YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

<midm λͺ¨λΈμ˜ μ„±λŠ₯을 높이기 μœ„ν•΄ 좔가적인 λ―Έμ„ΈνŠœλ‹μ„ μ§„ν–‰ν•˜μ˜€λ‹€.>

#첫번째 μ‹œλ„

  • μ‹€ν—˜ λ‚΄μš©: 주어진 μ½”λ“œ κ·ΈλŒ€λ‘œ 일단 μ‹€ν–‰ν•˜μ—¬ μ„±λŠ₯을 ν™•μΈν–ˆλ‹€.

  • μ‹€ν—˜ κ²°κ³Ό: TP: 412 TN: 439 PP: 69 PN: 80 Accuracy: 0.851

#λ‘λ²ˆμ§Έ μ‹œλ„

  • μ‹€ν—˜ λ‚΄μš©:
  1. 이전 ν”„λ‘œμ νŠΈλ₯Ό μ§„ν–‰ν•˜λ‹€λ³΄λ©΄ 보톡 epoch 수λ₯Ό 늘리면 μ„±λŠ₯이 쒋아짐을 확인할 수 μžˆμ—ˆλ‹€. 더 λ§Žμ€ 데이터λ₯Ό ν•™μŠ΅ν•΄λ³΄κΈ° μœ„ν•΄ μ¦κ°€μ‹œμΌ°λ‹€. (num_train_epochs: 1-> 1000둜 λ³€κ²½)
  2. max_grad_norm이 크면 λ°œμ‚°μ˜ μœ„ν—˜μ΄ 크고 수렴이 μ–΄λ €μ›Œμ§€λ©°, 정확도가 κ°μ†Œλ  수 있기 λ•Œλ¬Έμ— μ–΄λŠμ •λ„ 값을 μ€„μ—¬μ£Όμ—ˆλ‹€. (max_grad_norm: 0.3-> 0.1둜 λ³€κ²½)
  • μ‹€ν—˜ κ²°κ³Ό: TP: 442 TN: 465 PP: 41 PN: 52 Accuracy: 0.907

<정확도가 μ–΄λŠμ •λ„λŠ” κ°œμ„ λ˜μ—ˆμŒμ„ ν™•μΈν•˜μ˜€λ‹€.>

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference API
Unable to determine this model's library. Check the docs .