Hugging Face's logo

imdatta0
/
qwen_cnn_dailymail

Generated from Trainer
Model card Files Community
qwen_cnn_dailymail / README.md
imdatta0's picture
imdatta0
qwen-final-cnn_dailymail
ca17263
preview code
|
raw
history blame
3.88 kB
metadata
base_model: Qwen/Qwen-14B
tags:
  - generated_from_trainer
datasets:
  - cnn_dailymail
model-index:
  - name: final_cnn_dailymail
    results: []

final_cnn_dailymail

This model is a fine-tuned version of Qwen/Qwen-14B on the cnn_dailymail dataset. It achieves the following results on the evaluation set:

  • Loss: 2.2127

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: 1e-05
  • train_batch_size: 1
  • eval_batch_size: 1
  • seed: 42
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 0.01
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss
1.9757 0.02 100 1.9261
1.9258 0.04 200 1.8833
1.8977 0.06 300 1.8657
1.8903 0.08 400 1.8630
1.8858 0.1 500 1.8638
1.89 0.12 600 1.8636
1.873 0.14 700 1.8637
1.8908 0.16 800 1.8637
1.8791 0.18 900 1.8626
1.8851 0.2 1000 1.8634
1.89 0.22 1100 1.8651
1.8889 0.24 1200 1.8681
1.8896 0.26 1300 1.8708
1.8817 0.28 1400 1.8739
1.9003 0.3 1500 1.8791
1.9005 0.32 1600 1.8825
1.9024 0.34 1700 1.8864
1.9204 0.36 1800 1.8929
1.9182 0.38 1900 1.8955
1.9289 0.4 2000 1.9035
1.9348 0.42 2100 1.9157
1.9453 0.44 2200 1.9277
1.9689 0.46 2300 1.9457
1.9829 0.48 2400 1.9596
1.9874 0.5 2500 1.9803
2.0148 0.52 2600 1.9991
2.0391 0.54 2700 2.0249
2.0619 0.56 2800 2.0477
2.0736 0.58 2900 2.0678
2.0957 0.6 3000 2.0825
2.1223 0.62 3100 2.1097
2.1357 0.64 3200 2.1164
2.1759 0.66 3300 2.1524
2.168 0.68 3400 2.1650
2.1842 0.7 3500 2.1637
2.1956 0.72 3600 2.1775
2.2131 0.74 3700 2.1888
2.198 0.76 3800 2.1953
2.2231 0.78 3900 2.1994
2.2292 0.8 4000 2.2080
2.2343 0.82 4100 2.2093
2.2261 0.84 4200 2.2009
2.2104 0.86 4300 2.2015
2.2255 0.88 4400 2.2077
2.2299 0.9 4500 2.2099
2.2253 0.92 4600 2.2100
2.2239 0.94 4700 2.2116
2.2322 0.96 4800 2.2122
2.2457 0.98 4900 2.2127
2.2325 1.0 5000 2.2127

Framework versions

  • Transformers 4.32.0
  • Pytorch 2.1.0
  • Datasets 2.14.7
  • Tokenizers 0.13.3