llama_finetune_cs_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: 3.0872

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
1.6881	1.0	150	1.6028
1.3532	2.0	300	1.6693
0.9084	3.0	450	1.8783
0.4352	4.0	600	2.0825
0.374	5.0	750	2.1988
0.2596	6.0	900	2.2869
0.2316	7.0	1050	2.3945
0.1754	8.0	1200	2.5528
0.1698	9.0	1350	2.6226
0.1371	10.0	1500	2.5997
0.1702	11.0	1650	2.7153
0.1391	12.0	1800	2.6888
0.1894	13.0	1950	2.7324
0.1504	14.0	2100	2.8534
0.132	15.0	2250	2.8804
0.1585	16.0	2400	2.9148
0.1497	17.0	2550	2.9550
0.1099	18.0	2700	3.0002
0.1321	19.0	2850	3.0532
0.11	20.0	3000	3.0872

Framework versions

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