Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / abetlen/llama-cpp-python issues and pull requests

#105 - llama-cpp-python (vicuna 13b) producing extremely poor embeddings with chromadb/langchain

Issue - State: closed - Opened by khimaros over 1 year ago - 6 comments
Labels: performance

#104 - AVX / AVX512 / SSE3 optimizations not detected

Issue - State: closed - Opened by tye-singwa over 1 year ago - 3 comments

#103 - Count/truncate number of tokens before processing

Issue - State: open - Opened by jakvb over 1 year ago - 7 comments
Labels: enhancement, question

#102 - (windows) model doesn't generate nothing and stays running

Issue - State: closed - Opened by fcalabrow over 1 year ago - 2 comments
Labels: bug, windows

#101 - No cuBLAS

Issue - State: closed - Opened by Priestru over 1 year ago - 34 comments
Labels: windows, oobabooga

#100 - Compatibility issues with Chinese and slow response speed.

Issue - State: closed - Opened by B1gM8c over 1 year ago - 5 comments
Labels: bug

#99 - how to run llama.py/llama_ccp.py?

Issue - State: closed - Opened by sussyboiiii over 1 year ago - 6 comments

#98 - Improved unit tests

Pull Request - State: closed - Opened by diangamichael over 1 year ago

#97 - 如何动态去设置temperature，top_p和repeat_penalty

Issue - State: closed - Opened by yuxuan2015 over 1 year ago - 1 comment
Labels: enhancement

#96 - CDLL llama_init_from_file function

Issue - State: closed - Opened by Holpak over 1 year ago - 2 comments
Labels: model, windows

#95 - Cache Feature Request

Issue - State: closed - Opened by snxraven over 1 year ago - 6 comments

#94 - import llama_cpp ERROR

Issue - State: closed - Opened by 823863429 over 1 year ago - 18 comments
Labels: bug, build

#93 - Mikes

Pull Request - State: closed - Opened by diangamichael over 1 year ago

#92 - llama_tokenize: too many tokens

Issue - State: closed - Opened by ylchin over 1 year ago - 2 comments
Labels: model, quality

#91 - Accessing the low-level API

Issue - State: closed - Opened by andersRuge over 1 year ago - 2 comments

#90 - [Request] Test LoRA API

Issue - State: closed - Opened by abetlen over 1 year ago - 5 comments
Labels: help wanted

#89 - Default value for the number of threads

Issue - State: closed - Opened by avdosev over 1 year ago - 5 comments
Labels: hardware

#88 - Add LoRA support for low level API

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: enhancement

#87 - Fix TypeError in low_level chat

Pull Request - State: closed - Opened by SagsMug over 1 year ago

#86 - Performance Improvement and Error Handling for API

Issue - State: closed - Opened by djaffer over 1 year ago - 1 comment

#85 - Server with LangChain and Streaming throws JSONDecodeError

Issue - State: closed - Opened by ShreyBiswas over 1 year ago - 2 comments
Labels: bug, openai

#84 - Error running on huggingface space with dockerfile

Issue - State: closed - Opened by djaffer over 1 year ago

#83 - Improved unit tests

Pull Request - State: closed - Opened by diangamichael over 1 year ago - 1 comment

#82 - Changing parameters once Llama is initialized makes it not consistent

Issue - State: closed - Opened by adriacabeza over 1 year ago - 3 comments
Labels: bug

#81 - [Windows] Exception when trying to load a model

Issue - State: closed - Opened by Holpak over 1 year ago - 5 comments
Labels: windows

#80 - Openplayground Suport

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: bug, enhancement, server, openai

#78 - Issue template proposal

Issue - State: closed - Opened by gjmulder over 1 year ago - 4 comments

#76 - Create bugs/performance issue tracking template

Pull Request - State: closed - Opened by MillionthOdin16 over 1 year ago - 1 comment
Labels: enhancement

#74 - Create a GitHub ISSUE_TEMPLATE for Bugs / Performance Regressions

Issue - State: closed - Opened by abetlen over 1 year ago - 1 comment
Labels: documentation, good first issue, high-priority

#73 - Add Dockerfile + build workflow

Pull Request - State: closed - Opened by Niek over 1 year ago - 11 comments

#72 - Getting `Illegal instruction (core dumped)` when running an openblas made install

Issue - State: closed - Opened by Bloob-beep over 1 year ago - 6 comments
Labels: build, hardware

#71 - Why is the latest version 2x slower?

