flan-t5-large-extraction-all-dm_8000-ep10-nonstop

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

• Loss: 1.5334
• Hint Hit Num: 2.2682
• Hint Precision: 0.4261
• Num: 5.2682
• Gen Len: 18.776

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: 12
• eval_batch_size: 96
• seed: 1799
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Hint Hit Num	Hint Precision	Num	Gen Len
2.1988	0.3	200	1.5870	2.6768	0.4798	5.6224	18.9782
1.976	0.6	400	1.5613	2.5624	0.4663	5.495	18.9114
1.9367	0.9	600	1.5303	2.4822	0.4551	5.4574	18.9418
1.8626	1.2	800	1.5336	2.3896	0.4403	5.3966	18.9096
1.8278	1.5	1000	1.5110	2.5016	0.4514	5.5236	18.9486
1.8115	1.8	1200	1.5116	2.2886	0.4269	5.3196	18.9194
1.776	2.1	1400	1.5212	2.3278	0.4326	5.3394	18.8936
1.7332	2.4	1600	1.5172	2.2982	0.4323	5.2878	18.828
1.7543	2.7	1800	1.5003	2.473	0.4522	5.4414	18.9048
1.7212	3.0	2000	1.5051	2.3878	0.4389	5.4032	18.854
1.6915	3.3	2200	1.5083	2.3352	0.4347	5.3186	18.836
1.6808	3.6	2400	1.5065	2.3414	0.4367	5.321	18.8136
1.6812	3.9	2600	1.5047	2.3422	0.4376	5.3144	18.812
1.6408	4.2	2800	1.5158	2.3108	0.4297	5.33	18.8116
1.6266	4.5	3000	1.5086	2.2752	0.4227	5.329	18.8472
1.6144	4.8	3200	1.5120	2.2434	0.4192	5.283	18.8684
1.6164	5.1	3400	1.5135	2.3636	0.4356	5.3754	18.8526
1.5981	5.4	3600	1.5202	2.245	0.4201	5.2762	18.8574
1.5923	5.7	3800	1.5190	2.2462	0.4208	5.28	18.8358
1.5835	6.0	4000	1.5182	2.2812	0.4249	5.3042	18.8182
1.577	6.3	4200	1.5268	2.2928	0.4254	5.335	18.8268
1.5572	6.6	4400	1.5229	2.261	0.4237	5.276	18.7788
1.5522	6.9	4600	1.5153	2.3372	0.4323	5.3516	18.8326
1.5095	7.2	4800	1.5334	2.2108	0.4195	5.2086	18.7338
1.5568	7.5	5000	1.5243	2.302	0.4305	5.2964	18.7742
1.5373	7.8	5200	1.5277	2.2502	0.4204	5.2868	18.8176
1.5191	8.1	5400	1.5321	2.2716	0.4247	5.2856	18.7934
1.5261	8.4	5600	1.5300	2.2938	0.4273	5.3064	18.7828
1.5202	8.7	5800	1.5337	2.2744	0.4236	5.3086	18.8092
1.4942	9.0	6000	1.5351	2.2522	0.4239	5.257	18.7704
1.4816	9.3	6200	1.5349	2.2528	0.4247	5.2518	18.7682
1.5169	9.6	6400	1.5339	2.2698	0.4265	5.2646	18.7736
1.5007	9.9	6600	1.5334	2.269	0.4263	5.2664	18.776

Framework versions

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