skilltext

This model is a fine-tuned version of ai-forever/ruT5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0396
• Rouge1: 35.5496
• Rouge2: 22.9927
• Rougel: 33.7986
• Rougelsum: 33.9427
• Bleu: 3.0002
• Gen Len: 18.7273

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: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bleu	Gen Len
No log	0.5882	50	2.0006	22.8478	9.3528	21.5245	21.4195	1.3965	19.0
No log	1.1765	100	1.5029	26.0894	12.5184	22.7242	22.8568	1.7386	18.9545
No log	1.7647	150	1.4072	24.1385	9.8714	22.0278	22.0679	2.009	18.9545
No log	2.3529	200	1.3292	27.642	12.2998	26.3455	25.9994	1.2632	18.7727
No log	2.9412	250	1.2788	32.096	12.3806	30.9883	30.6962	1.6429	18.7273
No log	3.5294	300	1.1847	31.8602	21.2094	31.1454	30.9145	1.5913	18.8636
No log	4.1176	350	1.2193	22.6777	11.7225	22.1941	22.1638	1.4306	18.7727
No log	4.7059	400	1.1527	23.4161	11.2979	22.9918	23.0266	1.7552	18.8636
No log	5.2941	450	1.1200	28.9205	15.5233	27.153	27.2644	1.8557	18.7273
2.1495	5.8824	500	1.1426	28.2199	13.8386	26.9115	26.5472	2.3855	18.7273
2.1495	6.4706	550	1.1053	32.432	18.9395	30.9397	31.1198	2.2867	18.7727
2.1495	7.0588	600	1.0777	38.285	23.5443	35.0994	35.3165	2.6353	18.7727
2.1495	7.6471	650	1.0900	38.5934	21.6941	36.5629	36.9151	2.2212	18.7727
2.1495	8.2353	700	1.0931	41.2586	27.5923	40.1612	40.1672	2.5568	18.8182
2.1495	8.8235	750	1.0691	38.3785	25.0231	38.453	38.5248	2.4491	18.7273
2.1495	9.4118	800	1.0627	36.3073	20.703	35.2405	35.3787	2.3678	18.8636
2.1495	10.0	850	1.0528	39.1894	24.8355	39.3713	39.483	1.9687	18.8636
2.1495	10.5882	900	1.0628	40.0052	23.746	38.8726	39.077	2.0485	18.8636
2.1495	11.1765	950	1.0371	34.4982	23.4663	34.1685	34.1247	2.0922	18.8636
1.046	11.7647	1000	1.0368	38.0619	19.7898	36.4367	36.8115	2.3387	18.8636
1.046	12.3529	1050	1.0427	38.9055	25.1615	38.8253	38.9385	2.5522	18.8182
1.046	12.9412	1100	1.0255	36.5256	21.2328	34.8816	35.2236	2.4057	18.8182
1.046	13.5294	1150	1.0237	36.0048	25.3977	35.9471	35.9807	2.4804	18.8182
1.046	14.1176	1200	0.9918	32.6697	21.3968	30.8639	31.0221	2.4669	18.7727
1.046	14.7059	1250	1.0598	37.7878	20.6971	36.6794	36.7289	2.5767	18.7727
1.046	15.2941	1300	1.0130	34.549	24.4177	34.0376	34.1226	2.1773	18.8182
1.046	15.8824	1350	1.0256	32.774	19.6047	31.6125	31.9067	2.0504	18.7727
1.046	16.4706	1400	1.0232	31.4885	18.4703	30.0937	30.5529	2.514	18.8182
1.046	17.0588	1450	1.0210	33.4684	20.7982	31.7789	32.0023	2.4881	18.7273
0.7674	17.6471	1500	1.0419	37.4914	20.9444	35.0519	35.2368	3.0058	18.7727
0.7674	18.2353	1550	1.0328	36.5606	21.0215	35.2548	35.4748	2.7878	18.7273
0.7674	18.8235	1600	1.0376	31.3516	18.5826	29.6759	29.8435	2.3192	18.8182
0.7674	19.4118	1650	1.0414	37.4725	22.3216	35.6306	35.7383	2.477	18.8182
0.7674	20.0	1700	1.0513	39.5759	23.2665	39.2332	39.3667	2.4322	18.7273
0.7674	20.5882	1750	1.0518	36.1526	23.8263	34.5677	34.6173	2.8518	18.7727
0.7674	21.1765	1800	1.0446	41.5192	23.3064	39.3799	39.6548	3.0326	18.8182
0.7674	21.7647	1850	1.0150	40.5093	21.8683	38.2773	38.6063	2.6653	18.8636
0.7674	22.3529	1900	1.0364	34.2216	20.2095	32.5945	32.6999	2.6078	18.8182
0.7674	22.9412	1950	1.0148	39.8173	20.6247	37.2954	37.6752	3.0336	18.8636
0.6485	23.5294	2000	1.0429	40.2889	21.1598	37.7657	38.0596	2.9108	18.8182
0.6485	24.1176	2050	1.0423	39.2679	20.8842	36.7395	36.9295	2.845	18.8636
0.6485	24.7059	2100	1.0358	39.086	20.7799	36.2138	36.3741	2.9429	18.8182
0.6485	25.2941	2150	1.0219	38.754	22.4097	36.9752	37.121	2.831	18.8182
0.6485	25.8824	2200	1.0450	38.3531	22.3593	36.4439	36.6304	2.9804	18.7727
0.6485	26.4706	2250	1.0482	40.6921	23.617	39.298	39.5895	3.0971	18.7727
0.6485	27.0588	2300	1.0495	39.6761	22.7969	37.0805	37.4949	3.2639	18.7727
0.6485	27.6471	2350	1.0412	40.8199	23.7109	38.9222	39.2493	3.0267	18.7273
0.6485	28.2353	2400	1.0453	39.9504	23.888	38.0725	38.3121	3.2191	18.7727
0.6485	28.8235	2450	1.0400	36.205	23.1356	34.6087	34.6263	3.028	18.7727
0.5501	29.4118	2500	1.0402	35.033	22.2393	33.3754	33.4477	3.0299	18.7273
0.5501	30.0	2550	1.0396	35.5496	22.9927	33.7986	33.9427	3.0002	18.7273

Framework versions

• Transformers 4.40.0
• Pytorch 2.2.2
• Datasets 2.12.0
• Tokenizers 0.19.1