William Thurston 0c942a8c00 Add tie_embeddings support and configurable logging interval ... Implement weight tying between token embeddings and lm_head to reduce parameter count. When enabled, logits are scaled by 1/√d_model, lm_head zeroing is skipped, and optimizer groups are deduplicated. Param counting uses unique parameters while Chinchilla ratio calculation adds back the would-be lm_head size for comparability. Also adds boolean flag parsing (--flag without =value) to the configurator, an auto-derived log_every interval, and minor shell script fixes. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>		2026-02-22 14:42:58 -08:00
..
base_eval.py	Enhance model tagging support in training and evaluation scripts	2025-11-10 19:45:02 -08:00
base_loss.py	Add scripts for running evaluations and training with W&B integration	2025-11-05 11:49:50 -08:00
base_train.py	Add tie_embeddings support and configurable logging interval	2026-02-22 14:42:58 -08:00
chat_cli.py	upgrading all other files to be able to use cpu/mps as well as cuda. various minor other changes ,e.g. changing max_iterations to num_iterations in sft script for consistency in naming	2025-10-20 10:15:17 -07:00
chat_eval.py	Add scripts for running evaluations and training with W&B integration	2025-11-05 11:49:50 -08:00
chat_rl.py	initial commit	2025-10-13 06:49:24 -07:00
chat_sft.py	Enhance model tagging support in training and evaluation scripts	2025-11-10 19:45:02 -08:00
chat_web.py	upgrading all other files to be able to use cpu/mps as well as cuda. various minor other changes ,e.g. changing max_iterations to num_iterations in sft script for consistency in naming	2025-10-20 10:15:17 -07:00
mid_train.py	Enhance model tagging support in training and evaluation scripts	2025-11-10 19:45:02 -08:00
tok_eval.py	initial commit	2025-10-13 06:49:24 -07:00
tok_train.py	initial commit	2025-10-13 06:49:24 -07:00