| .. |
|
ext_server
|
fb9cdfa723
Fix server.cpp for the new cuda build macros
|
10 mēneši atpakaļ |
|
generate
|
0577af98f4
More parallelism on windows generate
|
10 mēneši atpakaļ |
|
llama.cpp @ 7c26775adb
|
9b5b69c00f
llm: update llama.cpp submodule to `7c26775`
|
10 mēneši atpakaļ |
|
patches
|
ce0dc33cb8
llm: patch to fix qwen 2 temporarily on nvidia (#4897)
|
10 mēneši atpakaļ |
|
filetype.go
|
d6f692ad1a
Add support for IQ1_S, IQ3_S, IQ2_S, IQ4_XS. IQ4_NL (#4322)
|
11 mēneši atpakaļ |
|
ggla.go
|
171eb040fc
simplify safetensors reading
|
11 mēneši atpakaļ |
|
ggml.go
|
6fd04ca922
Improve multi-gpu handling at the limit
|
10 mēneši atpakaļ |
|
gguf.go
|
7bdcd1da94
Revert "Merge pull request #4938 from ollama/mxyng/fix-byte-order"
|
10 mēneši atpakaļ |
|
llm.go
|
829ff87bd1
revert tokenize ffi (#4761)
|
11 mēneši atpakaļ |
|
llm_darwin_amd64.go
|
58d95cc9bd
Switch back to subprocessing for llama.cpp
|
1 gadu atpakaļ |
|
llm_darwin_arm64.go
|
58d95cc9bd
Switch back to subprocessing for llama.cpp
|
1 gadu atpakaļ |
|
llm_linux.go
|
58d95cc9bd
Switch back to subprocessing for llama.cpp
|
1 gadu atpakaļ |
|
llm_windows.go
|
058f6cd2cc
Move nested payloads to installer and zip file on windows
|
1 gadu atpakaļ |
|
memory.go
|
17df6520c8
Remove mmap related output calc logic
|
10 mēneši atpakaļ |
|
memory_test.go
|
6f351bf586
review comments and coverage
|
10 mēneši atpakaļ |
|
payload.go
|
6f351bf586
review comments and coverage
|
10 mēneši atpakaļ |
|
server.go
|
da3bf23354
Workaround gfx900 SDMA bugs
|
10 mēneši atpakaļ |
|
status.go
|
58d95cc9bd
Switch back to subprocessing for llama.cpp
|
1 gadu atpakaļ |
jmorganca
9b5b69c00f
llm: update llama.cpp submodule to `7c26775`