EXPLORE BY TASK
AI Tools
27 tools in this collection. Choose a task below or search the complete FeuTex toolkit.
Browse all tools
Adam Training State Memory
Account for parameter, gradient, master-weight and two optimizer-moment arrays.
Open tool →AI InfrastructureAI API Cost Comparison
Compare providers with separate uncached input, cached input, output prices and a batch discount factor.
Open tool →AI InfrastructureAI Workflow Cost Planner
Sum model calls across a multi-step workflow and project monthly token cost.
Open tool →AI InfrastructureAttention Score Tensor Memory
Estimate explicit dense attention score storage before memory-efficient attention.
Open tool →AI InfrastructureEmbedding API Batch Plan
Estimate request count under per-request token and item limits.
Open tool →AI InfrastructureEmbeddings Storage Estimator
Plan embeddings storage estimator with transparent inputs, a detailed result and clearly stated assumptions.
Open tool →AI InfrastructureGradient Accumulation Effective Batch
Calculate effective global batch and optimizer steps from microbatches and data parallel workers.
Open tool →AI InfrastructureHNSW Base Layer Link Memory
Budget base-layer neighbor identifiers separately from vector payloads.
Open tool →AI InfrastructureLLM Context Window Budget Planner
Plan llm context window budget planner with transparent inputs, a detailed result and clearly stated assumptions.
Open tool →AI InfrastructureLLM KV Cache Memory
Estimate key and value cache memory from layers, KV heads, head width and live tokens.
Open tool →AI InfrastructureLLM Prefill and Decode Latency
Estimate request latency from measured prompt throughput and output-token rate.
Open tool →AI InfrastructureLLM Prefix Cache Token Savings
Estimate repeated prefill tokens avoided by caching an identical shared prefix.
Open tool →AI InfrastructureLLM Weight Memory Estimator
Plan llm weight memory estimator with transparent inputs, a detailed result and clearly stated assumptions.
Open tool →AI InfrastructureLoRA Adapter Parameter Count
Count trainable low-rank matrices for repeated linear projections.
Open tool →AI InfrastructureMixture of Experts Active Weights
Compare stored parameters with per-token active expert parameters.
Open tool →AI InfrastructurePipeline Parallel Bubble Fraction
Estimate fill and drain overhead for a simple pipeline of equal stages.
Open tool →AI InfrastructurePrompt Brief Builder
Build a structured task brief with audience, context, constraints, output format and success criteria.
Open tool →AI InfrastructureQuantized Model Scale Overhead
Include per-group scale metadata in low-bit weight storage.
Open tool →AI InfrastructureRAG Retrieved Context Packing
Budget retrieved chunks after reserving instructions, user question and response.
Open tool →AI InfrastructureRing All Reduce Traffic
Estimate bytes sent per worker and bandwidth-only duration for ring all-reduce.
Open tool →AI InfrastructureServer Power and PUE Cost Planner
Plan server power and pue cost planner with transparent inputs, a detailed result and clearly stated assumptions.
Open tool →AI InfrastructureSpeculative Decoding Token Yield
Estimate accepted draft-prefix tokens for a fixed draft length and independent acceptance rate.
Open tool →AI InfrastructureTensor Parallel Shard Fit
Estimate minimum equal shards for weights after reserving device memory for runtime.
Open tool →AI InfrastructureTraining Checkpoint IO Budget
Estimate checkpoint pause overhead and storage occupied by retained checkpoints.
Open tool →AI InfrastructureTraining Sequence Padding Waste
Estimate token compute wasted by padding a batch to a common sequence length.
Open tool →AI InfrastructureTransformer Projection FLOPs
Estimate QKV and output projection arithmetic for a dense attention block.
Open tool →AI InfrastructureVector Search Brute Force Work
Estimate dot-product arithmetic for exact full-scan vector search.
Open tool →