topicalchat-multiturn

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

Loss: 2.5260

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: 64
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 100.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	73	4.2992
No log	2.0	146	3.4433
No log	3.0	219	3.1606
No log	4.0	292	3.0366
No log	5.0	365	2.9679
No log	6.0	438	2.9131
4.1401	7.0	511	2.8752
4.1401	8.0	584	2.8391
4.1401	9.0	657	2.8118
4.1401	10.0	730	2.7871
4.1401	11.0	803	2.7659
4.1401	12.0	876	2.7489
4.1401	13.0	949	2.7331
2.9768	14.0	1022	2.7196
2.9768	15.0	1095	2.7071
2.9768	16.0	1168	2.6940
2.9768	17.0	1241	2.6854
2.9768	18.0	1314	2.6728
2.9768	19.0	1387	2.6647
2.9768	20.0	1460	2.6562
2.7864	21.0	1533	2.6482
2.7864	22.0	1606	2.6439
2.7864	23.0	1679	2.6326
2.7864	24.0	1752	2.6107
2.7864	25.0	1825	2.6043
2.7864	26.0	1898	2.5970
2.7864	27.0	1971	2.5908
2.6568	28.0	2044	2.5862
2.6568	29.0	2117	2.5828
2.6568	30.0	2190	2.5765
2.6568	31.0	2263	2.5742
2.6568	32.0	2336	2.5682
2.6568	33.0	2409	2.5656
2.6568	34.0	2482	2.5614
2.5489	35.0	2555	2.5605
2.5489	36.0	2628	2.5552
2.5489	37.0	2701	2.5541
2.5489	38.0	2774	2.5494
2.5489	39.0	2847	2.5491
2.5489	40.0	2920	2.5455
2.5489	41.0	2993	2.5452
2.475	42.0	3066	2.5433
2.475	43.0	3139	2.5397
2.475	44.0	3212	2.5386
2.475	45.0	3285	2.5400
2.475	46.0	3358	2.5339
2.475	47.0	3431	2.5327
2.4144	48.0	3504	2.5327
2.4144	49.0	3577	2.5312
2.4144	50.0	3650	2.5338
2.4144	51.0	3723	2.5314
2.4144	52.0	3796	2.5309
2.4144	53.0	3869	2.5289
2.4144	54.0	3942	2.5290
2.3642	55.0	4015	2.5270
2.3642	56.0	4088	2.5270
2.3642	57.0	4161	2.5263
2.3642	58.0	4234	2.5267
2.3642	59.0	4307	2.5273
2.3642	60.0	4380	2.5258
2.3642	61.0	4453	2.5253
2.3216	62.0	4526	2.5244
2.3216	63.0	4599	2.5256
2.3216	64.0	4672	2.5227
2.3216	65.0	4745	2.5241
2.3216	66.0	4818	2.5244
2.3216	67.0	4891	2.5236
2.3216	68.0	4964	2.5251
2.2879	69.0	5037	2.5231
2.2879	70.0	5110	2.5254
2.2879	71.0	5183	2.5242
2.2879	72.0	5256	2.5254
2.2879	73.0	5329	2.5253
2.2879	74.0	5402	2.5228
2.2879	75.0	5475	2.5247
2.261	76.0	5548	2.5243
2.261	77.0	5621	2.5247
2.261	78.0	5694	2.5250
2.261	79.0	5767	2.5248
2.261	80.0	5840	2.5236
2.261	81.0	5913	2.5264
2.261	82.0	5986	2.5249
2.2396	83.0	6059	2.5256
2.2396	84.0	6132	2.5267
2.2396	85.0	6205	2.5258
2.2396	86.0	6278	2.5242
2.2396	87.0	6351	2.5233
2.2396	88.0	6424	2.5249
2.2396	89.0	6497	2.5253
2.2238	90.0	6570	2.5252
2.2238	91.0	6643	2.5255
2.2238	92.0	6716	2.5263
2.2238	93.0	6789	2.5261
2.2238	94.0	6862	2.5257
2.2238	95.0	6935	2.5253
2.213	96.0	7008	2.5267
2.213	97.0	7081	2.5258
2.213	98.0	7154	2.5258
2.213	99.0	7227	2.5259
2.213	100.0	7300	2.5260

Framework versions

Transformers 4.8.1
Pytorch 1.8.1+cu111
Datasets 1.8.0
Tokenizers 0.10.3