finetune_race_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: 3.4669

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
1.2823	1.0	150	1.4463
0.7207	2.0	300	1.5249
0.4983	3.0	450	1.6580
0.5956	4.0	600	1.8347
0.1806	5.0	750	2.0078
0.1439	6.0	900	2.2482
0.1399	7.0	1050	2.3858
0.0826	8.0	1200	2.4455
0.0878	9.0	1350	2.5889
0.0564	10.0	1500	2.7918
0.0641	11.0	1650	2.7968
0.0522	12.0	1800	2.7144
0.0491	13.0	1950	2.9539
0.0436	14.0	2100	2.9602
0.0431	15.0	2250	3.1613
0.0449	16.0	2400	3.2197
0.0516	17.0	2550	3.3331
0.0432	18.0	2700	3.4157
0.0448	19.0	2850	3.4384
0.0416	20.0	3000	3.4669

Framework versions

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