{ "item_title" : "The Hugging Face Evolution", "item_author" : [" Merritt Alden "], "item_description" : "Turn Hugging Face models into practical AI applications you can inspect, test, and improve.The Hugging Face Evolution guides you through building and fine-tuning agentic models, exploring ShadowPEFT, and designing compact AI architectures. Across ten chapters, it connects model development with the engineering decisions that shape useful applications: data preparation, controlled tool access, retrieval, inference, evaluation, and deployment.Begin with the Hugging Face ecosystem and progress toward complete Python projects. Explore training pipelines, instruction tuning, and parameter-efficient adaptation with LoRA, QLoRA, and DoRA. Learn how prompts, tokenization, supervision, and model artifacts fit together-and why a successful training run still requires careful evaluation.Discover how ShadowPEFT introduces an accompanying shadow network around a frozen foundation model. Examine attached and standalone prediction paths, checkpoint export, and independent evaluation. Explore knowledge distillation, quantization, structured pruning, and adaptive routing to understand when smaller models can meet an application's requirements.Build agents that retrieve evidence, maintain bounded conversational context, and use tools through explicit application controls. Practical examples connect Hugging Face embeddings, FAISS, and smolagents while addressing source attribution, argument validation, authorization, and execution limits. Follow the distinction between generating a plausible response and completing a task correctly.Move into model serving with FastAPI, request validation, authentication, health checks, and performance measurement. Learn to interpret latency, throughput, and memory results, detect behavioral regressions, and organize versioned releases with evaluation gates and rollback procedures.The final integrated project brings training, retrieval, model routing, bounded agent workflows, and API serving into one application. Supporting appendices provide shared helpers, dependencies, teaching fixtures, and a clear account of what the local verification tested.Complete code listings accompany the explanations, making the implementation available for close inspection. The examples state their assumptions and evaluation limits, helping you adapt each workflow to your own data, hardware, and operational needs.Written for Python developers, AI engineers, and technically minded readers, this book offers a practical foundation for moving from individual model experiments toward AI systems with clearer contracts, measurable behavior, and controlled releases. It gives you concrete workflows to study, reproduce, and extend as you develop your own applications.", "item_img_path" : "https://covers4.booksamillion.com/covers/bam/9/79/816/175/9798161756775_b.jpg", "price_data" : { "retail_price" : "17.99", "online_price" : "17.99", "our_price" : "17.99", "club_price" : "17.99", "savings_pct" : "0", "savings_amt" : "0.00", "club_savings_pct" : "0", "club_savings_amt" : "0.00", "discount_pct" : "10", "store_price" : "" } }

Overview

Customers Also Bought

Details