lukeleeai's picture
End of training
7068374 verified
metadata
base_model: meta-llama/Llama-2-7b-hf
tags:
  - generated_from_trainer
model-index:
  - name: sparse_llama_7b_hf_refined_web_70p_2024-03-25
    results: []

sparse_llama_7b_hf_refined_web_70p_2024-03-25

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 2.1856

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: 0
  • distributed_type: multi-GPU
  • num_devices: 4
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 32
  • total_eval_batch_size: 4
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 1000

Training results

Training Loss Epoch Step Validation Loss
2.3676 0.01 25 2.5933
2.3572 0.02 50 2.5793
2.4503 0.02 75 2.5568
2.3803 0.03 100 2.5265
2.4451 0.04 125 2.4951
2.2793 0.05 150 2.4778
2.2444 0.06 175 2.4667
2.406 0.06 200 2.4572
2.3583 0.07 225 2.4508
2.3262 0.08 250 2.4538
2.258 0.09 275 2.4476
2.2841 0.1 300 2.4456
2.3232 0.1 325 2.4379
2.2974 0.11 350 2.4353
2.2216 0.12 375 2.4379
2.3179 0.13 400 2.4340
2.3006 0.14 425 2.4333
2.2603 0.14 450 2.4333
2.3371 0.15 475 2.4384
2.3453 0.16 500 2.4328
2.254 0.17 525 2.4306
2.2423 0.18 550 2.4298
2.3666 0.18 575 2.4293
2.259 0.19 600 2.4298
2.2786 0.2 625 2.4290
2.3493 0.21 650 2.4275
2.2532 0.22 675 2.4255
2.2698 0.22 700 2.4233
2.2949 0.23 725 2.4277
2.1918 0.24 750 2.4268
2.2762 0.25 775 2.4243
2.3221 0.26 800 2.4256
2.278 0.26 825 2.4273
2.2406 0.27 850 2.4223
2.2466 0.28 875 2.4252
2.2199 0.29 900 2.4247
2.4064 0.3 925 2.4259
2.3672 0.3 950 2.4237
2.3096 0.31 975 2.4226
2.1979 0.32 1000 2.4257

Framework versions

  • Transformers 4.36.2
  • Pytorch 2.1.1+cu121
  • Datasets 2.15.0
  • Tokenizers 0.15.2