| # ClassTrackClassify |
|
|
| A fine-tuned **DistilBERT** model for **single-label text classification**. The model predicts one of four intent-style labels: `action`, `question`, `recall`, or `statement`. |
|
|
| > [!IMPORTANT] |
| > This model is part of a personal project and is provided for experimentation and learning purposes. No further support or revisions guranteed. |
|
|
| ## Labels |
|
|
| | ID | Label | |
| | -- | --------- | |
| | 0 | action | |
| | 1 | question | |
| | 2 | recall | |
| | 3 | statement | |
|
|
| ## Model Details |
|
|
| * Architecture: DistilBertForSequenceClassification |
| * Base model: DistilBERT |
| * Hidden size: 768 |
| * Layers: 6 |
| * Heads: 12 |
| * Max length: 512 |
| * Precision: float32 |
|
|
| ## Usage |
|
|
| ```python |
| from transformers import AutoTokenizer, AutoModelForSequenceClassification |
| import torch |
| |
| model_id = "AaryanK/ClassTrackClassify" |
| tokenizer = AutoTokenizer.from_pretrained(model_id) |
| model = AutoModelForSequenceClassification.from_pretrained(model_id) |
| |
| text = "What did we talk about earlier?" |
| inputs = tokenizer(text, return_tensors="pt", truncation=True) |
| |
| with torch.no_grad(): |
| logits = model(**inputs).logits |
| |
| label_id = logits.argmax(dim=-1).item() |
| print(model.config.id2label[str(label_id)]) |
| ``` |
|
|
| ## Intended Use |
|
|
| Lightweight intent and utterance-type classification for conversational systems. |
|
|
| --- |
|
|
