Edit model card

Llama-31-8B_task-3_60-samples_config-2

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the GaetanMichelet/chat-60_ft_task-3 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4068

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: 1
  • eval_batch_size: 1
  • seed: 42
  • distributed_type: multi-GPU
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss
2.4485 0.6957 2 2.4418
2.3693 1.7391 5 2.0528
1.7056 2.7826 8 1.2637
1.1524 3.8261 11 0.6909
0.568 4.8696 14 0.5483
0.5339 5.9130 17 0.4959
0.4507 6.9565 20 0.4587
0.3073 8.0 23 0.4425
0.4037 8.6957 25 0.4284
0.3191 9.7391 28 0.4162
0.3306 10.7826 31 0.4068
0.2342 11.8261 34 0.4073
0.2395 12.8696 37 0.4246
0.1509 13.9130 40 0.4373
0.2333 14.9565 43 0.4344
0.1348 16.0 46 0.4878
0.1198 16.6957 48 0.5004
0.0922 17.7391 51 0.5110

Framework versions

  • PEFT 0.12.0
  • Transformers 4.44.0
  • Pytorch 2.1.2+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
6
Inference API
Unable to determine this model’s pipeline type. Check the docs .

Model tree for GaetanMichelet/Llama-31-8B_task-3_60-samples_config-2

Adapter
(452)
this model

Collections including GaetanMichelet/Llama-31-8B_task-3_60-samples_config-2