Model Card for mt5-base-multi-label-all-cs-iv

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

This model is fine-tuned for multi-label seq2seq text classification of Supportive Interactions in Instant Messenger dialogs of Adolescents.

Model Description

The model was fine-tuned on a dataset of Czech Instant Messenger dialogs of Adolescents. The classification is multi-label. For each of the utterances in the input, the model outputs any combination of the tags:'NO TAG', 'Informační podpora', 'Emocionální podpora', 'Začlenění do skupiny', 'Uznání', 'Nabídka pomoci': as a string joined with ', ' (ordered alphabetically). Each label indicates the presence of that category of Supportive Interactions: 'no tag', 'informational support', 'emocional support', 'social companionship', 'appraisal', 'instrumental support' in each of the utterances of the input. The inputs of the model is a sequence of utterances joined with ';'. The outputs are a sequence of per-utterance labels such as: 'NO TAG; Informační podpora, Uznání; NO TAG'

Model Sources

<!-- Provide the basic links for the model. -->

Usage

Here is how to use this model to classify a context-window of a dialogue:

import itertools
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
import torch

# Target dialog context window
test_texts = ['Utterance1;Utterance2;Utterance3']

# Load the model and tokenizer
checkpoint_path = "chi2024/mt5-base-multi-label-all-cs-iv"
model = AutoModelForSeq2SeqLM.from_pretrained(checkpoint_path)\
    .to("cuda" if torch.cuda.is_available() else "cpu")
tokenizer = AutoTokenizer.from_pretrained(checkpoint_path)

# Define helper functions
def predict_one(text):
    inputs = tokenizer(text, return_tensors="pt", padding=True,
                       truncation=True, max_length=256).to(model.device)
    outputs = model.generate(**inputs)
    decoded = [text.split(",")[0].strip() for text in
               tokenizer.batch_decode(outputs, skip_special_tokens=True)]
    predicted_sequence = list(
        itertools.chain(*(pred_one.split("; ") for pred_one in decoded)))
    return predicted_sequence

# Run the prediction
dec = predict_one(test_texts[0])
print(dec)