Skip to main content

Your Dashboard

Select a persona to get personalized recommendations from the AI ecosystem index.

Select your persona

Trending Entities

Scriptai-scripts

PyTorch Image Classification Script

by PyTorch

A Python script using PyTorch for training and evaluating image classification models. It provides a modular structure for defining datasets, models, training loops, and evaluation metrics, enabling researchers and practitioners to quickly prototype and deploy image classification solutions.

image classificationpytorchdeep learning
70B+
Scriptai-scripts

Hugging Face Transformers Training Script

by Hugging Face

The Hugging Face Transformers training script simplifies the process of training and fine-tuning transformer models for various NLP tasks. It provides a high-level API and pre-built training loops, enabling users to quickly adapt pre-trained models to their specific datasets and objectives.

transformersnlptraining
69B
Scriptai-scripts

LangChain Expression Language (LCEL) Script

by LangChain

LCEL is a declarative way to compose chains of language models and other primitives in LangChain. This script demonstrates how to use LCEL to build complex AI pipelines with features like streaming, parallel execution, and retry mechanisms, enabling developers to create robust and scalable AI applications.

langchainchainingexpression language
67B
Scriptai-scripts

Scikit-learn Model Evaluation Script

by Scikit-learn

A Python script leveraging scikit-learn to comprehensively evaluate machine learning models. It calculates various performance metrics (e.g., accuracy, precision, recall, F1-score, AUC) and generates visualizations (e.g., confusion matrices, ROC curves) to provide insights into model behavior and facilitate informed decision-making.

model evaluationscikit-learnmachine learning
66B
Scriptai-scripts

TensorFlow Model Optimization Toolkit Script

by Google

The TensorFlow Model Optimization Toolkit script provides tools and techniques to optimize TensorFlow models for deployment, including quantization, pruning, and clustering. It reduces model size and improves inference speed, making models more suitable for edge devices and resource-constrained environments.

tensorflowmodel-optimizationquantization
65B
Integrationai-integrations

Databricks Feature Store - MLflow Integration

by Databricks

The Databricks Feature Store provides a centralized repository for managing and sharing machine learning features. Its integration with MLflow enables seamless tracking of feature usage in ML models, ensuring reproducibility and simplifying model deployment workflows by automatically packaging feature dependencies.

feature-storemlopsmodel-tracking
64B
Integrationai-integrations

MLflow Databricks Integration

by Databricks

The MLflow integration with Databricks provides a managed MLflow service within the Databricks platform. It simplifies the process of tracking experiments, managing models, and deploying them to production by leveraging Databricks' scalable infrastructure and collaborative environment.

mlopsmodel trackingexperiment management
60B
Integrationai-integrations

TensorFlow Privacy

by Google

TensorFlow Privacy is a library that makes it easier to train machine learning models with differential privacy. It provides TensorFlow optimizers that implement differentially private stochastic gradient descent (DP-SGD), allowing developers to protect the privacy of training data while still achieving good model performance.

differential privacyprivacy-preserving MLtensorflow
56C+
Integrationai-integrations

Databricks Feature Store - Feast Integration

by Databricks

The Databricks Feature Store integrates with Feast, an open-source feature store, to streamline feature engineering and management for machine learning workflows. This integration allows users to define, store, and serve features consistently across training and inference, reducing data skew and improving model performance within the Databricks environment.

feature-storefeastmlops
56C+
Integrationai-integrations

Hugging Face Optimum Intel Extension

by Hugging Face / Intel

Hugging Face Optimum Intel Extension is a toolkit designed to accelerate inference and training of transformer models on Intel CPUs and GPUs. It leverages Intel's Deep Learning Boost (DL Boost) and other hardware features to optimize model performance within the Hugging Face ecosystem.

hugging faceinteloptimization
55C+
Skillcustomer-success-foundry

Upsell Identification

by AaaS

Analyzes customer usage patterns, feature adoption gaps, and engagement trends to identify accounts ready for upsell or cross-sell conversations. Produces ranked expansion opportunities with supporting evidence — which features they are actively using, which limits they are approaching, and the timing signals that indicate readiness.

