t5-v1_1-large-gramatika-final-e8-b16

This model is a fine-tuned version of google/t5-v1_1-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1652
• Rouge1: 43.5873
• Rouge2: 34.5612
• Rougel: 43.3549
• Rougelsum: 43.3701
• Gen Len: 18.9261

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: 0.001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adafactor
• lr_scheduler_type: linear
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
1.021	0.37	300	0.3972	34.0058	21.9508	32.8363	32.8413	18.7689
0.4571	0.73	600	0.3349	33.8082	23.1941	33.2259	33.2092	18.9198
0.3649	1.1	900	0.2819	37.177	26.2961	36.1081	36.1	18.9251
0.2812	1.46	1200	0.2162	41.57	31.5746	41.2196	41.2329	18.9303
0.2348	1.83	1500	0.1934	42.498	32.9335	42.2362	42.2357	18.9313
0.1883	2.2	1800	0.1900	42.8109	33.4358	42.5931	42.6018	18.9308
0.1607	2.56	2100	0.1790	42.8463	33.5315	42.6271	42.6405	18.9329
0.1553	2.93	2400	0.1764	43.0997	34.0805	42.9099	42.9121	18.9313
0.1143	3.29	2700	0.1735	43.2635	34.1534	43.0452	43.0553	18.9334
0.1078	3.66	3000	0.1652	43.5873	34.5612	43.3549	43.3701	18.9261
0.1029	4.02	3300	0.1729	43.7576	35.0653	43.5811	43.5982	18.9292
0.0669	4.39	3600	0.1773	43.6613	34.8016	43.4742	43.4793	18.9266
0.0671	4.76	3900	0.1710	43.92	35.1941	43.7078	43.7307	18.9313
0.0584	5.12	4200	0.1883	44.0979	35.4031	43.9062	43.9378	18.9334
0.0408	5.49	4500	0.1946	44.1764	35.6678	44.0201	44.0305	18.9319
0.0405	5.85	4800	0.1901	44.2228	35.7124	44.0157	44.0352	18.9319
0.0311	6.22	5100	0.2098	44.4711	36.1284	44.3122	44.3301	18.9251
0.0239	6.59	5400	0.2087	44.4937	36.2131	44.3726	44.3922	18.9271
0.0241	6.95	5700	0.2111	44.48	36.0938	44.3196	44.3404	18.9256
0.0162	7.32	6000	0.2235	44.4869	36.1867	44.3554	44.3615	18.9251
0.0147	7.68	6300	0.2281	44.618	36.3734	44.5021	44.5151	18.9240

Framework versions

• Transformers 4.30.1
• Pytorch 1.11.0a0+b6df043
• Datasets 2.12.0
• Tokenizers 0.13.3