Hugging Face's logo

lenashamseldin
/
git-base-floors615images

Image-Text-to-Text
Transformers
TensorBoard
Safetensors
git
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
git-base-floors615images / README.md
lenashamseldin's picture
lenashamseldin
End of training
c659b21 verified
preview code
|
raw
history blame
5.73 kB
metadata
license: mit
base_model: microsoft/git-base
tags:
  - generated_from_trainer
model-index:
  - name: git-base-floors615images
    results: []

git-base-floors615images

This model is a fine-tuned version of microsoft/git-base on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0139
  • Wer Score: 0.5837

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: 4
  • eval_batch_size: 4
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 50
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer Score
6.955 0.72 50 4.0586 1.6636
1.911 1.44 100 0.2120 3.1636
0.0574 2.16 150 0.0130 4.5642
0.0118 2.88 200 0.0126 2.8213
0.0099 3.6 250 0.0095 2.2693
0.0082 4.32 300 0.0089 5.0186
0.0077 5.04 350 0.0088 5.3560
0.0068 5.76 400 0.0091 2.8026
0.0064 6.47 450 0.0082 3.4589
0.0063 7.19 500 0.0086 4.6660
0.006 7.91 550 0.0082 4.8546
0.0056 8.63 600 0.0085 3.1033
0.0062 9.35 650 0.0087 4.0475
0.0057 10.07 700 0.0086 5.4452
0.0054 10.79 750 0.0084 4.8614
0.0056 11.51 800 0.0084 5.4922
0.005 12.23 850 0.0084 5.4324
0.0054 12.95 900 0.0083 5.6768
0.0051 13.67 950 0.0086 5.1234
0.0052 14.39 1000 0.0091 6.0069
0.0053 15.11 1050 0.0085 5.0867
0.0049 15.83 1100 0.0092 5.4711
0.0051 16.55 1150 0.0087 5.7106
0.0048 17.27 1200 0.0086 4.0867
0.0047 17.99 1250 0.0087 5.0113
0.0048 18.71 1300 0.0088 5.9662
0.0046 19.42 1350 0.0091 4.5553
0.005 20.14 1400 0.0088 6.0235
0.0054 20.86 1450 0.0085 6.0230
0.005 21.58 1500 0.0088 6.0230
0.0052 22.3 1550 0.0084 6.0186
0.0049 23.02 1600 0.0093 3.9496
0.005 23.74 1650 0.0089 3.9275
0.0046 24.46 1700 0.0085 4.7311
0.0049 25.18 1750 0.0089 5.2463
0.0045 25.9 1800 0.0088 3.8296
0.0044 26.62 1850 0.0088 4.5010
0.0043 27.34 1900 0.0090 2.7282
0.0045 28.06 1950 0.0089 3.4383
0.0038 28.78 2000 0.0094 1.3452
0.0042 29.5 2050 0.0089 3.1126
0.0039 30.22 2100 0.0092 2.4177
0.0042 30.94 2150 0.0092 2.7204
0.0037 31.65 2200 0.0092 2.6229
0.0039 32.37 2250 0.0095 2.3673
0.0038 33.09 2300 0.0102 1.4265
0.0034 33.81 2350 0.0098 2.4819
0.0032 34.53 2400 0.0110 1.9114
0.0034 35.25 2450 0.0111 1.4285
0.0031 35.97 2500 0.0108 2.0573
0.0027 36.69 2550 0.0109 1.6538
0.0028 37.41 2600 0.0114 1.9789
0.0029 38.13 2650 0.0108 2.1675
0.0026 38.85 2700 0.0112 2.3497
0.0026 39.57 2750 0.0124 2.1665
0.0026 40.29 2800 0.0119 1.4652
0.0024 41.01 2850 0.0119 2.0720
0.0022 41.73 2900 0.0122 1.2238
0.0022 42.45 2950 0.0128 1.5343
0.0021 43.17 3000 0.0121 1.3310
0.0019 43.88 3050 0.0131 1.3859
0.0018 44.6 3100 0.0131 0.8487
0.0017 45.32 3150 0.0133 0.9540
0.0016 46.04 3200 0.0131 1.0446
0.0014 46.76 3250 0.0135 0.6459
0.0014 47.48 3300 0.0137 0.6562
0.0013 48.2 3350 0.0135 0.5901
0.0012 48.92 3400 0.0139 0.5857
0.0012 49.64 3450 0.0139 0.5837

Framework versions

  • Transformers 4.38.2
  • Pytorch 2.0.1
  • Datasets 2.18.0
  • Tokenizers 0.15.2