t5-small-fine-tuned_model_4

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

• Loss: 0.5774
• Rouge1: 36.746
• Rouge2: 27.845
• Rougel: 33.0926
• Rougelsum: 33.2212
• Gen Len: 1103.6667

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 200
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.9231	25	2.8693	12.3052	0.5791	11.7049	9.5975	60.6667
No log	3.8462	50	2.4096	10.877	0.8174	10.3456	8.2477	60.5
No log	5.7692	75	2.0937	9.5364	0.8174	8.9546	7.0331	77.5
No log	7.6923	100	1.9212	9.4414	0.7372	8.7846	6.8375	77.5
No log	9.6154	125	1.7628	9.4796	0.8174	8.9051	7.0161	77.8333
No log	11.5385	150	1.6492	9.3248	0.7389	8.7894	6.8375	77.8333
No log	13.4615	175	1.5335	7.5301	0.6238	7.2128	4.7712	399.6667
No log	15.3846	200	1.4494	7.4893	0.6238	7.1749	5.0595	399.6667
No log	17.3077	225	1.3791	3.6219	0.2407	3.4618	3.4618	1383.8333
No log	19.2308	250	1.3344	2.3996	0.3987	1.9002	1.9002	1061.6667
No log	21.1538	275	1.2727	4.0949	0.2788	3.5988	3.5988	1061.6667
No log	23.0769	300	1.2295	4.4727	0.2404	4.0127	4.0344	739.5
No log	25.0	325	1.1843	4.4115	0.1604	3.6418	2.8481	1080.5
No log	26.9231	350	1.1238	4.3344	0.1606	3.7237	2.8672	1080.5
No log	28.8462	375	1.0926	4.4134	0.1604	3.6827	2.8672	1080.5
No log	30.7692	400	1.0685	4.4134	0.1604	3.6827	2.8672	1080.5
No log	32.6923	425	1.0397	4.4996	0.2589	3.9347	2.9913	1080.5
No log	34.6154	450	1.0090	5.0042	0.7844	4.4025	3.5377	1080.5
No log	36.5385	475	0.9822	5.282	1.0387	4.7809	4.3066	1080.5
1.8477	38.4615	500	0.9604	5.1772	0.9156	4.3204	3.8574	1080.5
1.8477	40.3846	525	0.9432	5.1273	0.9511	4.6355	4.0124	1080.5
1.8477	42.3077	550	0.9267	4.8886	0.8802	4.3344	3.7113	1080.5
1.8477	44.2308	575	0.9095	5.1853	0.962	4.4102	4.669	1080.5
1.8477	46.1538	600	0.8909	5.4135	0.9609	4.4226	3.8166	1080.5
1.8477	48.0769	625	0.8768	5.2497	0.882	4.3238	3.7122	1079.0
1.8477	50.0	650	0.8723	5.0982	0.7553	4.3749	4.3948	1080.6667
1.8477	51.9231	675	0.8593	5.1037	0.7584	4.1406	4.1194	1080.5
1.8477	53.8462	700	0.8546	2.8445	0.6707	2.3994	2.3994	1080.5
1.8477	55.7692	725	0.8369	3.0259	0.8755	2.5487	2.5491	1080.5
1.8477	57.6923	750	0.8265	2.9364	0.8348	2.4559	2.4559	1080.5
1.8477	59.6154	775	0.8177	12.9616	5.7206	12.0543	12.3145	743.3333
1.8477	61.5385	800	0.8044	7.5774	3.6548	7.474	7.5189	1065.8333
1.8477	63.4615	825	0.7919	5.5517	1.0848	5.2334	4.1167	1058.6667
1.8477	65.3846	850	0.7907	15.4548	7.5458	13.4559	12.6255	518.6667
1.8477	67.3077	875	0.7854	12.3467	5.8196	11.7847	10.6211	490.6667
1.8477	69.2308	900	0.7807	19.1325	10.1098	17.0462	16.2033	179.0
1.8477	71.1538	925	0.7679	10.5941	4.8942	10.0005	8.6433	736.6667
1.8477	73.0769	950	0.7609	16.2234	10.4765	15.0	14.4657	499.6667
1.8477	75.0	975	0.7566	17.9094	12.6672	16.8839	16.2986	800.3333
1.067	76.9231	1000	0.7485	18.6028	12.7736	16.6168	16.1664	813.6667
1.067	78.8462	1025	0.7483	17.2067	12.1333	15.5985	15.2003	867.1667
1.067	80.7692	1050	0.7326	25.0325	19.4385	23.4785	23.4473	1113.6667
1.067	82.6923	1075	0.7222	24.0064	18.9078	22.4609	22.297	1113.6667
1.067	84.6154	1100	0.7171	29.6848	22.1186	27.4607	27.9158	1091.8333
1.067	86.5385	1125	0.7158	23.6259	17.0707	21.4673	20.8102	1386.5
1.067	88.4615	1150	0.7011	22.7916	17.2723	21.5803	20.3957	1064.3333
1.067	90.3846	1175	0.7069	30.4225	20.1908	25.9389	24.8933	1252.5
1.067	92.3077	1200	0.6928	24.476	18.5105	21.7682	20.6288	1419.1667
1.067	94.2308	1225	0.6924	34.3412	23.4853	30.2292	28.6861	1402.6667
1.067	96.1538	1250	0.6900	36.8626	24.9817	31.8421	30.1301	1402.6667
1.067	98.0769	1275	0.6838	36.4213	25.4676	32.0357	30.8773	1098.0
1.067	100.0	1300	0.6786	27.8196	18.9336	24.1778	23.6849	1419.3333
