bert

🧠 Accelerate your NLP projects with the Zenith NLP Framework, a simple and efficient toolkit for building and deploying language models.

Amazon Sales Analytics full pipeline for forecasting and insights 🐙. Python, MySQL, Power BI, and Azure Blob handle generation, storage, dashboards, and forecasts.

Autism Companion: Empowering Neurodiverse Children with AI Technology Our AI platform creates personalized support for children with Autism Spectrum Disorder. Using generative AI, we develop custom social stories and conversation simulations tailored to each child's unique sensitivities

Neural Network Compression Framework for enhanced OpenVINO™ inference

🤖 MockAI is an AI-powered career interview coach that uses multimodal AI to evaluate interview performance through 📝 NLP, 🎙️ speech, and 👤 facial analysis.

Easy and lightning fast training of 🤗 Transformers on Habana Gaudi processor (HPU)

Achieved 96.3% F1-score on FakeNewsNet benchmark with <200ms inference

Save tokens with persistent code memory and structural navigation for Claude Code

Deliver financial sentiment analysis for Indian markets using AI-powered models with high accuracy and fast response for 592 NSE tickers.

Enhance BERT fine-tuning for intent classification using supervised contrastive learning, LoRA, and layer-wise learning rate decay for better accuracy and efficiency

📂 Extract, embed, cluster, and securely store Korean text from documents using BERT, enhancing research efficiency and organization.

📊 Compare few-shot text classification with DistilBERT and TF-IDF + SVM using IMDB data, analyzing performance across various sample sizes.

Production-grade NLP system for parsing Indian addresses using IndicBERTv2-CRF

An Efficient Low-Resource Bootstrap for Topic Model Refinement in Zero-Shot Settings

Trained models & code to predict toxic comments on all 3 Jigsaw Toxic Comment Challenges. Built using ⚡ Pytorch Lightning and 🤗 Transformers. For access to our API, please email us at contact@unitary.ai.

Benchmarking TF-IDF, fine-tuned Bio_ClinicalBERT, and a LoRA-tuned LLM for clinical note classification

Achieved 96.3% F1-score on FakeNewsNet benchmark with <200ms inference

Research and Materials on Hardware implementation of Transformer Model

Türkçe müşteri geri bildirimlerinden duygu analizi yapan mikroservis mimarisi — Spring Boot, FastAPI, BERTurk ve Claude API kademeli yönlendirmesiyle