Downloads · 30 days
3
5% of all-time downloads
besstie_sarcasm_bart_finetune_lora_5runs
tlam25/besstie_sarcasm_bart_finetune_lora_5runs
besstie_sarcasm_bart_finetune_lora_5runs is a machine learning model from tlam25. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Best checkpoint from 15 runs on BESSTIE Sarcasm. Best (topic, seed) selected by val F1 macro: - topic = en-AUReddit - seed = 999 - val F1 macro = 72.78
All-time downloads
58
Public
Parameters
406M
813 MB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors813 MB · 100%
At a glance
- License
- apache-2.0
- Model type
- bart
- Access
- Public
- Created
- Apr 27, 2026
- Updated
- Apr 27, 2026
- SHA
- 6947588a
Model card
From the Hugging Face model README
BESSTIE Sarcasm - LoRA fine-tuning (BART-large)
Best checkpoint from 15 runs on BESSTIE Sarcasm.
Best (topic, seed) selected by val F1 macro:
- topic = en-AU_Reddit
- seed = 999
- val F1 macro = 72.78
Test results (mean ± std)
| Group | Accuracy | P (macro) | R (macro) | F1 (macro) | P (sarcastic) | R (sarcastic) | F1 (sarcastic) |
|---|---|---|---|---|---|---|---|
| en-AU_Reddit | 73.69 ± 0.47 | 73.38 ± 0.80 | 71.95 ± 0.46 | 72.28 ± 0.42 | 72.01 ± 2.31 | 61.19 ± 3.24 | 66.07 ± 1.07 |
| en-IN_Reddit | 86.52 ± 0.00 | 43.26 ± 0.00 | 50.00 ± 0.00 | 46.39 ± 0.00 | 0.00 ± 0.00 | 0.00 ± 0.00 | 0.00 ± 0.00 |
| en-UK_Reddit | 78.01 ± 0.00 | 39.01 ± 0.00 | 50.00 ± 0.00 | 43.82 ± 0.00 | 0.00 ± 0.00 | 0.00 ± 0.00 | 0.00 ± 0.00 |
| ALL Sarcasm (combined) | 79.41 ± 5.52 | 51.88 ± 15.84 | 57.32 ± 10.71 | 54.16 ± 13.31 | 24.00 ± 35.16 | 20.40 ± 29.91 | 22.02 ± 32.24 |
Aggregations:
- Per topic: 5 seeds.
- ALL Sarcasm (combined): 3 topics × 5 seeds = 15 runs.
Topics: ['en-AU_Reddit', 'en-IN_Reddit', 'en-UK_Reddit'] Seeds: [2025, 42, 123, 7, 999]
- Type
- bart
- License
- apache-2.0
- Languages
- en
- Created
- Apr 27, 2026
- Updated
- Apr 27, 2026