COMMITS
/ llama_cpp/llama.py July 20, 2023
A
Merge pull request #481 from c0sogi/main
Andrei committed
C
Now the last token sent when `stream=True`
Carlos Tejada committed
July 16, 2023
C
Added `RouteErrorHandler` for server
c0sogi committed
July 15, 2023
A
Re-order Llama class params
Andrei Betlen committed
A
Merge branch main into custom_rope
Andrei Betlen committed
July 9, 2023
R
Add bindings for custom_rope
randoentity committed
A
bugfix: truncate completion max_tokens to fit context length by default
Andrei Betlen committed
July 8, 2023
A
Merge branch 'main' into add_unlimited_max_tokens
Andrei committed
A
Merge pull request #453 from wu-qing-157/main
Andrei committed
A
bugfix: fix compatibility bug with openai api on last token
Andrei Betlen committed
A
perf: convert pointer to byref
Andrei Betlen committed
July 7, 2023
A
perf: avoid allocating new buffers during sampling
Andrei Betlen committed
A
perf: assign to candidates data structure instead
Andrei Betlen committed
W
fix indexing token_logprobs after sorting
wu-qing-157 committed
June 29, 2023
A
Hotfix: logits_all bug
Andrei Betlen committed
A
Load logits directly into scores buffer
Andrei Betlen committed
A
Use pre-allocated buffers to store input_ids and scores
Andrei Betlen committed
A
Free model when llama is unloaded. Closes #434
Andrei Betlen committed
June 26, 2023
A
Merge branch 'main' of github.com:abetlen/llama_cpp_python into main
Andrei Betlen committed
A
Update llama.cpp
Andrei Betlen committed
A
Update type signature
Andrei Betlen committed
A
Merge branch 'main' into fix-state-pickle
Andrei committed
June 24, 2023
S
Only concatenate after all batches are done
samfundev committed
June 23, 2023
A
Merge branch 'main' into fix-state-pickle
Andrei committed
June 17, 2023
A
Update docs. Closes #386
Andrei Betlen committed
June 16, 2023
I
Update llama.py: Added how many input tokens in ValueError exception
imaprogrammer committed
June 15, 2023
A
Add low_vram parameter
Andrei Betlen committed
June 13, 2023
O
fix: Make LLamaState pickable for disk cache
Okabintaro committed
June 10, 2023
A
Re-enable cache
Andrei Betlen committed
June 9, 2023
T
Add support for logit_bias and logit_bias_type parameters
Tanner Hobson committed
A
Temporarily disable cache until save state bug is fixed.
Andrei Betlen committed
A
Truncate max_tokens if it exceeds context length
Andrei Betlen committed
June 8, 2023
A
Fix cache implementation breaking changes
Andrei Betlen committed
June 6, 2023
A
Merge pull request #289 from Maximilian-Winter/main
Andrei committed
A
Fix resize issue. Closes #330
Andrei Betlen committed
May 31, 2023
M
Added both LlamaChache classes Disk and RAM.
Maximilian-Winter committed
M
Merge branch 'abetlen:main' into main
Maximilian Winter committed
May 28, 2023
M
Diskcache implementation for llama state.
Maximilian-Winter committed
May 27, 2023
A
Align dtype to match c structs
Andrei Betlen committed
A
Merge branch 'main' into add-numpy-support
Andrei Betlen committed
A
Fix stop sequence performance bug.
Andrei Betlen committed
A
Remove usage of eval_tokens for cache check
Andrei Betlen committed
A
Replace eval_logits and eval_tokens with numpy arrays
Andrei Betlen committed
May 26, 2023
A
Add support for numpy
Andrei Betlen committed
A
Bugfix for logits_processor and stopping_criteria
Andrei Betlen committed
A
Add extra logits_processor and stopping_criteria
Andrei Betlen committed
A
Fix streaming hang on last token when cache is on.
Andrei Betlen committed