finetune_arc_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: 2.8229

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
1.2708	1.0	150	1.2201
0.8854	2.0	300	1.2765
0.567	3.0	450	1.4685
0.3112	4.0	600	1.7395
0.1762	5.0	750	2.0026
0.1228	6.0	900	2.0326
0.1002	7.0	1050	2.1066
0.0931	8.0	1200	2.1262
0.1434	9.0	1350	2.2417
0.0746	10.0	1500	2.3327
0.069	11.0	1650	2.3327
0.0804	12.0	1800	2.5652
0.0586	13.0	1950	2.4866
0.0652	14.0	2100	2.5962
0.0471	15.0	2250	2.6461
0.054	16.0	2400	2.6890
0.0602	17.0	2550	2.7081
0.0562	18.0	2700	2.7800
0.064	19.0	2850	2.8103
0.0509	20.0	3000	2.8229

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0