Issue - State: closed - Opened by Bloob-beep over 1 year ago - 7 comments
Labels: performance

#70 - Problems when i try to use this inside the default python 3.10 docker container

Issue - State: closed - Opened by LLukas22 over 1 year ago - 28 comments

#69 - Improve error message when model file is not found

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: documentation, enhancement

#68 - Implement chat continuation

Issue - State: closed - Opened by abetlen over 1 year ago - 8 comments
Labels: enhancement, high-priority

#67 - Text generation only one line with Vicuna

Issue - State: closed - Opened by JeroenAdam over 1 year ago - 5 comments

#66 - kv model size is twice large then llama.cpp/main

Issue - State: closed - Opened by hengjiUSTC over 1 year ago - 6 comments

#65 - AttributeError occured

Issue - State: closed - Opened by slucky001 over 1 year ago - 1 comment

#64 - Add unlimited max_tokens

Pull Request - State: closed - Opened by jm12138 over 1 year ago - 4 comments
Labels: enhancement

#63 - Low level chat: Added iterative search to prevent instructions from being echoed

Pull Request - State: closed - Opened by SagsMug over 1 year ago - 1 comment

#62 - Thread bug in server code

Issue - State: closed - Opened by hengjiUSTC over 1 year ago - 4 comments
Labels: bug

#61 - Add UTF-8 Encoding in read_text.

Pull Request - State: closed - Opened by jm12138 over 1 year ago
Labels: windows

#60 - make: *** No rule to make target `libllama.so'. Stop.

Issue - State: closed - Opened by streamnsight over 1 year ago - 7 comments

#59 - Chat does not remember initial prompt as well as llama.cpp

Issue - State: closed - Opened by vashat over 1 year ago

#58 - Text generation stops prematurely

Issue - State: closed - Opened by vashat over 1 year ago - 2 comments

#57 - Issue with emoji decoding in steaming mode, only

Issue - State: closed - Opened by CyberTimon over 1 year ago - 2 comments
Labels: bug, quality

#56 - Long duration until generation starts with big context

Issue - State: closed - Opened by CyberTimon over 1 year ago

#55 - Bugfix: Fix broken: UnicodeDecodeError: 'utf-8' codec can't decode

Pull Request - State: closed - Opened by riverzhou over 1 year ago - 14 comments
Labels: bug

#54 - [Windows] [Windows] "Failed building wheel for llama-cpp-python

Issue - State: closed - Opened by kerbi over 1 year ago - 8 comments
Labels: bug, build, windows

#53 - [WinError 193] When trying to run the high level API example with vicuna

Issue - State: closed - Opened by ChilliSawse over 1 year ago - 9 comments
Labels: build, model, windows

#50 - Add abetlen's wheel to new /Wheels/ dir (removed from git ignore) + add amd64 c310 wheel

Pull Request - State: closed - Opened by Loufe over 1 year ago - 4 comments
Labels: windows, oobabooga

#49 - Incredibly slow response time

Issue - State: closed - Opened by MartinPJB over 1 year ago - 35 comments

#48 - Running server gives error when using huggingface model

Issue - State: closed - Opened by djaffer over 1 year ago - 6 comments
Labels: model

#47 - Wrong size of embeddings

Issue - State: closed - Opened by djpadbit over 1 year ago - 1 comment
Labels: bug

#46 - simple python api example?

Issue - State: closed - Opened by maschinenzeitmaschine over 1 year ago - 7 comments

#45 - Installation on Windows failed in building wheel,UnicodeDecodeError

Issue - State: closed - Opened by jeff31415 over 1 year ago - 3 comments
Labels: windows

#44 - Implement caching for evaluated prompts

Issue - State: closed - Opened by abetlen over 1 year ago - 40 comments
Labels: enhancement, high-priority

#42 - Use https for the gitmodule instead of ssh

Pull Request - State: closed - Opened by nsarrazin over 1 year ago - 4 comments

#41 - Fix v1/chat/completions Gibberish API Responses

Pull Request - State: closed - Opened by keldenl over 1 year ago - 11 comments
Labels: bug

#40 - Installation on Windows failed because Visual Studio is not installed

Issue - State: closed - Opened by oobabooga over 1 year ago - 45 comments
Labels: oobabooga

#39 - Investigate model aliasing

Issue - State: closed - Opened by abetlen over 1 year ago - 3 comments
Labels: enhancement, server

#38 - Add `/models/{model}` endpoint

Issue - State: closed - Opened by abetlen over 1 year ago - 1 comment
Labels: enhancement, server

#37 - Add cli options to server

Issue - State: closed - Opened by abetlen over 1 year ago

#36 - Fix unicode decoding error

Issue - State: closed - Opened by abetlen over 1 year ago - 2 comments
Labels: bug

#35 - Test other package managers in Github workflow

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: enhancement, build

#34 - Test other operating systems in Github workflow

Issue - State: closed - Opened by abetlen over 1 year ago

#33 - Why is only the MSVC compiler required to build wheels?

