Update README.md
Browse files
README.md
CHANGED
@@ -15,9 +15,31 @@ The primary purpose of this fine-tuned model is to perform sentiment analysis on
|
|
15 |
|
16 |
# results
|
17 |
|
18 |
-
This model
|
19 |
-
|
20 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
|
22 |
## Model description
|
23 |
|
|
|
15 |
|
16 |
# results
|
17 |
|
18 |
+
This section outlines the performance of the model on the sentiment analysis task using the IMDb movie reviews dataset, both before and after the fine-tuning process. The results highlight the effectiveness of fine-tuning in enhancing model accuracy and generalization.
|
19 |
+
|
20 |
+
Pre-Fine-Tuning Evaluation:
|
21 |
+
|
22 |
+
Before fine-tuning, the model was evaluated on the IMDb dataset to establish a baseline for its performance. The initial evaluation yielded the following metrics:
|
23 |
+
|
24 |
+
Loss: 0.6518
|
25 |
+
Evaluation Runtime: 42.6174 seconds
|
26 |
+
Samples per Second: 58.662
|
27 |
+
These results indicate the model's performance with the original distilbert-base-uncased training, without any adjustments for the specific task of sentiment analysis on movie reviews.
|
28 |
+
|
29 |
+
Post-Fine-Tuning Evaluation:
|
30 |
+
|
31 |
+
After fine-tuning, the model was re-evaluated on the same dataset to assess improvements from the training adjustments:
|
32 |
+
|
33 |
+
Loss: 6.091e-06
|
34 |
+
Evaluation Runtime: 39.3821 seconds
|
35 |
+
Samples per Second: 63.481
|
36 |
+
The significant reduction in loss demonstrates a substantial increase in the model's accuracy and its capability to correctly classify sentiment in movie reviews. The improvement in processing speed (samples per second) also suggests enhanced efficiency post-fine-tuning.
|
37 |
+
|
38 |
+
Discussion:
|
39 |
+
|
40 |
+
The dramatic decrease in evaluation loss post-fine-tuning highlights the effectiveness of adapting the DistilBERT model to a specific dataset and task. This adaptation has markedly improved the model's predictive accuracy, making it a valuable tool for applications involving sentiment analysis of English text, particularly movie reviews.
|
41 |
+
|
42 |
+
These results illustrate the potential of fine-tuning pre-trained models on specific subsets of data to enhance their applicability to specialized tasks.
|
43 |
|
44 |
## Model description
|
45 |
|