gpt2_sm_gen1

This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.8305
• Accuracy: 0.798
• Precision: 0.6346
• Recall: 0.1528
• F1: 0.2463
• D-index: 1.4919

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 8000
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	D-index
No log	1.0	250	1.2636	0.706	0.25	0.1806	0.2097	1.3706
1.4845	2.0	500	0.7018	0.734	0.25	0.1157	0.1582	1.3887
1.4845	3.0	750	0.6251	0.784	0.5	0.0093	0.0182	1.4233
0.5185	4.0	1000	0.5824	0.786	0.6	0.0278	0.0531	1.4326
0.5185	5.0	1250	0.5811	0.789	0.6667	0.0463	0.0866	1.4432
0.4299	6.0	1500	0.6123	0.793	0.8	0.0556	0.1039	1.4520
0.4299	7.0	1750	0.5759	0.784	0.5	0.1389	0.2174	1.4677
0.3603	8.0	2000	0.7418	0.79	0.6154	0.0741	0.1322	1.4541
0.3603	9.0	2250	0.6740	0.766	0.4392	0.3009	0.3571	1.4963
0.2662	10.0	2500	0.8520	0.789	0.5385	0.1620	0.2491	1.4824
0.2662	11.0	2750	1.4823	0.79	0.6154	0.0741	0.1322	1.4541
0.1838	12.0	3000	1.2440	0.789	0.5203	0.2963	0.3776	1.5267
0.1838	13.0	3250	1.3491	0.781	0.4854	0.2315	0.3135	1.4944
0.1362	14.0	3500	1.6458	0.786	0.5093	0.2546	0.3395	1.5089
0.1362	15.0	3750	1.7559	0.794	0.5532	0.2407	0.3355	1.5155
0.1176	16.0	4000	2.3472	0.801	0.7073	0.1343	0.2257	1.4899
0.1176	17.0	4250	2.1587	0.793	0.5849	0.1435	0.2305	1.4818
0.0812	18.0	4500	2.1713	0.793	0.5584	0.1991	0.2935	1.5003
0.0812	19.0	4750	2.1631	0.793	0.5413	0.2731	0.3631	1.5247
0.065	20.0	5000	2.8305	0.798	0.6346	0.1528	0.2463	1.4919

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3