3 Commits

Include renames

Author	SHA1	Message	Date
Vijay Janapa Reddi	dadd105db1	feat(vol3,vol4): refine GQA per-GPU head sizing and add Physical AI empirical figures ... - Vol 3 Ch 05: Compute per-GPU KV cache token footprint directly from partitioned heads (n_kv_heads_gpu = max(1, n_kv_heads // tp_degree)) avoiding float division. - Vol 4 Ch 01, 07, 13, 16: Integrate empirical Physical AI figures and data curves (causal boundary latency vs velocity, sim-to-real proxy transfer law with SIMPLER citation, 50-year microprocessor trend & robotics-aware DSE, California DMV disengagement decade trend). - Vol 4 Ch 03: Ensure resilient path resolution for brain frequency parameters dataset during code execution. - Validator: Enforce headless Agg backend for matplotlib in code_exec check.	2026-09-22 21:30:15 +02:00
Vijay Janapa Reddi	87a953e026	fix(validator): handle SystemExit in code_exec check and enforce byte unit in Ch 09	2026-09-22 21:08:06 +02:00
Vijay Janapa Reddi	06ca2b4a4b	feat(vol3): unify agentic KV cache formulas and add live code execution validator ... - Add exec scope to binder check code and test_all_qmd_execution.py to verify live execution across all volumes - Unify per-token and aggregate KV cache calculations across Volume 3 chapters using calc_kv_cache_bytes_per_token - Standardize reference model definitions with CodingAgents.SWE_Bench_Runner in Chapter 04	2026-09-22 20:37:45 +02:00