Instructions to use dnzblgn/Customer-Reviews-Classification with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use dnzblgn/Customer-Reviews-Classification with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="dnzblgn/Customer-Reviews-Classification")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("dnzblgn/Customer-Reviews-Classification") model = AutoModelForSequenceClassification.from_pretrained("dnzblgn/Customer-Reviews-Classification", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -40,11 +40,17 @@ A custom synthetic dataset was created for this task, focusing on the structural
|
|
| 40 |
|
| 41 |
### Training Hyperparameters
|
| 42 |
Model: distilbert/distilbert-base-uncased
|
|
|
|
| 43 |
Learning Rate: 3e-5
|
|
|
|
| 44 |
Epochs: 7
|
|
|
|
| 45 |
Train Batch Size: 16
|
|
|
|
| 46 |
Gradient Accumulation Steps: 2
|
|
|
|
| 47 |
Weight Decay: 0.015
|
|
|
|
| 48 |
Warm-up Ratio: 0.1
|
| 49 |
|
| 50 |
### Evaluation
|
|
@@ -52,8 +58,11 @@ The model was evaluated using a custom dataset representing the same six documen
|
|
| 52 |
|
| 53 |
### Metrics
|
| 54 |
Accuracy: 0.947
|
|
|
|
| 55 |
Precision: 0.948
|
|
|
|
| 56 |
Recall: 0.948
|
|
|
|
| 57 |
F1-Score: 0.948
|
| 58 |
|
| 59 |
|
|
|
|
| 40 |
|
| 41 |
### Training Hyperparameters
|
| 42 |
Model: distilbert/distilbert-base-uncased
|
| 43 |
+
|
| 44 |
Learning Rate: 3e-5
|
| 45 |
+
|
| 46 |
Epochs: 7
|
| 47 |
+
|
| 48 |
Train Batch Size: 16
|
| 49 |
+
|
| 50 |
Gradient Accumulation Steps: 2
|
| 51 |
+
|
| 52 |
Weight Decay: 0.015
|
| 53 |
+
|
| 54 |
Warm-up Ratio: 0.1
|
| 55 |
|
| 56 |
### Evaluation
|
|
|
|
| 58 |
|
| 59 |
### Metrics
|
| 60 |
Accuracy: 0.947
|
| 61 |
+
|
| 62 |
Precision: 0.948
|
| 63 |
+
|
| 64 |
Recall: 0.948
|
| 65 |
+
|
| 66 |
F1-Score: 0.948
|
| 67 |
|
| 68 |
|