finetune_mc_20

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: 4.1545

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
0.9427	1.0	70	1.9064
0.2001	2.0	140	2.6308
0.1129	3.0	210	2.9004
0.0803	4.0	280	3.0336
0.0665	5.0	350	3.0398
0.0396	6.0	420	3.0769
0.0429	7.0	490	3.1504
0.0318	8.0	560	3.2690
0.03	9.0	630	3.4818
0.0258	10.0	700	3.6011
0.0247	11.0	770	3.7578
0.0287	12.0	840	3.8834
0.0257	13.0	910	3.9492
0.0267	14.0	980	3.9646
0.0205	15.0	1050	4.0157
0.0202	16.0	1120	4.0518
0.0222	17.0	1190	4.0854
0.0203	18.0	1260	4.1229
0.0231	19.0	1330	4.1433
0.0199	20.0	1400	4.1545

Framework versions

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