8 Commits

Author	SHA1	Message	Date
Jeffrey Morgan	3490e9590b	model/qwen3next: avoid crash in in DeltaNet when offloading (#14541) ... Co-authored-by: Yossi Ovadia <jabadia@gmail.com>	2026-03-01 18:44:04 -08:00
Jeffrey Morgan	8da09b1e7e	qwen3next: add compatibility with imported GGUF models (#14517)	2026-02-28 14:21:42 -08:00
Jeffrey Morgan	7f9efd53df	model: add support for qwen3.5-27b model (#14415)	2026-02-25 01:09:58 -08:00
Jeffrey Morgan	da70c3222e	model: support for qwen3.5 architecture (#14378)	2026-02-24 20:08:05 -08:00
Michael Yang	f1373193dc	move tokenizers to separate package (#13825)	2026-02-05 17:44:11 -08:00
Jeffrey Morgan	d25535c3f3	qwen3next: avoid inplace sigmoid for shared gate (#14077)	2026-02-04 15:50:02 -08:00
Jeffrey Morgan	255579aaa7	qwen3next: fix issue in delta net (#14075) ... gDiffExp was being broadcast across the wrong axis when multiplying with k. This fix reshapes gDiffExp to [1, chunkSize, nChunks, ...]	2026-02-04 13:40:38 -08:00
Jeffrey Morgan	77eb2ca619	model: add qwen3-next architecture (#14051)	2026-02-03 23:27:21 -08:00