jalwa-1.5B / README.md
adarshxs's picture
Update README.md (#1)
7c03ff7 verified
---
library_name: transformers
license: other
base_model: tinycompany/ShawtyIsBad-nomic1.5
tags:
- llama-factory
- full
- generated_from_trainer
model-index:
- name: pretrain
results: []
---
| Task | Shot | Accuracy | Precision | Recall | F1 Score |
|------------------|------|----------|-----------|--------|----------|
| indicxparaphrase | 5 | 0.6119 | 0.9516 | 0.2358 | 0.3779 |
| indicxparaphrase | 0 | 0.7268 | 0.8650 | 0.5375 | 0.6630 |
| indicxnli | 0 | 0.4689 | 0.6390 | 0.4689 | 0.3782 |
| indicxnli | 5 | 0.5022 | 0.5180 | 0.5022 | 0.4651 |
| indicsentiment | 0 | 0.8367 | 0.7511 | 1.0000 | 0.8579 |
| indicsentiment | 5 | 0.9689 | 0.9894 | 0.9472 | 0.9678 |
| indiccopa | 0 | 0.6102 | 0.5701 | 0.8604 | 0.6858 |
| indiccopa | 5 | 0.7194 | 0.7000 | 0.7568 | 0.7273 |
Todo Tasks:
- [ ] `indicheadline`
- [ ] `indicqa`
- [ ] `in22` Dataset [link](https://indictrans2-public.objectstore.e2enetworks.net/IN22_benchmark.tar.gz) link doesnt work for now.
- [ ] Comparisons with other Indic models.