PinnedMukul Ranjan·Aug 28, 2024Combining Pruning and Knowledge DistillationEfficient LLM Part 3: Paper Summary of the paper “Compact Language Models via Pruning and Knowledge Distillation”
PinnedMukul Ranjan·Jun 30, 2022How does Lasso regression(L1) encourage zero coefficients but not the L2?We often read almost everywhere that Lasso regression encourages zero coefficient, and hence provide a great tool for variable selection as…A response icon1A response icon1
PinnedMukul Ranjan·Jan 18, 2025Understanding Tokenizers in LLM — Part 1 : Byte Pair Encoding and BeyondIf you have seen Karpathy Sensei’s video[1], you must be aware of the fact that tokenizer is one of the most “problematically important”…
PinnedMukul Ranjan·Jun 30, 2025Understanding the Pruning of Large Language ModelsEfficient LLM Part 1: Fundamental of Pruning
Mukul Ranjan·Jan 16Serving Thousands of Concurrent LoRA Adapters[Paper Review] S-LoRA: Serving Thousands of Concurrent LoRA Adapters
Mukul Ranjan·Jan 14Register Buffer and Register Parameter in PyTorchWhen building neural networks in PyTorch, we often need to store tensors that aren’t meant to be trained. The most common examples being…
Mukul Ranjan·Jan 4Complete Sony Camera Guide: Every Series ExplainedAcknowledgement: Most content of this blog is based on this youtube video[1] by Jason Vong.
Mukul Ranjan·Jan 4How Cars Work: Complete Beginner’s Guide to Car FundamentalsCars 101: Learning how to talk carA response icon2A response icon2
Mukul Ranjan·Aug 9, 2025Understanding PerplexityThis blog is extracted from my previous blog (Combining Pruning and Knowledge Distillation).