Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
Commits on main
api: implement stringer for ToolFunctionParameters (#12038)
d3450dd Jeffrey Morgan committed 1 year ago
tools: avoid matching braces that are part of tool content (#12039)
4bcb04a Jeffrey Morgan committed 1 year ago
Merge pull request #12021 from ollama/drifkin/thinking-double-emit
e3d5708 Devon Rifkin committed 1 year ago
server: skip parsing initial <think> if provided in the prompt (#12024)
4be4dc8 Jeffrey Morgan committed 1 year ago
kvcache: Use Cast instead of Copy for flash attention masks
05ccb17 Jesse Gross committed 1 year ago
Merge pull request #11910 from ollama/drifkin/harmony-fn-names
b517bb1 Devon Rifkin committed 1 year ago
Merge pull request #11875 from ollama/drifkin/print-template
ec8bf5e Devon Rifkin committed 1 year ago
readme: add Serene Pub to community integrations (#11946)
abeec24 Jody Doolittle committed 1 year ago
cli: show the default context length env setting in online help (#11928)
026bc29 Patrick Devine committed 1 year ago
docs: added missing comma in 'Ollama's Javascript library'' (#11915)
883d031 Thomas Pelster committed 1 year ago
test: improve scheduler/concurrency stress tests (#11906)
d6f7233 Daniel Hiltgen committed 1 year ago
server: add debug option for printing out prompt instead of calling model
8de1da4 Devon Rifkin committed 1 year ago
Revert "cuda: leverage JIT for smaller footprint (#11635)" (#11913)
d925b53 Daniel Hiltgen committed 1 year ago