Commit
•
a5241b2
1
Parent(s):
085be34
Update README.md
Browse files
README.md
CHANGED
|
@@ -4,4 +4,44 @@ datasets:
|
|
| 4 |
- squad
|
| 5 |
language:
|
| 6 |
- id
|
| 7 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
- squad
|
| 5 |
language:
|
| 6 |
- id
|
| 7 |
+
license: apache-2.0
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 11 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 12 |
+
|
| 13 |
+
# idt5-base-finetune-squad
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [muchad/idt5-base](https://huggingface.co/muchad/idt5-base) on on translated Indonesian SQUAD dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Exact: 49.19
|
| 18 |
+
- F1: 67.36
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
## Training procedure
|
| 22 |
+
|
| 23 |
+
### Training hyperparameters
|
| 24 |
+
|
| 25 |
+
The following hyperparameters were used during training:
|
| 26 |
+
- learning_rate: 0.0001
|
| 27 |
+
- train_batch_size: 4
|
| 28 |
+
- eval_batch_size: 64
|
| 29 |
+
- seed: 42
|
| 30 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 31 |
+
- lr_scheduler_type: linear
|
| 32 |
+
- num_epochs: 2
|
| 33 |
+
|
| 34 |
+
### Training results
|
| 35 |
+
|
| 36 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 37 |
+
|:-------------:|:-----:|:-----:|:---------------:|:
|
| 38 |
+
| 0.5682 | 1.1 | 36000 | 1.3641 |
|
| 39 |
+
| 0.5153 | 1.93 | 63000 | 1.2890 |
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
### Framework versions
|
| 43 |
+
|
| 44 |
+
- Transformers 4.35.0
|
| 45 |
+
- Pytorch 2.1.0+cu118
|
| 46 |
+
- Datasets 2.14.6
|
| 47 |
+
- Tokenizers 0.14.1
|