xlm-roberta-large-czech-parlspeech-cap-v3

Model description

An xlm-roberta-large model fine-tuned on english training data containing parliamentary speeches (oral questions, interpellations, bill debates, other plenary speeches, urgent questions) labeled with major topic codes from the Comparative Agendas Project.

How to use the model

from transformers import AutoTokenizer, pipeline

tokenizer = AutoTokenizer.from_pretrained("xlm-roberta-large")
pipe = pipeline(
    model="poltextlab/xlm-roberta-large-czech-parlspeech-cap-v3",
    task="text-classification",
    tokenizer=tokenizer,
    use_fast=False,
    token="<your_hf_read_only_token>"
)

text = "We will place an immediate 6-month halt on the finance driven closure of beds and wards, and set up an independent audit of needs and facilities."
pipe(text)

The translation table from the model results to CAP codes is the following:

CAP_NUM_DICT = {
    0: 1,
    1: 2,
    2: 3,
    3: 4,
    4: 5,
    5: 6,
    6: 7,
    7: 8,
    8: 9,
    9: 10,
    10: 12,
    11: 13,
    12: 14,
    13: 15,
    14: 16,
    15: 17,
    16: 18,
    17: 19,
    18: 20,
    19: 21,
    20: 23,
    21: 999,
}

Gated access

Due to the gated access, you must pass the token parameter when loading the model. In earlier versions of the Transformers package, you may need to use the use_auth_token parameter instead.

Model performance

The model was evaluated on a test set of 96127 examples.
Model accuracy is 0.85.

label	precision	recall	f1-score	support
0	0.73	0.78	0.76	30742
1	0.7	0.54	0.61	7882
2	0.79	0.86	0.82	9461
3	0.75	0.77	0.76	7755
4	0.6	0.66	0.63	8293
5	0.88	0.82	0.85	8664
6	0.68	0.76	0.72	3922
7	0.76	0.74	0.75	4096
8	0.77	0.65	0.7	2374
9	0.83	0.83	0.83	5652
10	0.76	0.72	0.74	11487
11	0.67	0.63	0.65	6985
12	0.7	0.65	0.67	5956
13	0.74	0.62	0.68	10150
14	0.85	0.74	0.79	4826
15	0.83	0.73	0.77	3760
16	0.57	0.48	0.52	2121
17	0.71	0.73	0.72	13222
18	0.66	0.7	0.68	33867
19	0.61	0.64	0.62	4981
20	0.69	0.52	0.59	1887
21	0.97	0.97	0.97	196422
macro avg	0.74	0.71	0.72	384505
weighted avg	0.85	0.85	0.85	384505

Fine-tuning procedure

This model was fine-tuned with the following key hyperparameters:

Number of Training Epochs: 10
Batch Size: 40
Learning Rate: 5e-06
Early Stopping: enabled with a patience of 2 epochs

Inference platform

This model is used by the CAP Babel Machine, an open-source and free natural language processing tool, designed to simplify and speed up projects for comparative research.

Cooperation

Model performance can be significantly improved by extending our training sets. We appreciate every submission of CAP-coded corpora (of any domain and language) at poltextlab{at}poltextlab{dot}com or by using the CAP Babel Machine.

Reference

Sebők, M., Máté, Á., Ring, O., Kovács, V., & Lehoczki, R. (2024). Leveraging Open Large Language Models for Multilingual Policy Topic Classification: The Babel Machine Approach. Social Science Computer Review, 0(0). https://doi.org/10.1177/08944393241259434

Debugging and issues

This architecture uses the sentencepiece tokenizer. In order to use the model before transformers==4.27 you need to install it manually.

If you encounter a RuntimeError when loading the model using the from_pretrained() method, adding ignore_mismatched_sizes=True should solve the issue.

poltextlab
/

xlm-roberta-large-czech-parlspeech-cap-v3

You need to agree to share your contact information to access this model