FazBrowse GitHub Viewer | Trending | | Home

Tools: [Download Repo ZIP] [Original HTTPS Page]

History for llama_cpp - willtejeda/llama-cpp-python · GitHub

Commits

History for

Commits on Apr 13, 2023

Add field to disable reseting between generations
abetlen
committed
6595ad8
Revert "Deprecate generate method"
abetlen
committed
22fa5a6
Formatting
abetlen
committed
4f5f99e

Commits on Apr 12, 2023

Enable logprobs on completion endpoint
abetlen
committed
0daf16d
Fix threading bug. Closes #62
abetlen
committed
19598ac
Update llama.cpp
abetlen
committed
005c78d
Don't serialize stateful parameters
abetlen
committed
c854c25
Style fix
abetlen
committed
2f9b649
Deprecate generate method
abetlen
committed
6cf5876
Implement logprobs parameter for text completion. Closes #2
abetlen
committed
b3805bb

Commits on Apr 11, 2023

Update llama.cpp
abetlen
committed
9f1e565
Remove async from function signature to avoid blocking the server
abetlen
committed
213cc5c

Commits on Apr 10, 2023

Changed the environment variable name into "LLAMA_CPP_LIB"
Mug
committed
2559e5a
Make windows users happy (hopefully)
Mug
committed
ee71ce8
Better custom library debugging
Mug
committed
cf339c9
Merge branch 'main' of https://github.com/abetlen/llama-cpp-python into local-lib
Mug
committed
4132293
Use environment variable for library override
Mug
committed
76131d5
Add use_mmap option
abetlen
committed
1f67ad2
Update llama.cpp
abetlen
committed
c3c2623

Commits on Apr 8, 2023

Fix cpu count default
abetlen
committed
314ce7d
Formatting
abetlen
committed
3fbc063
Formatting
abetlen
committed
0067c1a
Bugfix: Wrong size of embeddings. Closes #47
abetlen
committed
38f442d
Update shared library extension for macos
abetlen
committed
ae3e9c3

Commits on Apr 7, 2023

Safer calculation of default n_threads
abetlen
committed
da539cc
Merge branch 'main' of github.com:abetlen/llama_cpp_python into main
abetlen
committed
930db37
Handle prompt list
abetlen
committed
55279b6

Commits on Apr 5, 2023

Set n_batch to default values and reduce thread count:
MillionthOdin16
committed
c283edd
Set n_batch to the default value of 8. I think this is leftover from when n_ctx was missing and n_batch was 2048.
MillionthOdin16
committed
76a82ba
Add server as a subpackage
abetlen
committed
44448fb
Allow local llama library usage
Mug
committed
e3ea354
Make Llama instance pickleable. Closes #27
abetlen
committed
e96a5c5
Bugfix for Python3.7
abetlen
committed
7643f66
Add runtime check to ensure embedding is enabled if trying to generate embeddings
abetlen
committed
cefc69e
Remove workaround
abetlen
committed
5c50af7

Back | FazBrowse Home | New Git URL