COMMITS
/ llama_cpp/llama.py January 18, 2024
A
Merge branch 'main' into batch-processing
Andrei Betlen committed
January 17, 2024
A
Re-order classes in llama.py
Andrei Betlen committed
A
Move helper classes to _internals submodule
Andrei Betlen committed
A
Move cache classes to llama_cache submodule.
Andrei Betlen committed
January 15, 2024
A
Merge branch 'main' into batch-processing
Andrei Betlen committed
A
Add split_mode option. Closes #1085
Andrei Betlen committed
P
Implement GGUF metadata KV overrides (#1011)
Phil H committed
January 10, 2024
A
Use sampling context
Andrei Betlen committed
A
Merge branch 'main' into batch-processing
Andrei Betlen committed
S
Add ability to pass in penalize_nl param (#1068)
Stephen Hankinson committed
January 5, 2024
A
Merge branch 'main' into batch-processing
Andrei Betlen committed
December 22, 2023
T
fix text_offset of multi-token characters (#1037)
twaka committed
December 18, 2023
A
fix: float32 is not JSON serializable when streaming logits.
Andrei Betlen committed
A
Fix logits are not json serializable
Andrei Betlen committed
A
Merge branch 'main' into batch-processing
Andrei Betlen committed
A
Add offload_kqv option to llama and server
Andrei Betlen committed
A
Remove unnused import
Andrei Betlen committed
K
perf: Don't convert logprobs arrays to lists (#1021)
kddubey committed
B
Bugfix: Remove f16_kv, add offload_kqv field (#1019)
Brandon Roberts committed
December 16, 2023
D
Bug fixed with n_ctx=0 (#1015)
Daniele Morotti committed
K
Fix logits_to_logprobs for 2-D and 3-D logits (#1002)
kddubey committed
December 12, 2023
T
Replace logits_to_logprobs implementation with numpy equivalent to llama.cpp (#991)
Tanner Hobson committed
A
Merge branch 'main' into batch-processing
Andrei Betlen committed
December 11, 2023
A
Remove f16_kv
Andrei Betlen committed
November 30, 2023
A
Refactor _create_completion
Andrei Betlen committed
November 29, 2023
K
Fix #891 (#952)
kddubey committed
November 28, 2023
A
Add sampling context
Andrei Betlen committed
A
Clean up llama.py into seperate modules.
Andrei Betlen committed
November 26, 2023
A
docs: Update Llama docs
Andrei Betlen committed
November 24, 2023
A
docs: Update completion and chat_completion parameter docstrings
Andrei Betlen committed
November 23, 2023
A
docs: Add Llama class example
Andrei Betlen committed
November 21, 2023
A
Fix: Add logit_bias to all completion api methods
Andrei Betlen committed
A
Add support for logit_bias outside of server api. Closes #827
Andrei Betlen committed
T
Added support for min_p (#921)
TK-Master committed
November 10, 2023
A
Fix sampling bug when logits_all=False
Andrei Betlen committed
A
Potential bugfix for eval
Andrei Betlen committed
A
Fix: default max_tokens matches openai api (16 for completion, max length for chat completion)
Andrei Betlen committed
November 8, 2023
A
Add set_seed to Llama class
Andrei Betlen committed
A
Fix destructor NoneType is not callable error
Andrei Betlen committed
A
Add JSON mode support. Closes #881
Andrei Betlen committed
A
Add seed parameter support for completion and chat_completion requests. Closes #884
Andrei Betlen committed
D
Multimodal Support (Llava 1.5) (#821)
Damian Stewart committed
November 6, 2023
A
Fix type bug
Andrei Betlen committed
A
Refactor Llama class internals
Andrei Betlen committed
November 3, 2023
A
Clean up stdout / stderr suppression
Andrei Betlen committed