Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
Commits on main
ggml: Support heterogeneous KV cache layer sizes in memory estimation
f66216e Jesse Gross committed 1 year ago
kvcache: Sliding window cache only needs a single batch total
1feff61 Jesse Gross committed 1 year ago
docs: add flags to example linux log output command (#9852)
5e0b904 copeland3300 committed 1 year ago
server/internal/client/ollama: fix file descriptor management in Pull (#9931)
ce92998 Blake Mizerany committed 1 year ago
benchmark: performance of running ollama server (#8643)
fb6252d Bruce MacDonald committed 1 year ago
server/internal/client/ollama: persist through chunk download errors (#9923)
c794fef Blake Mizerany committed 1 year ago
Revert "parser: remove role validation from Modelfile parser" (#9917)
00ebda8 Parth Sareen committed 1 year ago
fix: show correct bool value for kv in verbose show information (#9928)
6d11030 Patrick Devine committed 1 year ago
ollamarunner: Provide mechanism for backends to report loading progress
0ff2875 Jesse Gross committed 1 year ago
kvcache: Account for source tensors in defrag operation count
d3e9ca3 Jesse Gross committed 1 year ago
parser: remove role validation from Modelfile parser (#9874)
ffbfe83 rylativity committed 1 year ago
server/internal/client/ollama: confirm all chunksums were received (#9893)
2ddacd7 Blake Mizerany committed 1 year ago
ggml: return error on failure to read tensor data (#9872)
df94175 Bruce MacDonald committed 1 year ago
convert: return name of unsupported architecture (#9862)
61a8825 Bruce MacDonald committed 1 year ago
ollamarunner: Check for minBatch of context space when shifting
bf24498 Jesse Gross committed 1 year ago
runner: remove cache prompt flag from ollama runner (#9826)
95e271d Bruce MacDonald committed 1 year ago