finetune_mc_20_cot

This model is a fine-tuned version of lmsys/vicuna-7b-v1.5 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.0612

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: 0.0001
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 5
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.8917	1.0	70	1.7411
0.2135	2.0	140	2.3727
0.1702	3.0	210	2.5470
0.1234	4.0	280	2.8131
0.0843	5.0	350	3.0446
0.0478	6.0	420	3.1141
0.0473	7.0	490	3.2304
0.0544	8.0	560	3.3806
0.0289	9.0	630	3.4231
0.0202	10.0	700	3.5503
0.0214	11.0	770	3.5892
0.0261	12.0	840	3.7211
0.0232	13.0	910	3.8148
0.0242	14.0	980	3.8177
0.0198	15.0	1050	3.9079
0.018	16.0	1120	3.9320
0.0196	17.0	1190	3.9807
0.0179	18.0	1260	4.0263
0.0194	19.0	1330	4.0520
0.0169	20.0	1400	4.0612

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.0+cu121
• Datasets 2.13.1
• Tokenizers 0.14.1