flan-t5-rouge-squad-qg-testc

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

• Loss: 0.3164
• Rouge1: 0.3601
• Rouge2: 0.1205
• Rougel: 0.3353
• Rougelsum: 0.3469

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: 0.0001
• train_batch_size: 80
• eval_batch_size: 80
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 320
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 160

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
62.2537	1.0	3	29.8253	0.0752	0.0204	0.0667	0.0672
51.2172	2.0	6	23.9235	0.0670	0.0190	0.0607	0.0609
42.2884	3.0	9	18.5434	0.0527	0.0175	0.0494	0.0496
33.4669	4.0	12	12.7318	0.0825	0.0407	0.0824	0.0827
25.4026	5.0	15	8.2492	0.0706	0.0377	0.0701	0.0704
20.0834	6.0	18	7.4425	0.0662	0.0349	0.0664	0.0666
16.9821	7.0	21	6.9817	0.0789	0.0347	0.0759	0.0776
14.7652	8.0	24	5.9311	0.0961	0.0435	0.0905	0.0937
12.7754	9.0	27	4.9313	0.1005	0.0444	0.0841	0.0916
11.6278	10.0	30	4.7801	0.1311	0.0577	0.1158	0.1214
10.6391	11.0	33	4.5698	0.0978	0.0372	0.0872	0.0905
9.9925	12.0	36	4.3618	0.1011	0.0440	0.0892	0.0934
9.4969	13.0	39	4.2002	0.1331	0.0582	0.1167	0.1221
9.0639	14.0	42	4.0571	0.1327	0.0564	0.1165	0.1217
8.7064	15.0	45	3.9041	0.1484	0.0628	0.1267	0.1337
8.3122	16.0	48	3.7252	0.1528	0.0543	0.1264	0.1345
8.0191	17.0	51	3.5352	0.1356	0.0502	0.1129	0.1204
7.7028	18.0	54	3.3741	0.1175	0.0426	0.0978	0.1045
7.3704	19.0	57	3.2478	0.1111	0.0460	0.0947	0.1004
7.059	20.0	60	3.1444	0.1052	0.0359	0.0842	0.0900
6.798	21.0	63	3.0483	0.1240	0.0444	0.0990	0.1057
6.6172	22.0	66	2.9450	0.1281	0.0416	0.1018	0.1100
6.397	23.0	69	2.8270	0.1294	0.0426	0.1058	0.1154
6.1434	24.0	72	2.6957	0.1192	0.0405	0.1011	0.1072
5.9183	25.0	75	2.5599	0.1207	0.0367	0.0974	0.1027
5.7236	26.0	78	2.4320	0.1124	0.0342	0.0950	0.1001
5.5052	27.0	81	2.3171	0.1121	0.0367	0.0961	0.1004
5.3234	28.0	84	2.2137	0.1322	0.0453	0.1132	0.1197
5.1292	29.0	87	2.1201	0.1464	0.0526	0.1248	0.1322
4.9497	30.0	90	2.0297	0.2406	0.0923	0.2102	0.2278
4.7775	31.0	93	1.9376	0.2581	0.0958	0.2229	0.2445
4.5872	32.0	96	1.8415	0.2783	0.0982	0.2442	0.2654
4.4228	33.0	99	1.7448	0.2993	0.1034	0.2699	0.2878
4.2818	34.0	102	1.6519	0.3106	0.1082	0.2810	0.2983
4.0818	35.0	105	1.5682	0.3280	0.1101	0.2963	0.3144
3.9575	36.0	108	1.4935	0.3307	0.1099	0.2992	0.3180
3.8176	37.0	111	1.4235	0.3464	0.1150	0.3140	0.3318
3.666	38.0	114	1.3551	0.3496	0.1166	0.3173	0.3349
3.5058	39.0	117	1.2865	0.3496	0.1166	0.3173	0.3349
3.3658	40.0	120	1.2200	0.3475	0.1166	0.3155	0.3334
3.2795	41.0	123	1.1562	0.3522	0.1179	0.3226	0.3386
3.1434	42.0	126	1.0954	0.3522	0.1179	0.3226	0.3386
3.0247	43.0	129	1.0422	0.3522	0.1179	0.3226	0.3386
2.9343	44.0	132	0.9925	0.3529	0.1185	0.3238	0.3393
2.8065	45.0	135	0.9465	0.3529	0.1185	0.3238	0.3393
2.7406	46.0	138	0.9023	0.3529	0.1185	0.3238	0.3393
2.6367	47.0	141	0.8608	0.3551	0.1166	0.3249	0.3427
2.4855	48.0	144	0.8197	0.3592	0.1172	0.3284	0.3457
2.4782	49.0	147	0.7803	0.3592	0.1172	0.3284	0.3457
2.3351	50.0	150	0.7463	0.3586	0.1149	0.3253	0.3446
