OpenSource/ollama

Auteur	SHA1 Message	Date
Jesse Gross	282bfaaa95 ollamarunner: Use a separate context per multimodal input	il y a 2 mois
Jesse Gross	9679f40146 ml: Allow models to constrain inputs to a single batch	il y a 2 mois
Michael Yang	5e2e0b46b1 fix: error if image requested without vision model	il y a 2 mois
Michael Yang	63a394068c use 2d pooling	il y a 2 mois
jmorganca	11bfa62796 add trailing \n\n after <end_of_image> to match reference implementation	il y a 2 mois
jmorganca	f63e62e546 reduce kernel size, add TODO for loading from config	il y a 2 mois
jmorganca	65b0f329d1 Revert "Allow models to force a new batch"	il y a 2 mois
Jesse Gross	06007c0a18 Allow models to force a new batch	il y a 2 mois
Jesse Gross	2c40c4d35e Fix follow up images and images split across batches	il y a 2 mois
Michael Yang	6b32a2d549 compat with upstream gguf	il y a 2 mois
Michael Yang	46bb0169c4 update model	il y a 2 mois
Michael Yang	8934324b72 use fast attention	il y a 2 mois
Michael Yang	0df1800436 set non-causal attention	il y a 2 mois
Michael Yang	4b037a97dc add gemma vision encoder	il y a 2 mois
Patrick Devine	5f74d1fd47 gemma2 impl	il y a 3 mois