4 Commits

Author	SHA1	Message	Date
Wing Lian	fec0c3a99e	chore: lint	2026-03-19 07:27:23 +00:00
Wing Lian	31d8d068bb	handle base+lora split kernel for older moe models	2026-03-19 07:11:30 +00:00
Wing Lian	66fea258c7	add correctness unit tests and benchmarks for scattermoe + lora	2026-03-19 06:40:04 +00:00
Wing Lian	163bd4dd5a	use custom triton kernels for entropy from logits and selective softmax (#3510) ... * use custom triton kernels for entropy from logits and selective softmax * PR comments fixes * fix out of bounds, include tests, include benchmarks * chore: lint	2026-03-19 02:02:43 -04:00