Downloads · 30 days
37
26% of all-time downloads
larc-iu/topdown_biaffine-gum-12.1.0
topdown_biaffine-gum-12.1.0 is a machine learning model from larc-iu. 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 dipo.
<p align="center" <a href="https://github.com/larc-iu/dipo"<img src="https://raw.githubusercontent.com/larc-iu/dipo/master/docs/logo.svg" alt="Dipo logo" width="200"</a </p
Downloads · 30 days
37
26% of all-time downloads
All-time downloads
141
Public
Repo size
1.5 GB
Likes
0
Public
Click a slice to open those files.
.pt1.5 GB · 100%
From the Hugging Face model README
A pretrained Top-down Biaffine RST parser trained with Dipo.
This model uses SpanBERT/spanbert-base-cased as its underlying encoder.
GUM 12.1.0 (Georgetown University Multilayer corpus) (English). A multi-genre English corpus (academic, biography, fiction, interview, news, reddit, speech, textbook, vlog, voyage, whow) annotated for RST among many other layers.
Relation labels: 32 labels:
adversative-antithesis, adversative-concession, adversative-contrast, attribution-negative, attribution-positive, causal-cause, causal-result, context-background, context-circumstance, contingency-condition, elaboration-additional, elaboration-attribute, evaluation-comment, explanation-evidence, explanation-justify, explanation-motivation, joint-disjunction, joint-list, joint-other, joint-sequence, mode-manner, mode-means, organization-heading, organization-phatic, organization-preparation, purpose-attribute, purpose-goal, restatement-partial, restatement-repetition, same-unit, topic-question, topic-solutionhood
| Split | span_f1 | nuc_f1 | rel_f1 | full_f1 |
|---|---|---|---|---|
| dev | 0.6364 | 0.4979 | 0.3740 | 0.3716 |
| test | 0.6669 | 0.5295 | 0.4033 | 0.4003 |
Install the library first with pip install dipo.
Programmatic:
from dipo.rst.parsers.topdown_biaffine.modeling_topdown_biaffine import TopdownBiaffineParser
parser = TopdownBiaffineParser.from_pretrained("larc-iu/topdown_biaffine-gum-12.1.0")
# This parser requires gold EDU segmentation, so the input must be an RS3/RS4 file.
from dipo.rst.data.reader import read_rst_file
gold = read_rst_file(
"doc.rs3",
relation_types=parser.config.relation_types,
relation_map=parser.config.relation_map,
)
tree = parser.predict(gold)
CLI:
python -m dipo topdown_biaffine predict --hub-id larc-iu/topdown_biaffine-gum-12.1.0 --input <doc.rs3> --output-dir out/
If you use this model, please cite both the paper this parser re-implements:
@inproceedings{topdown_biaffine_paper,
title = {A Simple and Strong Baseline for End-to-End Neural RST-style Discourse Parsing},
author = {Naoki Kobayashi, Tsutomu Hirao, Hidetaka Kamigaito, Manabu Okumura, Masaaki Nagata},
booktitle = {Findings of EMNLP 2022},
url = {https://aclanthology.org/2022.findings-emnlp.501/},
}
And the paper Dipo was released with:
@misc{gessler-zeldes-2026-worth,
title = {{Worth Their Weights? Evaluating Generative Models for Discourse Parsing}},
author = {Gessler, Luke and Zeldes, Amir},
year = {2026},
eprint = {2610.17658},
archivePrefix = {arXiv},
primaryClass = {cs.CL},
url = {https://arxiv.org/abs/2610.17658},
}
See below for the full training configuration this model was trained with.
{
"train_dir": "data/gum_12.1.0/train",
"dev_dir": "data/gum_12.1.0/dev",
"test_dir": "data/gum_12.1.0/test",
"relation_types": [
[
"adversative-antithesis",
"rst"
],
[
"adversative-concession",
"rst"
],
[
"adversative-contrast",
"multinuc"
],
[
"attribution-negative",
"rst"
],
[
"attribution-positive",
"rst"
],
[
"causal-cause",
"rst"
],
[
"causal-result",
"rst"
],
[
"context-background",
"rst"
],
[
"context-circumstance",
"rst"
],
[
"contingency-condition",
"rst"
],
[
"elaboration-additional",
"rst"
],
[
"elaboration-attribute",
"rst"
],
[
"evaluation-comment",
"rst"
],
[
"explanation-evidence",
"rst"
],
[
"explanation-justify",
"rst"
],
[
"explanation-motivation",
"rst"
],
[
"joint-disjunction",
"multinuc"
],
[
"joint-list",
"multinuc"
],
[
"joint-other",
"multinuc"
],
[
"joint-sequence",
"multinuc"
],
[
"mode-manner",
"rst"
],
[
"mode-means",
"rst"
],
[
"organization-heading",
"rst"
],
[
"organization-phatic",
"rst"
],
[
"organization-preparation",
"rst"
],
[
"purpose-attribute",
"rst"
],
[
"purpose-goal",
"rst"
],
[
"restatement-partial",
"rst"
],
[
"restatement-repetition",
"multinuc"
],
[
"same-unit",
"multinuc"
],
[
"topic-question",
"rst"
],
[
"topic-solutionhood",
"rst"
]
],
"relation_map": null,
"model_name": "SpanBERT/spanbert-base-cased",
"ffn_hidden_size": 512,
"dropout": 0.2,
"stride": 100,
"lr": 0.0002,
"encoder_lr": 1e-05,
"max_epochs": 30,
"grad_accum": 1,
"patience": 10,
"max_grad_norm": 1,
"weight_decay": 0.01,
"num_warmup_steps": 1000,
"log_every": 50,
"validate_every": null,
"checkpoint_every": null,
"checkpoint_dir": "checkpoints",
"run_name": null,
"seed": 42,
"val_metric_name": "span_f1"
}