This model is a fine-tuned version of
almanach/camemberta-base
on the HC3 FULL_FR_1.0_0.5_0.5 dataset with noise added.
It achieves the following results on the
Validation set:
Loss: 0.0430
F1: 0.9791
Test Set:
F1: 0.97
Adversarial:
F1: 0.45
Model description
This a model trained to detect text created by ChatGPT in French.
The training data is the
hc3_fr_full
subset of
almanach/hc3_multi
, but with added misspelling and homoglyph attacks.
Intended uses & limitations
This model is for research purposes only.
It is not intended to be used in production as we said in our paper:
We would like to emphasize that our study does not claim to have produced an universally accurate detector. Our strong results are based on in-domain testing and, unsurprisingly, do not generalize in out-of-domain scenarios. This is even more so when used on text specifically designed to fool language model detectors and on text intentionally stylistically similar to ChatGPT-generated text, especially instructional text.
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
learning_rate: 2e-05
train_batch_size: 8
eval_batch_size: 8
seed: 25
gradient_accumulation_steps: 4
total_train_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 5.0
mixed_precision_training: Native AMP
Training results
Training Loss
Epoch
Step
Validation Loss
F1
0.0199
1.0
4267
0.0430
0.9791
0.0104
2.0
8534
0.1457
0.9463
0.0026
3.0
12801
0.0805
0.9720
0.0
4.0
17068
0.2515
0.9419
0.0
5.0
21335
0.2000
0.9567
Framework versions
Transformers 4.26.1
Pytorch 1.11.0+cu115
Datasets 2.8.0
Tokenizers 0.13.2
Runs of almanach camemberta-chatgptdetect-noisy on huggingface.co
298
Total runs
0
24-hour runs
-35
3-day runs
-40
7-day runs
82
30-day runs
More Information About camemberta-chatgptdetect-noisy huggingface.co Model
camemberta-chatgptdetect-noisy huggingface.co
camemberta-chatgptdetect-noisy huggingface.co is an AI model on huggingface.co that provides camemberta-chatgptdetect-noisy's model effect (), which can be used instantly with this almanach camemberta-chatgptdetect-noisy model. huggingface.co supports a free trial of the camemberta-chatgptdetect-noisy model, and also provides paid use of the camemberta-chatgptdetect-noisy. Support call camemberta-chatgptdetect-noisy model through api, including Node.js, Python, http.
camemberta-chatgptdetect-noisy huggingface.co is an online trial and call api platform, which integrates camemberta-chatgptdetect-noisy's modeling effects, including api services, and provides a free online trial of camemberta-chatgptdetect-noisy, you can try camemberta-chatgptdetect-noisy online for free by clicking the link below.
almanach camemberta-chatgptdetect-noisy online free url in huggingface.co:
camemberta-chatgptdetect-noisy is an open source model from GitHub that offers a free installation service, and any user can find camemberta-chatgptdetect-noisy on GitHub to install. At the same time, huggingface.co provides the effect of camemberta-chatgptdetect-noisy install, users can directly use camemberta-chatgptdetect-noisy installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
camemberta-chatgptdetect-noisy install url in huggingface.co: