knowledgator/Qwen-encoder-0.5B

The knowledgator/Qwen-encoder-0.5B is a 0.5 billion parameter bi-directional text encoder model, adapted from Qwen2-0.5B by McGill-NLP's LLM2Vec recipe. It enables bidirectional attention and masked next token prediction, leveraging modern decoder architectures for enhanced context understanding. This model is specifically designed for discriminative NLP tasks such as text classification, question answering, and token classification, offering improved performance over classical encoders.