AI Systems · core
AdamW Optimizer
Moments, weight decay, gradient clipping.
Mental model
AdamW is the standard optimizer for training transformers. It adapts the learning rate per parameter (Adam) and applies weight decay separately from the gradient (the "W"). The default choice unless you have a reason to switch.
How to study AdamW Optimizer
Begin by restating the mental model in your own words, then connect it to a concrete system you have built or operated. Name the mechanism, the constraint it addresses, and the trade-off it introduces. Use Decoupled Weight Decay Regularization (Loshchilov & Hutter), CS231n — Neural Networks 3: Adam and learning-rate schedules, Neural Networks: Zero to Hero (Karpathy) to check details, but close the source before writing your explanation. Retrieval is the learning step; rereading is only preparation.
Next, compare AdamW Optimizer with Training & Debugging. Ask what changes in correctness, latency, resource use, operability, and failure recovery. Complete AdamW vs L2-on-gradient and preserve the command, input, output, and one failed attempt as evidence. Finish by explaining the idea without jargon to someone who has not studied the track.
Proof of understanding
- Explain the mechanism from first principles and identify the state it reads or changes.
- Give one situation where the concept is the right choice and one where it is not.
- Predict a realistic failure mode before running the drill, then compare the prediction with evidence.
- Connect the result to a roadmap or build artifact instead of treating the concept as isolated trivia.
Learn from primary sources
Practice and explain it back
AdamW vs L2-on-gradient
Weight w=2, grad=0.5, lr=0.1, decay=0.01. Compare one step: (a) L2 in gradient w−lr·(grad+decay·w) vs (b) AdamW-style w−lr·grad then w·(1−lr·decay).
Expected evidence: L2→1.33; decoupled→1.33 then shrink (slightly different effective decay).
Open the interactive drill →Review prompts
- What does the W in AdamW change, and why did plain Adam plus L2 not do the same thing?
Build evidence
Synthesize: AI Models & Training
Move from transformer foundations through pre-training, fine-tuning, post-training, compression, and evaluation. Produce one working system, benchmark, or evidence-backed design that integrates the path.
- Implements or precisely models the core mechanisms from all three milestones
- Includes at least one injected failure or adversarial case and demonstrates recovery
- Reports quality, latency, resource, reliability, or usability measurements relevant to the domain
- Ships a concise architecture note explaining decisions, trade-offs, and remaining risks