1.067	101.9231	1325	0.6758	36.6034	27.2136	31.8824	30.9734	1094.6667
1.067	103.8462	1350	0.6661	35.9341	26.937	31.4599	30.4341	1096.5
1.067	105.7692	1375	0.6649	37.8479	26.9005	33.634	32.6227	1402.6667
1.067	107.6923	1400	0.6592	39.6796	28.1931	33.4312	32.7645	1119.6667
1.067	109.6154	1425	0.6571	33.8316	23.0668	30.0589	29.8248	1444.0
1.067	111.5385	1450	0.6578	39.1579	29.4923	34.6578	34.0201	1150.8333
1.067	113.4615	1475	0.6500	42.4567	28.9735	37.3988	35.6678	835.3333
0.8748	115.3846	1500	0.6451	37.9141	25.7378	33.9276	31.9374	1192.0
0.8748	117.3077	1525	0.6422	33.8898	24.3709	30.1179	28.4935	1433.8333
0.8748	119.2308	1550	0.6368	41.9531	28.6251	36.0198	34.2882	1161.1667
0.8748	121.1538	1575	0.6344	37.3211	25.806	32.6797	30.9541	1426.0
0.8748	123.0769	1600	0.6304	42.411	29.4476	36.0243	34.2522	844.8333
0.8748	125.0	1625	0.6276	38.234	26.7376	31.9694	30.1662	1282.0
0.8748	126.9231	1650	0.6257	29.1	21.2848	26.7081	26.0839	1582.0
0.8748	128.8462	1675	0.6227	36.314	26.1402	31.5141	29.605	901.0
0.8748	130.7692	1700	0.6198	35.9501	25.86	31.4977	29.631	1426.0
0.8748	132.6923	1725	0.6224	35.7537	25.5328	30.8648	29.0449	1213.1667
0.8748	134.6154	1750	0.6172	26.1073	20.1223	23.4478	23.2837	1415.6667
0.8748	136.5385	1775	0.6207	29.6349	20.975	25.1467	23.1802	1183.1667
0.8748	138.4615	1800	0.6137	37.7711	27.3393	33.4935	32.0528	856.0
0.8748	140.3846	1825	0.6100	43.4868	32.9129	38.9293	37.6363	802.8333
0.8748	142.3077	1850	0.6091	33.0147	23.8658	29.2587	27.4546	1121.3333
0.8748	144.2308	1875	0.6075	43.7785	31.7055	38.7407	37.2803	1118.8333
0.8748	146.1538	1900	0.6028	26.047	19.8575	23.2564	22.9809	1413.6667
0.8748	148.0769	1925	0.6020	25.8242	19.4078	23.0936	22.7658	1307.6667
0.8748	150.0	1950	0.6017	29.0892	22.112	26.4017	26.2841	1203.8333
0.8748	151.9231	1975	0.5972	35.2639	27.0773	31.3233	31.6303	1110.3333
0.7957	153.8462	2000	0.5973	34.7757	25.617	31.3525	29.5239	909.6667
0.7957	155.7692	2025	0.5963	39.2421	28.6572	35.3195	33.3751	1141.5
0.7957	157.6923	2050	0.5971	40.705	28.8564	35.7408	34.4547	825.0
0.7957	159.6154	2075	0.5940	32.7885	24.3525	29.2913	28.7441	1137.0
0.7957	161.5385	2100	0.5918	36.1103	26.7078	32.6294	32.165	1109.3333
0.7957	163.4615	2125	0.5901	33.3945	24.3106	29.5642	29.1211	1122.0
0.7957	165.3846	2150	0.5887	33.0641	25.1483	28.5307	28.5341	1136.5
0.7957	167.3077	2175	0.5880	33.9058	25.3162	30.4182	30.4334	1122.3333
0.7957	169.2308	2200	0.5856	40.1053	29.2168	35.5522	35.4371	823.8333
0.7957	171.1538	2225	0.5831	40.3565	29.3126	35.7256	36.0169	822.5
0.7957	173.0769	2250	0.5841	32.7059	24.2819	28.9926	29.0977	1123.6667
0.7957	175.0	2275	0.5833	31.6615	23.8942	28.0682	27.8896	1131.8333
0.7957	176.9231	2300	0.5809	33.3448	25.3438	29.9738	29.9541	1123.3333
0.7957	178.8462	2325	0.5809	34.7749	25.4859	30.8064	30.966	1133.0
0.7957	180.7692	2350	0.5794	35.5149	25.6719	31.4748	31.7132	1125.5
0.7957	182.6923	2375	0.5798	37.8505	28.7851	33.1326	33.9326	1101.5
0.7957	184.6154	2400	0.5795	44.9411	31.929	39.3159	40.2195	814.8333
0.7957	186.5385	2425	0.5785	45.3771	32.5483	39.8906	40.6959	810.0
0.7957	188.4615	2450	0.5779	36.6623	27.845	32.6189	33.0543	1103.6667
0.7957	190.3846	2475	0.5774	36.6466	28.0217	32.5312	33.3439	843.3333
0.7481	192.3077	2500	0.5775	36.3607	27.8503	32.5271	32.9644	1103.6667
0.7481	194.2308	2525	0.5775	36.8397	27.9327	32.6315	33.0669	1103.6667
0.7481	196.1538	2550	0.5776	36.7963	27.8825	33.0842	33.2128	1103.6667
0.7481	198.0769	2575	0.5774	36.7436	27.845	33.0878	33.219	1103.6667
0.7481	200.0	2600	0.5774	36.746	27.845	33.0926	33.2212	1103.6667

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 3.0.2
• Tokenizers 0.19.1