dusersad12 commited on
Commit
f51f23c
·
verified ·
1 Parent(s): a01eddb

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +57 -0
README.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ library_name: transformers
4
+ tags:
5
+ - text-classification
6
+ - roberta
7
+ - sweep
8
+ ---
9
+ # SweepBestModel
10
+
11
+ <div align="center">
12
+ <img src="figures/training_curve.png" width="70%" alt="Training Curve" />
13
+ </div>
14
+
15
+ ## Overview
16
+
17
+ This model was selected from a hyperparameter sweep as the best-performing run based on validation accuracy. It is a RoBERTa-based sequence classifier fine-tuned on our internal dataset.
18
+
19
+ ## Training Configuration
20
+
21
+ - **Learning Rate:** 3e-5
22
+ - **Batch Size:** 64
23
+ - **Epochs:** 10
24
+ - **Best Validation Accuracy:** 0.864
25
+
26
+ ## Benchmark Results
27
+
28
+ <div align="center">
29
+
30
+ | Benchmark | Score |
31
+ |---|---|
32
+ | MNLI (m/mm) | 0.864 |
33
+ | SST-2 | 0.864 |
34
+ | QQP | 0.864 |
35
+ | QNLI | 0.864 |
36
+ | RTE | 0.864 |
37
+ | CoLA | 0.864 |
38
+ | STS-B | 0.864 |
39
+ | MRPC | 0.864 |
40
+
41
+ </div>
42
+
43
+ ## Usage
44
+
45
+ ```python
46
+ from transformers import AutoModelForSequenceClassification, AutoTokenizer
47
+
48
+ model = AutoModelForSequenceClassification.from_pretrained("SweepBestModel-Repo")
49
+ tokenizer = AutoTokenizer.from_pretrained("SweepBestModel-Repo")
50
+ ```
51
+
52
+ ## Figures
53
+
54
+ <div align="center">
55
+ <img src="figures/confusion_matrix.png" width="45%" alt="Confusion Matrix" />
56
+ <img src="figures/loss_curve.png" width="45%" alt="Loss Curve" />
57
+ </div>