diff options
| author | tslil <tslil@posteo.de> | 2026-09-11 12:56:12 +0100 |
|---|---|---|
| committer | tslil <tslil@posteo.de> | 2026-09-11 17:00:18 +0100 |
| commit | 6e90f1145b2c5c0f2dcc82d21ec45432274b8cad (patch) | |
| tree | 18db1c03ea7c3c4724028fc5154d4bca3d7200a5 /Makefile | |
| parent | 220a4b79e4b6d742a9ca7c278a1962b870500fbc (diff) | |
move from single candidate td-leaf with gating to tournament of a population
The goal in this change was to overcome the rock-paper-scissors phenomenon that was happening in capability space by having on a single candidate learn to beat the most recent version of itself. In some sense this lead to rotating around in capability space instead of chasing monotonic improvements.
I hope that flipping to a system in which many (N=5) candidates play each other and learn from their own set of diverse games in this pool, combined with --replay-iters, will lead to a given network having to have a much more comprehensive understanding of the space of tactics. I haven't yet had time to train this thorougly, will run those experiments on my desktop computer.
Diffstat (limited to 'Makefile')
| -rw-r--r-- | Makefile | 27 |
1 files changed, 17 insertions, 10 deletions
@@ -4,7 +4,7 @@ NATIVE_FLAGS=-march=native CTAK=include/tak.c CTAK_OBJS=$(CTAK:.c=.o) -CAND_SRCS=include/cand_weights_5.c include/cand_weights_6.c +CAND_SRCS=$(wildcard include/cand_*.c) SRCS=$(filter-out include/lcdlib.c $(CAND_SRCS),$(wildcard include/*.c)) OBJS=$(SRCS:.c=.o) @@ -27,24 +27,31 @@ cttei: src/cttei.o $(OBJS) $(CC) $(CFLAGS) src/cttei.o $(OBJS) -o cttei clean: - rm -f ctlm ct1986 cttei pptdb geminict selfplay match + rm -f ctlm ct1986 cttei pptdb geminict rm -f src/*.o include/*.o native: CFLAGS += $(NATIVE_FLAGS) native: ctlm cttei geminict -.PHONY: selfplay -selfplay: +cand_a = $(word 1,$(subst -, ,$1)) +cand_b = $(word 2,$(subst -, ,$1)) + +.PHONY: selfplay-% +selfplay-%: + rm -f selfplay-% $(MAKE) clean - $(MAKE) DEFINES=-DRECORD_PV selfplay-bin + $(MAKE) SELFPLAY_OUT=$@ \ + DEFINES="-DRECORD_PV -DCAND_A=$(call cand_a,$*) -DCAND_B=$(call cand_b,$*) \ + -DCAND_A_WEIGHTS_5='\"cand_$(call cand_a,$*)_weights_5.c\"' \ + -DCAND_B_WEIGHTS_5='\"cand_$(call cand_b,$*)_weights_5.c\"' \ + -DCAND_A_WEIGHTS_6='\"cand_$(call cand_a,$*)_weights_6.c\"' \ + -DCAND_B_WEIGHTS_6='\"cand_$(call cand_b,$*)_weights_6.c\"'" \ + selfplay-bin +SELFPLAY_OUT ?= selfplay selfplay-bin: CFLAGS += $(NATIVE_FLAGS) selfplay-bin: src/selfplay.o $(OBJS) - $(CC) $(CFLAGS) src/selfplay.o $(OBJS) -o selfplay - -match: CFLAGS += $(NATIVE_FLAGS) -match: src/match.o $(OBJS) - $(CC) $(CFLAGS) src/match.o $(OBJS) -o match + $(CC) $(CFLAGS) src/selfplay.o $(OBJS) -o $(SELFPLAY_OUT) $(CROSS_CC): ./resources/do_buildroot.sh $(BUILDROOT_DIR) |