psycho-emotion-detector_v2

Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->

Model for multilabel emotions classification. Trained on synthetic psychological dataset (Created by Claude Sonnet 4.6). Fine-tuned ai-forever/ruBert-base

Model Description

<!-- Provide a longer summary of what this model is. -->

10 classes: "neutral","joy","sadness","anger","enthusiasm","surprise","disgust","fear","guilt","shame"

Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

https://huggingface.co/datasets/VBadazhkov/therapy_emotions_ru

Training Hyperparameters

num_train_epochs=15

per_device_train_batch_size=16

label_smoothing_factor=0.1

learning_rate=2e-5

weight_decay=0.01

warmup_ratio=0.1

test_size=0.15

Evaluation

<!-- This section describes the evaluation protocols and provides the results. -->

Per-label F1:

surprise 0.723
guilt 0.745
enthusiasm 0.750
sadness 0.761
fear 0.827
joy 0.846
anger 0.880
neutral 0.889
shame 0.898
disgust 0.909