Hugging Face's logo

xshubhamx
/
tiny-llama-lora-new

TensorBoard
Safetensors
Generated from Trainer
Model card Files Metrics Community
tiny-llama-lora-new / README.md
xshubhamx's picture
xshubhamx
Upload README.md with huggingface_hub
928bee3 verified
preview code
|
raw
history blame contribute delete
6.12 kB
metadata
license: apache-2.0
tags:
  - generated_from_trainer
base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
metrics:
  - accuracy
  - precision
  - recall
model-index:
  - name: tiny-llama-lora-new
    results: []

tiny-llama-lora-new

This model is a fine-tuned version of TinyLlama/TinyLlama-1.1B-Chat-v1.0 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2252
  • Accuracy: 0.8203
  • Precision: 0.8184
  • Recall: 0.8203
  • Precision Macro: 0.7732
  • Recall Macro: 0.7380
  • Macro Fpr: 0.0162
  • Weighted Fpr: 0.0154
  • Weighted Specificity: 0.9743
  • Macro Specificity: 0.9863
  • Weighted Sensitivity: 0.8203
  • Macro Sensitivity: 0.7380
  • F1 Micro: 0.8203
  • F1 Macro: 0.7435
  • F1 Weighted: 0.8173

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: 5e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 15

Training results

Training Loss Epoch Step Validation Loss Accuracy Precision Recall Precision Macro Recall Macro Macro Fpr Weighted Fpr Weighted Specificity Macro Specificity Weighted Sensitivity Macro Sensitivity F1 Micro F1 Macro F1 Weighted
No log 1.0 160 0.6615 0.8002 0.8040 0.8002 0.7266 0.6678 0.0182 0.0175 0.9726 0.9848 0.8002 0.6678 0.8002 0.6790 0.7959
No log 2.0 321 0.6996 0.8064 0.8110 0.8064 0.7448 0.7207 0.0177 0.0169 0.9737 0.9853 0.8064 0.7207 0.8064 0.7235 0.8039
No log 3.0 482 0.8202 0.8125 0.8119 0.8125 0.7577 0.7080 0.0171 0.0162 0.9711 0.9856 0.8125 0.7080 0.8125 0.7180 0.8085
0.2932 4.0 643 0.9493 0.8141 0.8204 0.8141 0.7593 0.7327 0.0166 0.0160 0.9744 0.9859 0.8141 0.7327 0.8141 0.7415 0.8154
0.2932 5.0 803 1.0610 0.8110 0.8110 0.8110 0.7596 0.7427 0.0172 0.0164 0.9738 0.9857 0.8110 0.7427 0.8110 0.7413 0.8087
0.2932 6.0 964 1.1362 0.8149 0.8160 0.8149 0.7731 0.7380 0.0167 0.0160 0.9741 0.9859 0.8149 0.7380 0.8149 0.7408 0.8128
0.0107 7.0 1125 1.1713 0.8102 0.8123 0.8102 0.7734 0.7310 0.0171 0.0165 0.9736 0.9856 0.8102 0.7310 0.8102 0.7343 0.8085
0.0107 8.0 1286 1.1786 0.8156 0.8141 0.8156 0.7656 0.7349 0.0166 0.0159 0.9740 0.9860 0.8156 0.7349 0.8156 0.7374 0.8128
0.0107 9.0 1446 1.1960 0.8187 0.8170 0.8187 0.7693 0.7368 0.0163 0.0156 0.9743 0.9862 0.8187 0.7368 0.8187 0.7400 0.8157
0.0016 10.0 1607 1.2049 0.8156 0.8150 0.8156 0.7659 0.7353 0.0166 0.0159 0.9741 0.9860 0.8156 0.7353 0.8156 0.7376 0.8131
0.0016 11.0 1768 1.2137 0.8156 0.8147 0.8156 0.7661 0.7353 0.0166 0.0159 0.9741 0.9860 0.8156 0.7353 0.8156 0.7377 0.8130
0.0016 12.0 1929 1.2158 0.8156 0.8145 0.8156 0.7664 0.7353 0.0166 0.0159 0.9739 0.9860 0.8156 0.7353 0.8156 0.7379 0.8129
0.0011 13.0 2089 1.2202 0.8187 0.8169 0.8187 0.7720 0.7372 0.0163 0.0156 0.9741 0.9862 0.8187 0.7372 0.8187 0.7425 0.8158
0.0011 14.0 2250 1.2229 0.8187 0.8169 0.8187 0.7720 0.7372 0.0163 0.0156 0.9741 0.9862 0.8187 0.7372 0.8187 0.7425 0.8158
0.0011 14.93 2400 1.2252 0.8203 0.8184 0.8203 0.7732 0.7380 0.0162 0.0154 0.9743 0.9863 0.8203 0.7380 0.8203 0.7435 0.8173

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.19.0
  • Tokenizers 0.15.1