Update README.md
Browse files
README.md
CHANGED
@@ -2,7 +2,36 @@
|
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
|
5 |
+
**Hyperparameters:**
|
6 |
+
|
7 |
+
- learning rate: 2e-5
|
8 |
+
- weight decay: 0.01
|
9 |
+
- per_device_train_batch_size: 8
|
10 |
+
- per_device_eval_batch_size: 8
|
11 |
+
- gradient_accumulation_steps:1
|
12 |
+
- eval steps: 6000
|
13 |
+
- max_length: 512
|
14 |
+
- num_epochs: 2
|
15 |
+
|
16 |
+
**Dataset version:**
|
17 |
+
- “craffel/tasky_or_not”, “10xp3_10xc4”, “15f88c8”
|
18 |
+
|
19 |
+
**Checkpoint:**
|
20 |
+
|
21 |
+
- 48000 steps
|
22 |
+
|
23 |
+
**Results on Validation set:**
|
24 |
+
|
25 |
+
| Step | Training Loss | Validation Loss | Accuracy | Precision | Recall | F1 |
|
26 |
+
|-------|---------------|-----------------|----------|-----------|----------|----------|
|
27 |
+
| 6000 | 0.031900 | 0.163412 | 0.982194 | 0.999211 | 0.980462 | 0.989748 |
|
28 |
+
| 12000 | 0.014700 | 0.106132 | 0.976666 | 0.999639 | 0.973733 | 0.986516 |
|
29 |
+
| 18000 | 0.010700 | 0.043012 | 0.995743 | 0.999223 | 0.995918 | 0.997568 |
|
30 |
+
| 24000 | 0.007400 | 0.095047 | 0.984724 | 0.999857 | 0.982714 | 0.991211 |
|
31 |
+
| 30000 | 0.004100 | 0.087274 | 0.990400 | 0.999829 | 0.989217 | 0.994495 |
|
32 |
+
| 36000 | 0.003100 | 0.162909 | 0.981972 | 1.000000 | 0.979434 | 0.989610 |
|
33 |
+
| 42000 | 0.002200 | 0.148721 | 0.980454 | 0.999986 | 0.977717 | 0.988726 |
|
34 |
+
| 48000 | 0.001000 | 0.094455 | 0.990437 | 0.999943 | 0.989147 | 0.994516 |
|
35 |
+
|
36 |
+
|
37 |
+
|