wisalkhanmv
commited on
Commit
•
78887c2
1
Parent(s):
67b2869
Update README.md
Browse files
README.md
CHANGED
@@ -13,3 +13,107 @@ tags:
|
|
13 |
- rating
|
14 |
- lowerated
|
15 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
- rating
|
14 |
- lowerated
|
15 |
---
|
16 |
+
|
17 |
+
# Lowerated/deberta-v3-lm6
|
18 |
+
|
19 |
+
## Model Details
|
20 |
+
**Model Name:** Lowerated/deberta-v3-lm6
|
21 |
+
**Model Type:** Text Classification (Aspect-Based Sentiment Analysis)
|
22 |
+
**Language:** English
|
23 |
+
**Framework:** PyTorch
|
24 |
+
**License:** Apache 2.0
|
25 |
+
|
26 |
+
## Model Description
|
27 |
+
Lowerated/deberta-v3-lm6 is a DeBERTa-v3-based model fine-tuned for aspect-based sentiment analysis on IMDb movie reviews. The model is designed to classify sentiments across seven key aspects of filmmaking: Cinematography, Direction, Story, Characters, Production Design, Unique Concept, and Emotions.
|
28 |
+
|
29 |
+
## Dataset
|
30 |
+
**Dataset Name:** Lowerated/imdb-reviews-rated
|
31 |
+
**Dataset URL:** [IMDb Reviews Rated](https://huggingface.co/datasets/LOWERATED/imdb-reviews-rated)
|
32 |
+
**Dataset Description:** The dataset contains IMDb movie reviews with sentiment scores for seven aspects of filmmaking. Each review is labeled with sentiment scores for Cinematography, Direction, Story, Characters, Production Design, Unique Concept, and Emotions.
|
33 |
+
|
34 |
+
## Performance
|
35 |
+
**Evaluation Metric:** Mean Squared Error (MSE)
|
36 |
+
**MSE:** 0.08594679832458496
|
37 |
+
|
38 |
+
### Detailed Results
|
39 |
+
|
40 |
+
**Cinematography:**
|
41 |
+
- **Precision:** 0.96
|
42 |
+
- **Recall:** 0.97
|
43 |
+
- **F1-score:** 0.96
|
44 |
+
- **Accuracy:** 0.95
|
45 |
+
|
46 |
+
**Confusion Matrix:**
|
47 |
+
\[
|
48 |
+
\begin{bmatrix}
|
49 |
+
68 & 0 & 0 \\
|
50 |
+
1 & 377 & 37 \\
|
51 |
+
0 & 0 & 310 \\
|
52 |
+
\end{bmatrix}
|
53 |
+
\]
|
54 |
+
|
55 |
+
**Direction:**
|
56 |
+
- **Precision:** 0.93
|
57 |
+
- **Recall:** 0.97
|
58 |
+
- **F1-score:** 0.94
|
59 |
+
- **Accuracy:** 0.95
|
60 |
+
|
61 |
+
**Story:**
|
62 |
+
- **Precision:** 0.85
|
63 |
+
- **Recall:** 0.88
|
64 |
+
- **F1-score:** 0.85
|
65 |
+
- **Accuracy:** 0.85
|
66 |
+
|
67 |
+
|
68 |
+
**Characters:**
|
69 |
+
- **Precision:** 0.89
|
70 |
+
- **Recall:** 0.89
|
71 |
+
- **F1-score:** 0.89
|
72 |
+
- **Accuracy:** 0.90
|
73 |
+
|
74 |
+
**Production Design:**
|
75 |
+
- **Precision:** 0.95
|
76 |
+
- **Recall:** 0.98
|
77 |
+
- **F1-score:** 0.96
|
78 |
+
- **Accuracy:** 0.96
|
79 |
+
|
80 |
+
**Unique Concept:**
|
81 |
+
- **Precision:** 0.83
|
82 |
+
- **Recall:** 1.00
|
83 |
+
- **F1-score:** 0.89
|
84 |
+
- **Accuracy:** 1.00
|
85 |
+
|
86 |
+
**Emotions:**
|
87 |
+
- **Precision:** 0.76
|
88 |
+
- **Recall:** 0.87
|
89 |
+
- **F1-score:** 0.78
|
90 |
+
- **Accuracy:** 0.82
|
91 |
+
|
92 |
+
**Test Results:**
|
93 |
+
- **Eval Loss:** 0.08594681322574615
|
94 |
+
- **Eval Model Preparation Time:** 0.0011
|
95 |
+
- **Eval MSE:** 0.08594679832458496
|
96 |
+
- **Eval Runtime:** 23.1411
|
97 |
+
- **Eval Samples per Second:** 34.268
|
98 |
+
- **Eval Steps per Second:** 8.599
|
99 |
+
|
100 |
+
## Intended Use
|
101 |
+
This model is intended for rating of movies across seven aspects of filmmaking. It can be used to provide a more nuanced understanding of viewer opinions and improve movie rating systems.
|
102 |
+
|
103 |
+
## Limitations
|
104 |
+
While the model performs well on the evaluation dataset, its performance may vary on different datasets. Continuous monitoring and retraining with diverse data are recommended to maintain and improve its accuracy.
|
105 |
+
|
106 |
+
## Future Work
|
107 |
+
Future improvements could focus on exploring alternative methods for handling neutral values, investigating advanced techniques for addressing missing ratings, enhancing sentiment analysis methods, and expanding the range of aspects analyzed.
|
108 |
+
|
109 |
+
## Citation
|
110 |
+
If you use this model in your research, please cite it as follows:
|
111 |
+
|
112 |
+
```bibtex
|
113 |
+
@model{lowerated_deberta-v3-lm6,
|
114 |
+
author = {LOWERATED},
|
115 |
+
title = {deberta-v3-lm6},
|
116 |
+
year = {2024},
|
117 |
+
url = {https://huggingface.co/Lowerated/deberta-v3-lm6},
|
118 |
+
}
|
119 |
+
```
|