Distilbert Base Uncased Sst2

🚀 DistilBERT-Base-Uncased for Sentiment Analysis

This model is a fine - tuned DistilBERT for sentiment analysis, offering efficient and accurate sentiment classification.

🚀 Quick Start

This model is a fine - tuned version of [distilbert - base - uncased](https://huggingface.co/distilbert - base - uncased), originally introduced in "DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter". It was trained on the Stanford Sentiment Treebank v2 (SST2), which is part of the General Language Understanding Evaluation (GLUE) benchmark. The fine - tuning was carried out by the team at AssemblyAI, and it comes with a corresponding blog post.

💻 Usage Examples

Basic Usage

import torch.nn.functional as F 
from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("assemblyai/distilbert-base-uncased-sst2")
model = AutoModelForSequenceClassification.from_pretrained("assemblyai/distilbert-base-uncased-sst2")

tokenized_segments = tokenizer(["AssemblyAI is the best speech-to-text API for modern developers with performance being second to none!"], return_tensors="pt", padding=True, truncation=True)
tokenized_segments_input_ids, tokenized_segments_attention_mask = tokenized_segments.input_ids, tokenized_segments.attention_mask
model_predictions = F.softmax(model(input_ids=tokenized_segments_input_ids, attention_mask=tokenized_segments_attention_mask)['logits'], dim=1)

print("Positive probability: "+str(model_predictions[0][1].item()*100)+"%")
print("Negative probability: "+str(model_predictions[0][0].item()*100)+"%")

If you have any questions about how to use this model, feel free to contact the team at AssemblyAI!