ollama

Commit Graph

Author	SHA1	Message	Date
Michael Yang	971d62595a	fix: qwen2.5 vl rope (#13486 ) * qwen25vl: bump max pixels * qwen25vl: mrope fix qwen2.5vl window * qwen25vl: vision rope	2025-12-15 17:30:33 -08:00
Michael Yang	603ceefaa6	refactor rope change to a flatter directory structure and group the options with the function update models to call rope in one place	2025-12-08 14:42:22 -08:00
Michael Yang	564b558c92	fix(llama): other llama flavours (#12308 ) * fix(llama): rope scale * spm llama * skip moe models * cleanup	2025-09-17 12:12:21 -07:00
Michael Yang	ad95d5b30b	use split activations when possible (#12293 ) * use ggml__split activations when possible forward qkv	2025-09-16 09:51:19 -07:00
Michael Yang	9ed8bf14cb	ml: add more rope options (#10775 )	2025-05-20 15:51:08 -07:00
Jesse Gross	3c14461d5d	ollamarunner: Separate text and multimodal graphs For some multimodal models (such as gemma3), we create a single graph that generates the image embedding and then use this in the text model. The embedding tensor is completely opaque to the runner. However, this doesn't work if we need to use the embedding in multiple batches. This can arise if the embedding is larger than the batch size. In these cases (as with llama4), we would like to create views that are more appropriately sized. However, if we do this then the original source tensor is used in multiple graphs, which isn't allowed. To avoid that problem, models with this pattern compute the embedding tensor on first use and recreate the individual views. There is no longer a single vision and text graph. This codifies the pattern of separating vision and text graphs. The logic of computing tensors on demand is moved to the runner, so models no longer have to worry about this. It also gives the runner visibility into the multimodal tensors, which is important for memory management.	2025-05-15 13:46:20 -07:00
Bruce MacDonald	0aa8b371dd	model: add Qwen2.5-VL support (#10385 )	2025-05-13 20:58:02 -07:00

7 Commits