Hugging Face's logo

yunaseo
/
emotions_flan_tf

PEFT
Safetensors
Generated from Trainer
Model card Files Community
emotions_flan_tf / README.md
yunaseo's picture
yunaseo
yunase/flant5_QLORA_model_emotion_detection
9a3d0ae verified
preview code
|
raw
history blame contribute delete
No virus
5.51 kB
metadata
license: apache-2.0
library_name: peft
tags:
  - generated_from_trainer
metrics:
  - accuracy
base_model: google/flan-t5-base
model-index:
  - name: emotions_flan_tf
    results: []

emotions_flan_tf

This model is a fine-tuned version of google/flan-t5-base on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4970
  • F1 Micro: 0.6980
  • F1 Macro: 0.6126
  • Accuracy: 0.2188

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: 0.0001
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss F1 Micro F1 Macro Accuracy
0.8313 0.21 20 0.7916 0.4215 0.1586 0.0123
0.7813 0.41 40 0.7840 0.4717 0.2286 0.0201
0.7702 0.62 60 0.7485 0.4935 0.2379 0.0971
0.7218 0.83 80 0.6327 0.6045 0.3866 0.1256
0.6401 1.03 100 0.5907 0.6269 0.4310 0.1586
0.5951 1.24 120 0.5668 0.6459 0.4981 0.1502
0.5686 1.45 140 0.5458 0.6593 0.5372 0.1683
0.5576 1.65 160 0.5332 0.6675 0.5403 0.1722
0.5465 1.86 180 0.5224 0.6734 0.5667 0.1812
0.5436 2.07 200 0.5164 0.6807 0.5751 0.1877
0.5297 2.27 220 0.5149 0.6742 0.5793 0.1741
0.5109 2.48 240 0.5049 0.6845 0.5824 0.1929
0.5265 2.69 260 0.5070 0.6846 0.5859 0.1799
0.5028 2.89 280 0.5068 0.6847 0.5870 0.1864
0.5097 3.1 300 0.5025 0.6892 0.5940 0.2084
0.4971 3.31 320 0.5032 0.6843 0.5995 0.1890
0.4762 3.51 340 0.5069 0.6955 0.5928 0.2129
0.4811 3.72 360 0.4954 0.6898 0.5996 0.2026
0.5065 3.93 380 0.4961 0.6918 0.6038 0.1838
0.4746 4.13 400 0.4992 0.6956 0.6009 0.2142
0.4786 4.34 420 0.5013 0.6918 0.6018 0.2026
0.4832 4.55 440 0.4935 0.6904 0.6031 0.2155
0.465 4.75 460 0.4984 0.6938 0.6027 0.2071
0.4683 4.96 480 0.4977 0.6960 0.6011 0.2091
0.4573 5.17 500 0.4985 0.6915 0.6076 0.2006
0.4619 5.37 520 0.4952 0.6945 0.6044 0.2129
0.4535 5.58 540 0.4983 0.6927 0.6024 0.2078
0.4475 5.79 560 0.4967 0.6970 0.6064 0.2194
0.454 5.99 580 0.5027 0.6941 0.6090 0.1994
0.4479 6.2 600 0.4940 0.6919 0.6041 0.2117
0.4304 6.41 620 0.5002 0.6982 0.6114 0.2006
0.445 6.61 640 0.4970 0.6951 0.6098 0.2071
0.4434 6.82 660 0.4964 0.6976 0.6075 0.2136
0.4543 7.03 680 0.4904 0.6936 0.6086 0.2013
0.4474 7.24 700 0.4969 0.6960 0.6108 0.2071
0.4325 7.44 720 0.4998 0.7013 0.6123 0.2123
0.4362 7.65 740 0.4947 0.6953 0.6101 0.2091
0.4276 7.86 760 0.4978 0.6955 0.6119 0.2052
0.4392 8.06 780 0.4944 0.6967 0.6078 0.2104
0.4167 8.27 800 0.4987 0.6966 0.6080 0.2097
0.4309 8.48 820 0.4970 0.6980 0.6126 0.2188
0.42 8.68 840 0.4999 0.6977 0.6105 0.2129
0.423 8.89 860 0.5003 0.6975 0.6087 0.2142
0.4382 9.1 880 0.4977 0.6975 0.6115 0.2136
0.4182 9.3 900 0.4976 0.6981 0.6123 0.2155
0.4153 9.51 920 0.5000 0.6978 0.6108 0.2175
0.4277 9.72 940 0.5003 0.6982 0.6092 0.2168
0.4246 9.92 960 0.5000 0.6976 0.6093 0.2168

Framework versions

  • PEFT 0.10.0
  • Transformers 4.39.3
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2