Hugging Face's logo

tyzhu
/
lmind_hotpot_train8000_eval7405_v1_qa_5e-4_lora2

PEFT
Safetensors
Generated from Trainer
Eval Results
Model card Files Community
lmind_hotpot_train8000_eval7405_v1_qa_5e-4_lora2 / README.md
tyzhu's picture
tyzhu
End of training
9aff838 verified
preview code
|
raw
history blame
3.17 kB
metadata
license: llama2
base_model: meta-llama/Llama-2-7b-hf
tags:
  - generated_from_trainer
datasets:
  - tyzhu/lmind_hotpot_train8000_eval7405_v1_qa
metrics:
  - accuracy
model-index:
  - name: lmind_hotpot_train8000_eval7405_v1_qa_5e-4_lora2
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: tyzhu/lmind_hotpot_train8000_eval7405_v1_qa
          type: tyzhu/lmind_hotpot_train8000_eval7405_v1_qa
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5813164556962025

lmind_hotpot_train8000_eval7405_v1_qa_5e-4_lora2

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on the tyzhu/lmind_hotpot_train8000_eval7405_v1_qa dataset. It achieves the following results on the evaluation set:

  • Loss: 2.9420
  • Accuracy: 0.5813

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: 0.0005
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 4
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 32
  • total_eval_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: constant
  • lr_scheduler_warmup_ratio: 0.05
  • num_epochs: 20.0

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.8732 1.0 250 2.0111 0.5939
1.6142 2.0 500 1.8443 0.6051
1.206 3.0 750 1.9818 0.6007
0.8693 4.0 1000 2.2100 0.5941
0.6023 5.0 1250 2.3756 0.5910
0.4717 6.0 1500 2.5421 0.5896
0.3938 7.0 1750 2.6587 0.5891
0.3697 8.0 2000 2.7532 0.5873
0.3617 9.0 2250 2.7664 0.5870
0.3607 10.0 2500 2.8514 0.5867
0.3414 11.0 2750 2.8932 0.5861
0.3439 12.0 3000 2.9545 0.5855
0.335 13.0 3250 2.8991 0.5843
0.3391 14.0 3500 2.8793 0.5840
0.328 15.0 3750 2.8954 0.5851
0.3351 16.0 4000 2.9140 0.5838
0.3252 17.0 4250 2.9297 0.5825
0.332 18.0 4500 2.9812 0.5834
0.324 19.0 4750 2.9823 0.5808
0.3329 20.0 5000 2.9420 0.5813

Framework versions

  • Transformers 4.34.0
  • Pytorch 2.1.0+cu121
  • Datasets 2.18.0
  • Tokenizers 0.14.1