Open Source Python LLM Inference Tools - Page 3
Sort By:
Python LLM Inference Tools
View 131 business solutionsBrowse free open source Python LLM Inference Tools and projects below. Use the toggles on the left to filter open source Python LLM Inference Tools by OS, license, language, programming language, and project status.
-
Our Free Plans just got better! | Auth0You asked, we delivered! Auth0 is excited to expand our Free and Paid plans to include more options so you can focus on building, deploying, and scaling applications without having to worry about your security. Auth0 now, thank yourself later.
-
MongoDB Atlas runs apps anywhereMongoDB Atlas gives you the freedom to build and run modern applications anywhere—across AWS, Azure, and Google Cloud. With global availability in over 115 regions, Atlas lets you deploy close to your users, meet compliance needs, and scale with confidence across any geography.
-
1
EconML
Python Package for ML-Based Heterogeneous Treatment Effects Estimation
EconML is a Python package for estimating heterogeneous treatment effects from observational data via machine learning. This package was designed and built as part of the ALICE project at Microsoft Research with the goal of combining state-of-the-art machine learning techniques with econometrics to bring automation to complex causal inference problems. One of the biggest promises of machine learning is to automate decision-making in a multitude of domains. At the core of many data-driven personalized decision scenarios is the estimation of heterogeneous treatment effects: what is the causal effect of an intervention on an outcome of interest for a sample with a particular set of features? In a nutshell, this toolkit is designed to measure the causal effect of some treatment variable(s) T on an outcome variable Y, controlling for a set of features X, W and how does that effect vary as a function of X. -
2
FastChat
Open platform for training, serving, and evaluating language models
FastChat is an open platform for training, serving, and evaluating large language model-based chatbots. If you do not have enough memory, you can enable 8-bit compression by adding --load-8bit to the commands above. This can reduce memory usage by around half with slightly degraded model quality. It is compatible with the CPU, GPU, and Metal backend. Vicuna-13B with 8-bit compression can run on a single NVIDIA 3090/4080/T4/V100(16GB) GPU. In addition to that, you can add --cpu-offloading to commands above to offload weights that don't fit on your GPU onto the CPU memory. This requires 8-bit compression to be enabled and the bitsandbytes package to be installed, which is only available on linux operating systems. -
3
FlashInfer
FlashInfer: Kernel Library for LLM Serving
FlashInfer is a kernel library designed to enhance the serving of Large Language Models (LLMs) by optimizing inference performance. It provides a high-performance framework that integrates seamlessly with existing systems, aiming to reduce latency and improve efficiency in LLM deployments. FlashInfer supports various hardware architectures and is built to scale with the demands of production environments. -
4
GPT-NeoX
Implementation of model parallel autoregressive transformers on GPUs
This repository records EleutherAI's library for training large-scale language models on GPUs. Our current framework is based on NVIDIA's Megatron Language Model and has been augmented with techniques from DeepSpeed as well as some novel optimizations. We aim to make this repo a centralized and accessible place to gather techniques for training large-scale autoregressive language models, and accelerate research into large-scale training. For those looking for a TPU-centric codebase, we recommend Mesh Transformer JAX. If you are not looking to train models with billions of parameters from scratch, this is likely the wrong library to use. For generic inference needs, we recommend you use the Hugging Face transformers library instead which supports GPT-NeoX models. -
Zenflow- The AI Workflow Engine for Software DevsZenflow is the AI workflow engine built for real teams. Parallel agents plan, code, test, and verify in one workflow. With spec-driven development and deep context, Zenflow turns requirements into production-ready output so teams ship faster and stay in flow.
-
5
GPflow
Gaussian processes in TensorFlow
GPflow is a package for building Gaussian process models in Python. It implements modern Gaussian process inference for composable kernels and likelihoods. GPflow builds on TensorFlow 2.4+ and TensorFlow Probability for running computations, which allows fast execution on GPUs. -
6
Genv
GPU environment management and cluster orchestration
Genv is an open-source environment and cluster management system for GPUs. Genv lets you easily control, configure, monitor and enforce the GPU resources that you are using in a GPU machine or cluster. It is intended to ease up the process of GPU allocation for data scientists without code changes. -
7
Hugging Face Transformer
CPU/GPU inference server for Hugging Face transformer models
Optimize and deploy in production Hugging Face Transformer models in a single command line. At Lefebvre Dalloz we run in-production semantic search engines in the legal domain, in the non-marketing language it's a re-ranker, and we based ours on Transformer. In that setup, latency is key to providing a good user experience, and relevancy inference is done online for hundreds of snippets per user query. Most tutorials on Transformer deployment in production are built over Pytorch and FastAPI. Both are great tools but not very performant in inference. Then, if you spend some time, you can build something over ONNX Runtime and Triton inference server. You will usually get from 2X to 4X faster inference compared to vanilla Pytorch. It's cool! However, if you want the best in class performances on GPU, there is only a single possible combination: Nvidia TensorRT and Triton. You will usually get 5X faster inference compared to vanilla Pytorch. -
8
LLMFlows
LLMFlows - Simple, Explicit and Transparent LLM Apps
LLMFlows is a framework for building simple, explicit, and transparent applications utilizing Large Language Models (LLMs). It emphasizes clarity and control in the development process, allowing developers to create LLM-powered applications with well-defined workflows and interactions. LLMFlows supports various LLMs and provides tools to manage prompts, responses, and application logic effectively. -
9
LazyLLM
Easiest and laziest way for building multi-agent LLMs applications
LazyLLM is an optimized, lightweight LLM server designed for easy and fast deployment of large language models. It is fully compatible with the OpenAI API specification, enabling developers to integrate their own models into applications that normally rely on OpenAI’s endpoints. LazyLLM emphasizes low resource usage and fast inference while supporting multiple models. -
Grafana: The open and composable observability platformGrafana is the open source analytics & monitoring solution for every database.
