Results for “megatron-lm”

Provides battle-tested PyTorch training recipes for LLMs, vision, diffusion, and biomedical domains, covering training loops, optimizer selection, LR scheduling, mixed precision, and debugging.

Implements production LLM operations: RAG pipelines, embeddings, vector databases, fine-tuning, advanced prompt engineering, cost estimation, quality evals, semantic caching, streaming, and agents.

Provides guidance and code for production AI workflows including RAG pipelines, vector databases, embedding indexing, prompt engineering, cost estimation, semantic caching, and quality evaluation.