Issue - State: closed - Opened by HCBlackFox over 1 year ago - 2 comments

#32 - Kudos on a great job! Need a little help with BLAS

Issue - State: closed - Opened by regstuff over 1 year ago - 9 comments

#31 - hi, I need help:

Issue - State: closed - Opened by elpewpew over 1 year ago - 4 comments

#30 - Shared library with base name 'llama' not found, windows

Issue - State: closed - Opened by Mohamedballouch over 1 year ago - 16 comments
Labels: bug

#29 - Fixes and Tweaks to Defaults

Pull Request - State: closed - Opened by MillionthOdin16 over 1 year ago - 3 comments

#28 - Allow local llama library usage

Pull Request - State: closed - Opened by SagsMug over 1 year ago - 2 comments

#27 - Support pickling the `Llama` instance

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: bug

#26 - Add performance optimization example

Issue - State: closed - Opened by abetlen over 1 year ago - 6 comments
Labels: documentation, enhancement

#24 - Add tests

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: enhancement

#22 - Interactive mode with Llama class

Issue - State: closed - Opened by abetlen over 1 year ago - 4 comments
Labels: enhancement

#21 - Standalone Server

Issue - State: closed - Opened by abetlen over 1 year ago - 9 comments

#20 - Investigate using `make` instead of `cmake` to build shared library

Issue - State: closed - Opened by abetlen over 1 year ago - 3 comments
Labels: help wanted, good first issue

#19 - Add Verbose Logging Support to Diagnose Performance Issues

Issue - State: closed - Opened by luminalle over 1 year ago - 6 comments
Labels: enhancement

#17 - [Feature] Dynamic Model Loading and Model Endpoint in FastAPI

Issue - State: closed - Opened by MillionthOdin16 over 1 year ago - 6 comments
Labels: enhancement, server, high-priority

#16 - AttributeError: 'Llama' object has no attribute 'embed'

Issue - State: closed - Opened by BayramAnnakov over 1 year ago - 2 comments
Labels: bug

#15 - llama.cpp chat example implementation

Pull Request - State: closed - Opened by SagsMug over 1 year ago - 5 comments

#14 - [Question] How to use kv cache?

Issue - State: closed - Opened by SagsMug over 1 year ago - 4 comments

#13 - [Question] Purpose of completion ID field

Issue - State: closed - Opened by MillionthOdin16 over 1 year ago - 1 comment

#12 - [Investigate] Custom `llama.dll` Dependency Resolution Issues on Windows

Issue - State: closed - Opened by MillionthOdin16 over 1 year ago - 8 comments

#11 - Possible Typo in fastapi_server.py: n_ctx vs n_batch

Issue - State: closed - Opened by MillionthOdin16 over 1 year ago - 1 comment
Labels: bug, documentation

#10 - Improve Shared Library Loading Mechanism

Pull Request - State: closed - Opened by MillionthOdin16 over 1 year ago - 2 comments

#9 - [Question] Drop in replacement for OpenAI

Issue - State: closed - Opened by MillionthOdin16 over 1 year ago - 19 comments

#8 - Add additional tests for `stop` sequences

Issue - State: closed - Opened by abetlen over 1 year ago - 5 comments
Labels: enhancement

#7 - Interactive mode/Session

Issue - State: closed - Opened by BlackLotus over 1 year ago - 7 comments
Labels: documentation, enhancement

#6 - Unexpected output

Issue - State: closed - Opened by jessejohnson over 1 year ago - 10 comments
Labels: bug

#5 - Error running on M1 Mac

Issue - State: closed - Opened by DrBenjamin over 1 year ago - 24 comments
Labels: bug

#4 - Add support for get embeddings

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: enhancement

#3 - Add support for tokenize / detokenize

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: enhancement

#2 - Implement `logprobs`

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: enhancement

#1 - Implement `stream` option in high-level api

Issue - State: closed - Opened by abetlen over 1 year ago
Labels: enhancement