-
10
Lightning Bolts
Toolbox of models, callbacks, and datasets for AI/ML researchers
Bolts package provides a variety of components to extend PyTorch Lightning, such as callbacks & datasets, for applied research and production. Torch ORT converts your model into an optimized ONNX graph, speeding up training & inference when using NVIDIA or AMD GPUs. We can introduce sparsity during fine-tuning with SparseML, which ultimately allows us to leverage the DeepSparse engine to see performance improvements at inference time. -
11
LitGPT
20+ high-performance LLMs with recipes to pretrain, finetune at scale
LitGPT is a collection of over 20 high-performance large language models (LLMs) accompanied by recipes to pretrain, finetune, and deploy them at scale. It provides implementations without abstractions, making it beginner-friendly while offering advanced features like flash attention and support for various precision levels. LitGPT is designed to run efficiently across multiple GPUs or TPUs, catering to both small-scale and large-scale deployments. -
12
Llama Recipes
Scripts for fine-tuning Meta Llama3 with composable FSDP & PEFT method
The 'llama-recipes' repository is a companion to the Meta Llama models. We support the latest version, Llama 3.1, in this repository. The goal is to provide a scalable library for fine-tuning Meta Llama models, along with some example scripts and notebooks to quickly get started with using the models in a variety of use-cases, including fine-tuning for domain adaptation and building LLM-based applications with Llama and other tools in the LLM ecosystem. The examples here showcase how to run Llama locally, in the cloud, and on-prem. -
13
LoRAX
Multi-LoRA inference server that scales to 1000s of fine-tuned LLMs
Lorax is a multi-LoRA (Low-Rank Adaptation) inference server that scales to thousands of fine-tuned Large Language Models (LLMs). It enables efficient deployment and management of numerous fine-tuned models, facilitating scalable AI applications. Lorax is designed to handle high concurrency and provides a robust infrastructure for serving multiple LLMs simultaneously. -
14
MMDeploy
OpenMMLab Model Deployment Framework
MMDeploy is an open-source deep learning model deployment toolset. It is a part of the OpenMMLab project. Models can be exported and run in several backends, and more will be compatible. All kinds of modules in the SDK can be extended, such as Transform for image processing, Net for Neural Network inference, Module for postprocessing and so on. Install and build your target backend. ONNX Runtime is a cross-platform inference and training accelerator compatible with many popular ML/DNN frameworks. Please read getting_started for the basic usage of MMDeploy. -
15
MMTracking
OpenMMLab Video Perception Toolbox
MMTracking is an open-source video perception toolbox by PyTorch. It is a part of OpenMMLab project. We are the first open-source toolbox that unifies versatile video perception tasks include video object detection, multiple object tracking, single object tracking and video instance segmentation. We decompose the video perception framework into different components and one can easily construct a customized method by combining different modules. MMTracking interacts with other OpenMMLab projects. It is built upon MMDetection that we can capitalize any detector only through modifying the configs. All operations run on GPUs. The training and inference speeds are faster than or comparable to other implementations. We reproduce state-of-the-art models and some of them even outperform the official implementations. -
16
Medusa
Framework for Accelerating LLM Generation with Multiple Decoding Heads
Medusa is a framework aimed at accelerating the generation capabilities of Large Language Models (LLMs) by employing multiple decoding heads. This approach allows for parallel processing during text generation, significantly enhancing throughput and reducing response times. Medusa is designed to be simple to implement and integrates with existing LLM infrastructures, making it a practical solution for scaling LLM applications. -
17
Mistral Inference
Official inference library for Mistral models
Open and portable generative AI for devs and businesses. We release open-weight models for everyone to customize and deploy where they want it. Our super-efficient model Mistral Nemo is available under Apache 2.0, while Mistral Large 2 is available through both a free non-commercial license, and a commercial license. -
18
NNCF
Neural Network Compression Framework for enhanced OpenVINO
NNCF (Neural Network Compression Framework) is an optimization toolkit for deep learning models, designed to apply quantization, pruning, and other techniques to improve inference efficiency. -
19
OpenFieldAI - AI Open Field Test Tracker
OpenFieldAI is an AI based Open Field Test Rodent Tracker
