Hugging Face's logo

saiteki-kai
/
QA-Llama-3.1

Text Classification
Transformers
Safetensors
llama
multi-label
question-answering
Generated from Trainer
Eval Results
text-generation-inference
Model card Files Community
QA-Llama-3.1 / README.md
saiteki-kai's picture
saiteki-kai
End of training
3d36d37 verified
preview code
|
raw
history blame contribute delete
3.13 kB
metadata
library_name: transformers
license: llama3.1
base_model: meta-llama/Llama-3.1-8B-Instruct
tags:
  - multi-label
  - question-answering
  - text-classification
  - generated_from_trainer
datasets:
  - beavertails
metrics:
  - accuracy
model-index:
  - name: QA-Llama-3.1-4155
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: saiteki-kai/BeaverTails-it
          type: beavertails
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.6964434241607612

QA-Llama-3.1-4155

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on the saiteki-kai/BeaverTails-it dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0781
  • Accuracy: 0.6964
  • Macro F1: 0.6445
  • Macro Precision: 0.7365
  • Macro Recall: 0.5970
  • Micro F1: 0.7539
  • Micro Precision: 0.8036
  • Micro Recall: 0.7100
  • Flagged/accuracy: 0.8560
  • Flagged/precision: 0.9050
  • Flagged/recall: 0.8283
  • Flagged/f1: 0.8649

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: 2.8362564501611134e-07
  • train_batch_size: 16
  • eval_batch_size: 32
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 4
  • total_train_batch_size: 64
  • total_eval_batch_size: 128
  • optimizer: Use 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
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss Accuracy Macro F1 Macro Precision Macro Recall Micro F1 Micro Precision Micro Recall Flagged/accuracy Flagged/precision Flagged/recall Flagged/f1
0.0688 1.0 8454 0.0799 0.6891 0.6367 0.7276 0.5931 0.7464 0.8015 0.6984 0.8491 0.8948 0.8260 0.8590
0.0745 2.0 16908 0.0777 0.6956 0.6295 0.7647 0.5680 0.7503 0.8171 0.6935 0.8532 0.9108 0.8160 0.8608
0.06 3.0 25362 0.0781 0.6965 0.6444 0.7361 0.5968 0.7539 0.8035 0.7100 0.8561 0.9050 0.8284 0.8650

Framework versions

  • Transformers 4.51.3
  • Pytorch 2.7.0+cu118
  • Datasets 3.5.1
  • Tokenizers 0.21.1