llama-68m-finetuned-wikitext2

This model is a fine-tuned version of JackFram/llama-68m on the None dataset. It achieves the following results on the evaluation set:

• Loss: 5.9290

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: 0.0003
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	175	3.1817
No log	2.0	350	3.2699
2.5169	3.0	525	3.4555
2.5169	4.0	700	3.6923
2.5169	5.0	875	3.9542
1.5624	6.0	1050	4.2089
1.5624	7.0	1225	4.4777
1.5624	8.0	1400	4.7426
0.841	9.0	1575	4.9898
0.841	10.0	1750	5.2385
0.841	11.0	1925	5.4648
0.3788	12.0	2100	5.6607
0.3788	13.0	2275	5.8074
0.3788	14.0	2450	5.8933
0.1474	15.0	2625	5.9290

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0