Baby-Llama-58M-RUN3_5

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 5.2656

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: 0.00025
• train_batch_size: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 50
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
287.9659	1.0	12	256.0041
230.7873	2.0	24	212.6014
207.1002	3.0	36	180.9384
121.5561	4.0	48	107.3193
81.2108	5.0	60	71.6529
45.9781	6.0	72	40.4501
24.5986	7.0	84	22.4212
15.2205	8.0	96	13.7469
10.1247	9.0	108	9.8119
7.975	10.0	120	7.8583
6.7087	11.0	132	7.0360
6.1988	12.0	144	6.4104
5.6752	13.0	156	6.1222
5.5155	14.0	168	5.8179
4.7754	15.0	180	5.5676
4.816	16.0	192	5.4583
4.817	17.0	204	5.3641
4.6966	18.0	216	5.3147
4.8322	19.0	228	5.2867
4.4875	20.0	240	5.2656

Framework versions

• Transformers 4.39.1
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0