llama_lora_20k_alpaca_begining

This model is a fine-tuned version of mifeng09/my_final_llama_model_v2_add_wiki_fix_resume on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 3.7591

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: 4
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 8
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • lr_scheduler_warmup_steps: 101
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss
4.2374 0.1484 50 4.0144
4.3818 0.2967 100 3.9270
3.944 0.4451 150 3.8766
3.3377 0.5935 200 3.8441
3.3489 0.7418 250 3.8238
3.5876 0.8902 300 3.8094
3.7506 1.0386 350 3.7967
3.9164 1.1869 400 3.7891
4.4806 1.3353 450 3.7829
4.078 1.4837 500 3.7769
3.8764 1.6320 550 3.7716
3.7083 1.7804 600 3.7681
3.5106 1.9288 650 3.7660
3.8436 2.0772 700 3.7631
3.8274 2.2255 750 3.7621
4.0989 2.3739 800 3.7603
3.7672 2.5223 850 3.7594
3.7825 2.6706 900 3.7592
3.8184 2.8190 950 3.7591
4.0853 2.9674 1000 3.7591

Framework versions

  • PEFT 0.18.0
  • Transformers 4.57.3
  • Pytorch 2.9.0+cpu
  • Datasets 4.0.0
  • Tokenizers 0.22.1
Downloads last month
255
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for mifeng09/llama_lora_20k_alpaca_begining

Evaluation results