Utsavtiwari·Mar 11, 2025Let's chase the MambaThe evolution of LLMs was almost impossible without Transformers, and people said ‘Attention is all you need’. But is it really true, what…
Utsavtiwari·Mar 11, 2025Mamba-2: The ‘Transform’ation of MambaTransformers gained popularity due to their ability to utilize modern hardware like GPUs and NPUs efficiently. Unlike sequential models…
Utsavtiwari·Aug 12, 2024Introduction to Llama2 : Part-2 SamplersDeep Neural networks are nothing but mathematical models in which output depends on the calculations of inputs with parameters. So, for…
Utsavtiwari·Jan 15, 2024Introduction to Llama2 : Part-1 Architectural AnalysisLlama is a decoder-only language model, which takes the input sentence as ordered tokens and predicts the next token. In this article, we…A response icon2A response icon2