OpenFieldAI use AI-CNN to track rodents movement with pretrained OFAI models , or user could create their own model with YOLOv8 for inferencing. The software generates Centroid graph, Heat map and Line path and a spreadsheet containing all calculated parameters like - Speed - Time in and out of ROI - Distance - Entries/Exits for single/multiple pre-recorded videos or live webcam video. The ROI is assigned automatically in multiple video input , and can be manually given in single input. - For Queries/ Reporting Bugs, contact: kabeermuzammil614@gmail.com - Available on WIndows OS -
20
OpenFold
Trainable, memory-efficient, and GPU-friendly PyTorch reproduction
OpenFold carefully reproduces (almost) all of the features of the original open source inference code (v2.0.1). The sole exception is model ensembling, which fared poorly in DeepMind's own ablation testing and is being phased out in future DeepMind experiments. It is omitted here for the sake of reducing clutter. In cases where the Nature paper differs from the source, we always defer to the latter. OpenFold is trainable in full precision, half precision, or bfloat16 with or without DeepSpeed, and we've trained it from scratch, matching the performance of the original. We've publicly released model weights and our training data — some 400,000 MSAs and PDB70 template hit files — under a permissive license. Model weights are available via scripts in this repository while the MSAs are hosted by the Registry of Open Data on AWS (RODA). -
21
OpenVINO Training Extensions
Trainable models and NN optimization tools
OpenVINO™ Training Extensions provide a convenient environment to train Deep Learning models and convert them using the OpenVINO™ toolkit for optimized inference. When ote_cli is installed in the virtual environment, you can use the ote command line interface to perform various actions for templates related to the chosen task type, such as running, training, evaluating, exporting, etc. ote train trains a model (a particular model template) on a dataset and saves results in two files. ote optimize optimizes a pre-trained model using NNCF or POT depending on the model format. NNCF optimization used for trained snapshots in a framework-specific format. POT optimization used for models exported in the OpenVINO IR format. -
22
PEFT
State-of-the-art Parameter-Efficient Fine-Tuning
Parameter-Efficient Fine-Tuning (PEFT) methods enable efficient adaptation of pre-trained language models (PLMs) to various downstream applications without fine-tuning all the model's parameters. Fine-tuning large-scale PLMs is often prohibitively costly. In this regard, PEFT methods only fine-tune a small number of (extra) model parameters, thereby greatly decreasing the computational and storage costs. Recent State-of-the-Art PEFT techniques achieve performance comparable to that of full fine-tuning. -
23
Pandas Profiling
Create HTML profiling reports from pandas DataFrame objects
pandas-profiling generates profile reports from a pandas DataFrame. The pandas df.describe() function is handy yet a little basic for exploratory data analysis. pandas-profiling extends pandas DataFrame with df.profile_report(), which automatically generates a standardized univariate and multivariate report for data understanding. High correlation warnings, based on different correlation metrics (Spearman, Pearson, Kendall, Cramér’s V, Phik). Most common categories (uppercase, lowercase, separator), scripts (Latin, Cyrillic) and blocks (ASCII, Cyrilic). File sizes, creation dates, dimensions, indication of truncated images and existance of EXIF metadata. Mostly global details about the dataset (number of records, number of variables, overall missigness and duplicates, memory footprint). Comprehensive and automatic list of potential data quality issues (high correlation, skewness, uniformity, zeros, missing values, constant values, between others). -
24
Petals
Run 100B+ language models at home, BitTorrent-style
Run 100B+ language models at home, BitTorrent‑style. Run large language models like BLOOM-176B collaboratively — you load a small part of the model, then team up with people serving the other parts to run inference or fine-tuning. Single-batch inference runs at ≈ 1 sec per step (token) — up to 10x faster than offloading, enough for chatbots and other interactive apps. Parallel inference reaches hundreds of tokens/sec. Beyond classic language model APIs — you can employ any fine-tuning and sampling methods, execute custom paths through the model, or see its hidden states. You get the comforts of an API with the flexibility of PyTorch. You can also host BLOOMZ, a version of BLOOM fine-tuned to follow human instructions in the zero-shot regime — just replace bloom-petals with bloomz-petals. Petals runs large language models like BLOOM-176B collaboratively — you load a small part of the model, then team up with people serving the other parts to run inference or fine-tuning. -
25
Phi-3-MLX
Phi-3.5 for Mac: Locally-run Vision and Language Models
Phi-3-Vision-MLX is an Apple MLX (machine learning on Apple silicon) implementation of Phi-3 Vision, a lightweight multi-modal model designed for vision and language tasks. It focuses on running vision-language AI efficiently on Apple hardware like M1 and M2 chips.