customer-successupsellexpansion
42C
Skillpeople-foundry

Training Schedule Generation

by AaaS

Generates customized 30/60/90-day training schedules for new employees based on their role, department, seniority, and onboarding goals. Sequences mandatory compliance training, role-specific tool walkthroughs, and team introductions into a coherent calendar that integrates with the employee's actual availability.

onboardingtrainingscheduling
42C
Skillpeople-foundry

Upskill Recommendation

by AaaS

Generates personalized upskilling and learning path recommendations for employees based on their current skill profile, career trajectory goals, and identified flight risk signals. Matches employees to specific courses, mentorship opportunities, and internal projects that address their development gaps and increase retention probability.

learningdevelopmentupskilling
41C
Skillpeople-foundry

Skill Adjacency Detection

by AaaS

Identifies non-obvious skill adjacencies and transferable capabilities between different roles and domains. Maps how expertise in one area translates to another — for example, how a military logistics background maps to supply chain management, or how academic research skills map to product analytics. Surfaces high-potential candidates that traditional keyword-based ATS systems routinely discard.

hrskillsadjacency
41C
Skillrevenue-foundry

Velocity Prediction

by AaaS

Forecasts engineering delivery timelines using historical sprint velocity, team composition, and dependency complexity. Identifies tasks at high risk of delay before they become blockers, producing adjusted delivery estimates and capacity recommendations that inform roadmap sequencing.

product-managementvelocityforecasting
40C
Paperacademic-papers

LLaMA

by Meta

Meta's efficient open-weight model family outperforming larger closed models.

llamametaopen-weights
36D
Modelembedding-models

text-embedding-3-small

by OpenAI

OpenAI's text-embedding-3-small offers a cost-optimized embedding model producing 1536-dimensional vectors. Despite its smaller size, it outperforms the older ada-002 model on most benchmarks. It is widely used for production RAG systems where cost per embedding is a key constraint.

embeddingsopenaicost-efficient
35D
Modelembedding-models

text-embedding-3-large

by OpenAI

OpenAI's text-embedding-3-large is the highest-performance model in OpenAI's embedding lineup, producing 3072-dimensional vectors with top results on MTEB English benchmarks. It supports dimensionality reduction via the dimensions parameter, allowing developers to trade off performance against storage and compute cost.

embeddingsopenaimteb
35D
Providerllm-providers

Together AI

by Together AI

Together AI provides a cloud platform for running, fine-tuning, and deploying open-source language models. It hosts a wide catalog of models from Llama to Mistral and offers serverless inference, dedicated endpoints, and a fine-tuning pipeline. Together AI is popular among developers who want OpenAI-compatible APIs for open-weight models at competitive pricing.

inferencefine-tuningopen-source
35D
Agentagent-platforms

Windsurf

by Codeium

Windsurf (by Codeium) is an agentic IDE that introduces Cascade, an AI agent with deep awareness of the developer's actions and codebase state. Cascade can autonomously write code, run commands, fix bugs, and maintain coherent multi-step workflows across files.

coding-agentidecascade
34D
Agentagent-platforms

Replit Agent

by Replit

Replit Agent is a browser-based AI software engineer that builds, deploys, and iterates on full-stack applications in the cloud. It handles environment setup, dependency installation, debugging, and deployment within Replit's managed cloud infrastructure.

coding-agentreplitcloud
34D
Agentagent-platforms

OpenAI Codex

by OpenAI

OpenAI Codex is a cloud-based software engineering agent that runs in isolated sandboxes, executing tasks in parallel from natural language instructions. It integrates with GitHub to read repositories, write code, run tests, and create PRs autonomously.

coding-agentopenaicloud
34D
Agentagent-platforms

Kiro

by Amazon

Kiro is Amazon's spec-driven AI IDE that introduces a structured approach to agentic development. It generates product requirements, design documents, and task lists before writing code, combining planning rigor with autonomous implementation and steering capability.

coding-agentamazonspec-driven
34D
Agentagent-platforms

GitHub Copilot

by GitHub

GitHub Copilot is Microsoft's AI coding assistant integrated across GitHub, VS Code, and other IDEs. It provides inline completions, chat-based assistance, and an autonomous coding agent mode (Copilot Workspace) for planning and implementing changes across repositories.

