Tag
Read 6 articles on Llm Training by Ashish Srivastava — practical guides covering implementation patterns, security, and production lessons.
6 posts
Learn how to implement effective dataset versioning and lineage tracking with DVC to maintain reproducibility and provenance for large language model training.
Part 5 · Dataset Engineering for LLMsLearn how to identify and remove duplicate data and leaked benchmarks to ensure your LLM training data is clean, distinct, and reliable.
Part 3 · Dataset Engineering for LLMsLearn how to generate synthetic data without triggering model collapse, ensuring diversity and robustness in your LLM training loops.
Part 2 · Dataset Engineering for LLMsWhy high-quality, curated datasets outperform massive raw collections. A look at filtering techniques and the LIMA approach for efficient model training.
Part 1 · Dataset Engineering for LLMsUnderstand the mechanics of RLHF, from collecting human preferences to training reward models and fine-tuning LLMs with PPO and KL constraints.
Part 2 · Post-Training & AlignmentCompare the memory requirements of full fine-tuning and PEFT. We break down the math on optimizer states, activation checkpointing, and GPU utilization.
Part 2 · Fine-Tuning LLMs Hands-On
