Modelpublic

vaitech/open-hermes-erotic-story-finetune

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face

0likes5downloads

Model Card

open-hermes-erotic-story-finetune

This model is a fine-tuned version of teknium/OpenHermes-2.5-Mistral-7B on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.1465

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: 2.5e-05
—trainbatchsize: 2
—evalbatchsize: 8
—seed: 42
—gradientaccumulationsteps: 4
—totaltrainbatch_size: 8
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—lrschedulerwarmup_steps: 5
—training_steps: 1500

Training results

Training Loss	Epoch	Step	Validation Loss
1.1158	0.1056	50	0.2470
0.2078	0.2112	100	0.1934
0.1866	0.3168	150	0.1867
0.1804	0.4224	200	0.1719
0.158	0.5280	250	0.1515
0.15	0.6336	300	0.1507
0.1489	0.7392	350	0.1499
0.1464	0.8448	400	0.1494
0.1485	0.9504	450	0.1487
0.1485	1.0560	500	0.1485
0.1459	1.1616	550	0.1484
0.1444	1.2672	600	0.1481
0.1469	1.3728	650	0.1477
0.148	1.4784	700	0.1477
0.1438	1.5839	750	0.1475
0.1454	1.6895	800	0.1472
0.149	1.7951	850	0.1471
0.1395	1.9007	900	0.1469
0.1416	2.0063	950	0.1468
0.1443	2.1119	1000	0.1468
0.1445	2.2175	1050	0.1467
0.143	2.3231	1100	0.1468
0.1416	2.4287	1150	0.1468
0.1418	2.5343	1200	0.1466
0.1384	2.6399	1250	0.1466
0.1382	2.7455	1300	0.1466
0.1423	2.8511	1350	0.1466
0.1414	2.9567	1400	0.1465
0.1398	3.0623	1450	0.1465
0.136	3.1679	1500	0.1465

Framework versions

—PEFT 0.11.2.dev0
—Transformers 4.42.4
—Pytorch 2.3.1+cu121
—Datasets 2.20.0
—Tokenizers 0.19.1