coding-agentgithubmicrosoft
34D
Modelembedding-models

Voyage AI voyage-3-large

by Voyage AI

Voyage AI's voyage-3-large is optimized for retrieval-augmented generation in code and long-form text contexts. It consistently achieves top rankings on MTEB benchmarks, particularly for code search and domain-specific retrieval. Anthropic partners with Voyage AI for recommended embedding use with Claude-based pipelines.

embeddingsvoyage-aicode-search
34D
Modelembedding-models

Snowflake Arctic Embed

by Snowflake

Snowflake Arctic Embed is a family of open-source text embedding models optimized for enterprise retrieval tasks. The family includes multiple sizes for different performance-latency tradeoffs, all trained on high-quality curated data. Arctic Embed models achieve competitive MTEB retrieval scores and are optimized for integration with Snowflake Cortex.

embeddingssnowflakeenterprise
34D
Modelllm-models

Qwen 3.6 Plus

by Alibaba Cloud

Qwen 3.6 Plus is Alibaba's latest large-scale model in the Qwen 3 series, featuring strong multilingual capabilities across 29+ languages alongside competitive coding and math performance. It uses a mixture-of-experts architecture for efficient inference and is available through DashScope API and as open weights on Hugging Face.

qwenmultilingualmixture-of-experts
34D
Datasetdatasets

UltraChat

by Tsinghua University

1.5M high-quality multi-turn dialogue dataset for instruction fine-tuning.

alignmentdialoguesft
34D
Datasetdatasets

The Pile

by EleutherAI

825GB diverse English pretraining corpus from 22 high-quality data sources.

pretrainingenglishdiverse
34D
Datasetdatasets

SWE-bench

by Princeton NLP

2.3K real GitHub issues requiring AI agents to write and verify code fixes.

benchmarkcodingagents
34D
Datasetdatasets

StarCoder Data

by BigCode / HuggingFace

The Stack — 6.4TB permissively licensed source code across 350+ languages.

pretrainingcodebigcode
34D
Datasetdatasets

SlimOrca

by Open-Orca

Curated 518K subset of OpenOrca GPT-4 explanations optimized for fine-tuning.

alignmentsftgpt4
34D
Benchmarkbenchmarks-evaluation

SWE-bench

by Princeton NLP

SWE-bench is a benchmark for evaluating AI systems' ability to resolve real GitHub issues from popular Python repositories. Each instance requires understanding a codebase, identifying the bug, and producing a correct patch. SWE-bench Verified is the curated subset accepted as the standard for coding agent evaluation by the AI industry.

benchmarkcodingsoftware-engineering
34D
Benchmarkbenchmarks-evaluation

MTEB

by Hugging Face / MTEB Team

MTEB (Massive Text Embedding Benchmark) is the standard benchmark for evaluating text embedding models across 8 task types (retrieval, clustering, classification, etc.) and 112 datasets. The MTEB leaderboard on Hugging Face is the primary reference for selecting embedding models and is updated continuously as new models are released.

benchmarkembeddingsretrieval
34D
Benchmarkbenchmarks-evaluation

MMLU

by UC Berkeley

MMLU (Massive Multitask Language Understanding) is a comprehensive benchmark covering 57 academic subjects from elementary to professional level, including STEM, law, medicine, and social sciences. It became the standard for measuring general knowledge breadth in LLMs and is included in virtually every model evaluation suite.

benchmarkknowledgemultitask
34D
Benchmarkbenchmarks-evaluation

LiveBench

by LiveBench OSS

LiveBench is a contamination-resistant benchmark that continuously updates with new questions sourced from recent math competitions, research papers, and news. By using only data post-dating model training cutoffs, LiveBench mitigates benchmark saturation and provides more reliable capability assessments of frontier models.

benchmarkcontamination-resistantlive
34D
Benchmarkbenchmarks-evaluation

HumanEval

by OpenAI

HumanEval is OpenAI's code generation benchmark consisting of 164 hand-written Python programming problems with unit tests. It measures a model's ability to generate syntactically correct and functionally complete code from docstring descriptions. HumanEval is the foundational coding benchmark that all subsequent code benchmarks build upon.

benchmarkcodingpython
34D
Paperacademic-papers

Test-Time Compute

by Google DeepMind

Scaling inference compute via verifiers and search improves reasoning without training.

