9991
1.6462	9.6	12000	1.7011	2.3492	0.4237	5.5193	18.9994
1.6302	10.4	13000	1.7008	2.3825	0.4279	5.5488	18.999
1.6108	11.2	14000	1.7058	2.3274	0.4217	5.496	18.9998
1.6031	12.0	15000	1.7092	2.3741	0.4265	5.5432	18.999
1.5798	12.8	16000	1.7112	2.3416	0.4217	5.5248	18.9981
1.5664	13.6	17000	1.7210	2.4102	0.4291	5.6002	18.9986
1.5521	14.4	18000	1.7193	2.3779	0.4236	5.5859	18.9992
1.5426	15.2	19000	1.7323	2.3727	0.4227	5.5887	18.9992
1.5318	16.0	20000	1.7214	2.3992	0.4274	5.593	18.999
1.5134	16.8	21000	1.7300	2.4111	0.4272	5.6176	18.9987
1.5031	17.6	22000	1.7363	2.3823	0.425	5.5836	18.999
1.4845	18.4	23000	1.7429	2.4123	0.428	5.6077	18.9987
1.4895	19.2	24000	1.7534	2.3726	0.4207	5.6104	18.9984
1.4687	20.0	25000	1.7552	2.4185	0.4268	5.6374	18.9987
1.4601	20.8	26000	1.7602	2.3924	0.4251	5.6007	18.9989
1.4486	21.6	27000	1.7687	2.3863	0.4234	5.6088	18.9984
1.4353	22.4	28000	1.7770	2.3885	0.4234	5.611	18.9982
1.4317	23.2	29000	1.7788	2.4084	0.4243	5.6463	18.9987
1.4269	24.0	30000	1.7786	2.428	0.4274	5.6495	18.9985
1.4135	24.8	31000	1.7883	2.3993	0.4244	5.6265	18.9981
1.4025	25.6	32000	1.7911	2.3972	0.4225	5.6432	18.9977
1.3874	26.4	33000	1.7930	2.3838	0.4207	5.6284	18.9989
1.4023	27.2	34000	1.7988	2.436	0.4277	5.6637	18.9981
1.3796	28.0	35000	1.8079	2.4162	0.4256	5.6432	18.9981
1.3729	28.8	36000	1.8124	2.3894	0.4225	5.6167	18.9975
1.3686	29.6	37000	1.8153	2.4301	0.4271	5.6606	18.9978
1.3603	30.4	38000	1.8174	2.4248	0.4253	5.6696	18.9973
1.3551	31.2	39000	1.8224	2.42	0.4243	5.67	18.9976
1.3504	32.0	40000	1.8246	2.4189	0.4254	5.6551	18.9977
1.3447	32.8	41000	1.8222	2.4234	0.425	5.6685	18.9969
1.3354	33.6	42000	1.8380	2.3975	0.422	5.6471	18.997
1.3304	34.4	43000	1.8416	2.4161	0.4247	5.6584	18.9975
1.3274	35.2	44000	1.8386	2.4271	0.4249	5.6804	18.9972
1.3238	36.0	45000	1.8361	2.4164	0.4241	5.6643	18.9972
1.3167	36.8	46000	1.8418	2.4359	0.426	5.6864	18.9973
1.3115	37.6	47000	1.8499	2.4068	0.4222	5.6682	18.9972
1.31	38.4	48000	1.8508	2.433	0.4256	5.6872	18.997
1.3085	39.2	49000	1.8477	2.4184	0.423	5.682	18.9967
1.3009	40.0	50000	1.8485	2.4182	0.4236	5.6753	18.9972
1.3028	40.8	51000	1.8547	2.4074	0.4222	5.6657	18.9977
1.2919	41.6	52000	1.8552	2.4199	0.4233	5.6825	18.9965
1.2945	42.4	53000	1.8652	2.4227	0.4238	5.6853	18.9975
1.2931	43.2	54000	1.8605	2.427	0.424	5.6907	18.9971
1.2838	44.0	55000	1.8647	2.4244	0.4244	5.6774	18.9972
1.2878	44.8	56000	1.8629	2.4209	0.4234	5.6848	18.997
1.2848	45.6	57000	1.8674	2.4291	0.4242	5.6931	18.9966
1.279	46.4	58000	1.8649	2.4253	0.4238	5.6898	18.9966
1.2862	47.2	59000	1.8643	2.4187	0.4228	5.686	18.9966
1.2798	48.0	60000	1.8643	2.4195	0.4231	5.6832	18.997
1.279	48.8	61000	1.8670	2.424	0.4239	5.6845	18.9966
1.2754	49.6	62000	1.8686	2.4234	0.4238	5.6842	18.9965

Framework versions

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