Edit model card

Horoscope Model Base - finetuned on weather

This model is a fine-tuned version of iulik-pisik/horoscope_model_base on the Vreme ProTV dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2407
  • Wer: 11.8989

Model description

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 1e-05
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 4000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.0227 6.02 1000 0.1744 13.2758
0.001 12.05 2000 0.2217 12.0222
0.0004 18.07 3000 0.2365 11.9194
0.0003 24.1 4000 0.2407 11.8989

Framework versions

  • Transformers 4.39.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
0
Safetensors
Model size
72.6M params
Tensor type
F32
·
Inference API
or
This model can be loaded on Inference API (serverless).

Finetuned from

Dataset used to train iulik-pisik/horoscop_vreme_base

Evaluation results