inferencereasoningscaling
34D
Paperacademic-papers

Scaling Laws

by OpenAI

Kaplan et al. power-law relationships between model size, data, compute, and loss.

scalingcomputeopenai
34D
Paperacademic-papers

ReAct

by Princeton / Google

Interleaved reasoning and acting pattern enabling LLMs to use tools iteratively.

agentsreasoningtools
34D
Paperacademic-papers

Mixtral

by Mistral AI

Mistral's sparse Mixture-of-Experts model matching GPT-3.5 at fraction of cost.

moemistralefficiency
34D
Hardwareai-hardware

Google TPU v7 Ironwood

by Google

Google's TPU v7 Ironwood is the seventh generation of Google's custom Tensor Processing Units, designed for large-scale AI inference at hyperscaler capacity. Ironwood pods target serving frontier models like Gemini at Google's internal scale and are available to cloud customers via Google Cloud's TPU v7 instances.

googletpuinference
34D
Hardwareai-hardware

Google TPU v6e Trillium

by Google

Google TPU v6e Trillium is Google's sixth-generation TPU with 4x the compute and 3x the memory bandwidth per chip compared to v5e. Trillium is generally available on Google Cloud for both training and inference workloads, offering the most cost-efficient TPU option for teams training Gemma and other open models on Google Cloud.

googletputraining
34D
Hardwareai-hardware

SambaNova SN40L RDU

by SambaNova Systems

SambaNova's SN40L is a Reconfigurable Dataflow Unit designed for high-throughput LLM inference and training. Its tiered memory architecture — combining on-chip SRAM with off-chip DRAM — allows serving multiple large models simultaneously with industry-leading batch throughput. The SN40L is the hardware underlying SambaNova Cloud's inference API.

sambanovarduinference
34D
Hardwareai-hardware

NVIDIA RTX 5090

by NVIDIA

The NVIDIA RTX 5090 is NVIDIA's flagship consumer/prosumer GPU in the Blackwell generation, featuring 32GB GDDR7 memory and massive compute for local AI inference and fine-tuning. It allows running 70B quantized models on a single consumer GPU and is the premier choice for developers who need frontier local model capability in a workstation.

nvidiablackwellconsumer-gpu
34D
Hardwareai-hardware

NVIDIA H200

by NVIDIA

The NVIDIA H200 is a Hopper-generation GPU with 141GB of HBM3e memory — nearly double the H100's bandwidth — targeting inference workloads for very large models. The additional memory enables running 70B+ parameter models on fewer GPUs, significantly reducing the cost per inference token for large-scale deployments.

nvidiahoppergpu
34D
Providerllm-providers

xAI

by xAI

xAI is Elon Musk's AI company and creator of the Grok model family. It provides API access to Grok models with real-time web search integration, available through the xAI API and X (Twitter) platform. Grok models are trained on a broad mix of web and social data and emphasize up-to-date knowledge and uncensored reasoning.

llmgrokreal-time
34D
Providergpu-compute

Vast.ai

by Vast.ai

Vast.ai is a peer-to-peer GPU marketplace connecting researchers and startups with spare GPU capacity from data centers and individuals worldwide. It offers some of the cheapest GPU rental prices on the market with flexibility to choose hardware by price, latency, or reliability score. Best suited for cost-sensitive experimentation and training runs.

gpu-cloudmarketplacepeer-to-peer
34D
Providergpu-compute

Together AI (GPU Compute)

by Together AI

Together AI's compute platform provides on-demand and reserved GPU clusters for training and fine-tuning open-source models. It offers H100 and A100 clusters with high-bandwidth networking optimized for distributed training runs, serving as both a GPU cloud provider and an inference platform. Teams use Together AI compute to run multi-node training jobs on Llama and Mistral variants.

gpu-cloudh100a100
34D
Providerllm-providers

SambaNova

by SambaNova Systems

SambaNova Systems builds custom AI hardware (Reconfigurable Dataflow Units) and offers cloud inference via SambaNova Cloud. It delivers some of the highest throughput speeds for large models including Llama 3 and Meta's frontier releases, targeting enterprises that need predictable, high-throughput inference at scale.

inferencerduhardware
34D

Discover something new