Update README.md
Browse files
README.md
CHANGED
|
@@ -11,7 +11,7 @@ tags: []
|
|
| 11 |
- **Output:** Joke or No-joke sentiment
|
| 12 |
|
| 13 |
## Training Data
|
| 14 |
-
- **Dataset:** 200k Short Texts for Humor Detection
|
| 15 |
- **Link:** https://www.kaggle.com/datasets/deepcontractor/200k-short-texts-for-humor-detection
|
| 16 |
- **Size:** 200,000 labeled short texts
|
| 17 |
- **Distribution:** Equally balanced between humor and non-humor
|
|
@@ -29,7 +29,6 @@ DistilBERT base model (uncased), a distilled version of BERT optimized for effic
|
|
| 29 |
| Batch Size | 32 (per device) |
|
| 30 |
| Learning Rate | 2e-4 |
|
| 31 |
| Weight Decay | 0.01 |
|
| 32 |
-
| Max Steps | Total training steps |
|
| 33 |
| Epochs | 2 |
|
| 34 |
| Warmup Steps | 100 |
|
| 35 |
| Best Model Selection | Based on eval_loss |
|
|
|
|
| 11 |
- **Output:** Joke or No-joke sentiment
|
| 12 |
|
| 13 |
## Training Data
|
| 14 |
+
- **Dataset:** 200k Short Texts for Humor Detection
|
| 15 |
- **Link:** https://www.kaggle.com/datasets/deepcontractor/200k-short-texts-for-humor-detection
|
| 16 |
- **Size:** 200,000 labeled short texts
|
| 17 |
- **Distribution:** Equally balanced between humor and non-humor
|
|
|
|
| 29 |
| Batch Size | 32 (per device) |
|
| 30 |
| Learning Rate | 2e-4 |
|
| 31 |
| Weight Decay | 0.01 |
|
|
|
|
| 32 |
| Epochs | 2 |
|
| 33 |
| Warmup Steps | 100 |
|
| 34 |
| Best Model Selection | Based on eval_loss |
|