Downloads · 30 days
0
mta-ttk/AstroSpeech-emotion-RoBERTa-large-MSP-Podcast-2.0-dimensions-LoRa
AstroSpeech-emotion-RoBERTa-large-MSP-Podcast-2.0-dimensions-LoRa is a machine learning model from mta-ttk. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers.
This model was fine-tuned for emotional analysis of texts, more precisely the prediction of emotion dimensions Active, Valuation, and Dominance. Training was conducted on the transcripts of MSP Podcast 2.0, a large na…
Downloads · 30 days
0
Access
Public
Updated Aug 13, 2026
Repo size
7.8 MB
Likes
0
Public
Click a slice to open those files.
.safetensors7.8 MB · 100%
From the Hugging Face model README
This model was fine-tuned for emotional analysis of texts, more precisely
the prediction of emotion dimensions Active, Valuation, and Dominance.
Training was conducted on the transcripts of MSP Podcast 2.0,
a large naturalistic speech emotion corpus
(Busso et al 2025).
The pretrained model is roberta-large, and Low-Rank
Adaptaters (LoRA) were
used for efficiency (compute, storage, and implicit regularization).
The model was made within the AstroSpeech project of the
MTA TTK Environmental Adaptation and Space Research
Group.
MSP Podcast is annotated for both emotion class and dimensions.
There are 8 emotion classes proper along with X (for annotator disagreement) and
O (for other emotions).
The examples belonging to these two special classes were
excluded from fine-tuning.
The emotion dimensions were rescaled with sklearn's StandardScaler.
The sequence representations are computed as the mean of the tokens.
The intermediate layer in the emotion dimension regressor is a ReLU
(instead of Dropout, wich is, as far as the author knows, transformers's default
for sequence classification).
LoRA rank is 2, and the training loss is weighted by the inverse frequency of
the emotion class of the training sample.
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
roberta-large[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
See load_model.py to get started with the model.
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
BibTeX:
[More Information Needed]
APA:
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]
[More Information Needed]