llama_finetune_mc_20

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.4015

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
0.9393	1.0	70	1.8240
0.2414	2.0	140	2.5937
0.126	3.0	210	2.8429
0.0836	4.0	280	2.9219
0.0872	5.0	350	3.2516
0.0575	6.0	420	3.1180
0.0482	7.0	490	3.4019
0.0356	8.0	560	3.3709
0.0305	9.0	630	3.5186
0.0272	10.0	700	3.8218
0.0243	11.0	770	3.7827
0.0312	12.0	840	3.9016
0.0259	13.0	910	4.0432
0.027	14.0	980	4.1255
0.0205	15.0	1050	4.1950
0.0199	16.0	1120	4.2793
0.0219	17.0	1190	4.3363
0.0197	18.0	1260	4.3627
0.0218	19.0	1330	4.3868
0.0201	20.0	1400	4.4015

Framework versions

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