prec_240730

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the perception_240723 dataset. It achieves the following results on the evaluation set:

• Loss: 1.1269

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: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• total_eval_batch_size: 2
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
1.1325	0.2222	100	1.0773
1.1674	0.4444	200	1.1166
1.0636	0.6667	300	1.0837
1.0874	0.8889	400	1.0411
0.5615	1.1111	500	1.0647
0.539	1.3333	600	1.0291
0.5426	1.5556	700	1.0026
0.4856	1.7778	800	0.9768
0.5153	2.0	900	0.9532
0.111	2.2222	1000	1.1014
0.1255	2.4444	1100	1.1229
0.0899	2.6667	1200	1.1248
0.1131	2.8889	1300	1.1272

Framework versions

• Transformers 4.43.3
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1