2.2519	51.0	153	0.7154	0.3655	0.1169	0.3281	0.3517
2.1864	52.0	156	0.6861	0.3676	0.1171	0.3293	0.3530
2.128	53.0	159	0.6597	0.3676	0.1171	0.3293	0.3530
2.0668	54.0	162	0.6343	0.3676	0.1171	0.3293	0.3530
2.013	55.0	165	0.6098	0.3658	0.1167	0.3277	0.3506
1.9364	56.0	168	0.5872	0.3658	0.1167	0.3277	0.3506
1.8327	57.0	171	0.5655	0.3658	0.1167	0.3277	0.3506
1.7749	58.0	174	0.5456	0.3659	0.1169	0.3282	0.3506
1.7399	59.0	177	0.5276	0.3659	0.1169	0.3282	0.3506
1.7449	60.0	180	0.5110	0.3659	0.1169	0.3282	0.3506
1.6973	61.0	183	0.4959	0.3659	0.1169	0.3282	0.3506
1.5943	62.0	186	0.4822	0.3658	0.1185	0.3280	0.3506
1.5571	63.0	189	0.4703	0.3666	0.1194	0.3308	0.3519
1.5806	64.0	192	0.4589	0.3666	0.1194	0.3308	0.3519
1.5002	65.0	195	0.4471	0.3666	0.1194	0.3308	0.3519
1.4634	66.0	198	0.4356	0.3666	0.1194	0.3308	0.3519
1.4553	67.0	201	0.4250	0.3697	0.1216	0.3344	0.3559
1.4035	68.0	204	0.4164	0.3701	0.1222	0.3350	0.3561
1.4084	69.0	207	0.4090	0.3739	0.1255	0.3378	0.3594
1.3806	70.0	210	0.4023	0.3739	0.1255	0.3378	0.3594
1.3048	71.0	213	0.3957	0.3728	0.1254	0.3373	0.3576
1.2709	72.0	216	0.3891	0.3728	0.1254	0.3373	0.3576
1.2735	73.0	219	0.3828	0.3728	0.1254	0.3373	0.3576
1.2733	74.0	222	0.3768	0.3728	0.1254	0.3373	0.3576
1.2215	75.0	225	0.3715	0.3728	0.1254	0.3373	0.3576
1.2225	76.0	228	0.3669	0.3732	0.1255	0.3374	0.3580
1.1829	77.0	231	0.3628	0.3732	0.1255	0.3374	0.3580
1.162	78.0	234	0.3591	0.3722	0.1244	0.3362	0.3570
1.097	79.0	237	0.3556	0.3715	0.1236	0.3355	0.3564
1.1702	80.0	240	0.3519	0.3715	0.1236	0.3355	0.3564
1.1309	81.0	243	0.3483	0.3764	0.1259	0.3400	0.3608
1.0986	82.0	246	0.3451	0.3563	0.1178	0.3261	0.3428
1.1109	83.0	249	0.3422	0.3563	0.1178	0.3261	0.3428
1.0752	84.0	252	0.3397	0.3553	0.1167	0.3257	0.3417
1.0475	85.0	255	0.3374	0.3553	0.1167	0.3257	0.3417
1.0736	86.0	258	0.3353	0.3553	0.1167	0.3257	0.3417
1.0723	87.0	261	0.3333	0.3552	0.1176	0.3249	0.3409
1.0326	88.0	264	0.3314	0.3536	0.1165	0.3236	0.3394
1.0742	89.0	267	0.3297	0.3536	0.1165	0.3236	0.3394
1.0081	90.0	270	0.3281	0.3583	0.1172	0.3282	0.3447
1.0158	91.0	273	0.3266	0.3583	0.1172	0.3282	0.3447
1.032	92.0	276	0.3252	0.3632	0.1213	0.3330	0.3497
0.9778	93.0	279	0.3239	0.3632	0.1213	0.3330	0.3497
0.9834	94.0	282	0.3228	0.3624	0.1221	0.3323	0.3492
0.9913	95.0	285	0.3218	0.3624	0.1221	0.3323	0.3492
0.9963	96.0	288	0.3210	0.3624	0.1221	0.3323	0.3492
0.9759	97.0	291	0.3202	0.3605	0.1208	0.3344	0.3465
1.0011	98.0	294	0.3195	0.3605	0.1208	0.3344	0.3465
0.9895	99.0	297	0.3189	0.3605	0.1208	0.3344	0.3465
0.926	100.0	300	0.3183	0.3605	0.1208	0.3344	0.3465
0.9347	101.0	303	0.3178	0.3605	0.1208	0.3344	0.3465
1.0039	102.0	306	0.3173	0.3605	0.1208	0.3344	0.3465
0.9693	103.0	309	0.3170	0.3603	0.1205	0.3353	0.3469
0.9754	104.0	312	0.3167	0.3601	0.1205	0.3353	0.3469
0.9872	105.0	315	0.3165	0.3601	0.1205	0.3353	0.3469
1.0003	106.0	318	0.3164	0.3601	0.1205	0.3353	0.3469
1.9782	106.8	320	0.3164	0.3601	0.1205	0.3353	0.3469

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0