metadata
model-index:
- name: poltextlab/media2-25-26-v1-1001
results:
- task:
type: text-classification
metrics:
- name: Accuracy
type: accuracy
value: 71%
- name: F1-Score
type: f1
value: 70%
tags:
- text-classification
- transformers
- roberta
metrics:
- accuracy
- f1_score
language:
- en
base_model:
- xlm-roberta-large
pipeline_tag: text-classification
library_name: transformers
license: cc-by-4.0
extra_gated_prompt: >-
Our models are intended for academic projects and academic research only. If
you are not affiliated with an academic institution, please reach out to us at
huggingface [at] poltextlab [dot] com for further inquiry. If we cannot
clearly determine your academic affiliation and use case based on your form
data, your request may be rejected. Please allow us a few business days to
manually review subscriptions.
extra_gated_fields:
Country: country
Institution: text
Institution Email: text
Full Name: text
Please specify your academic project/use case you want to use the models for: text
media2-25-26-v1-1001
This model uses the poltextLAB Media2 codebook built on top of the CAP codebook.
How to use the model
from transformers import AutoTokenizer, pipeline
tokenizer = AutoTokenizer.from_pretrained("xlm-roberta-large")
pipe = pipeline(
model="poltextlab/media2-25-26-v1-1001",
task="text-classification",
tokenizer=tokenizer,
use_fast=False,
token="<your_hf_read_only_token>"
)
text = "<text_to_classify>"
pipe(text)
Classification Report
Overall Performance:
Evaluated on a test set of 1601 English samples.
- Accuracy: 71%
- Macro Avg: Precision: 0.67, Recall: 0.62, F1-score: 0.62
- Weighted Avg: Precision: 0.74, Recall: 0.71, F1-score: 0.70
Per-Class Metrics:
| Label | Precision | Recall | F1-score | Support |
|---|---|---|---|---|
| 1 | 0.77 | 0.8 | 0.78 | 50 |
| 2 | 0.74 | 0.78 | 0.76 | 50 |
| 3 | 0.74 | 0.74 | 0.74 | 50 |
| 4 | 0.7 | 0.86 | 0.77 | 50 |
| 5 | 0.86 | 0.76 | 0.81 | 50 |
| 6 | 0.83 | 0.98 | 0.9 | 50 |
| 7 | 0.85 | 0.88 | 0.86 | 50 |
| 8 | 0.87 | 0.94 | 0.9 | 50 |
| 9 | 0.87 | 0.82 | 0.85 | 50 |
| 10 | 0.77 | 0.94 | 0.85 | 50 |
| 12 | 0.56 | 0.88 | 0.69 | 50 |
| 13 | 0.88 | 0.86 | 0.87 | 50 |
| 14 | 0.73 | 0.76 | 0.75 | 50 |
| 15 | 0.51 | 0.86 | 0.64 | 50 |
| 16 | 0.75 | 0.86 | 0.8 | 50 |
| 17 | 0.63 | 0.76 | 0.69 | 50 |
| 18 | 0.91 | 0.82 | 0.86 | 50 |
| 19 | 0.51 | 0.82 | 0.63 | 50 |
| 20 | 0.62 | 0.92 | 0.74 | 50 |
| 21 | 0.75 | 0.8 | 0.78 | 50 |
| 23 | 0.52 | 0.78 | 0.62 | 50 |
| 24 | 0.71 | 0.57 | 0.63 | 42 |
| 25 | 0.92 | 0.48 | 0.63 | 23 |
| 26 | 0.92 | 0.56 | 0.7 | 43 |
| 27 | 0 | 0 | 0 | 18 |
| 28 | 0 | 0 | 0 | 9 |
| 29 | 0.43 | 0.27 | 0.33 | 33 |
| 30 | 0.72 | 0.28 | 0.41 | 46 |
| 31 | 0.89 | 0.44 | 0.59 | 36 |
| 32 | 0 | 0 | 0 | 20 |
| 33 | 0.12 | 0.08 | 0.1 | 12 |
| 34 | 0.07 | 0.14 | 0.1 | 7 |
| 35 | 0.93 | 0.71 | 0.81 | 35 |
| 36 | 0 | 0 | 0 | 3 |
| 37 | 1 | 0.82 | 0.9 | 44 |
| 38 | 0.81 | 0.81 | 0.81 | 42 |
| 39 | 1 | 0.39 | 0.57 | 33 |
| 40 | 0.88 | 0.21 | 0.34 | 33 |
| 41 | 1 | 0.78 | 0.88 | 32 |
| 998 | 0.92 | 0.55 | 0.69 | 40 |
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.
Debugging and issues
This architecture uses the sentencepiece tokenizer. In order to run the model before transformers==4.27 you need to install it manually.