Was bedeutet Tensor?

"Tensor ist ein mehrdimensionales Array, die Grunddatenstruktur in ML-Frameworks."

Autor: Karl Kratz, 05.04.2025 (aktualisiert 05.04.2026) · Karl Kratz » Technologie » KI » Tensor

Wo kommt der Begriff Tensor auf karlkratz.de vor?

Ein eigenes LLM? Na klar!
Die Frage nach dem eigenen KI-Modell kommt regelmäßig, weil viele Unternehmen sich fragen, ob sie ihre Daten wirklich externen Anbietern anvertrauen wollen.

Begriffe, die mit Tensor eng verwandt sind

Zeige mir das als

Filter:

A/B Testing concept

Adaptation-Level-Theory concept

Agile concept

Amortisation concept

Ankerheuristik concept

Antwortzeit concept

Appreciative Inquiry concept

Auto-Scaling concept

Autopoiesis concept

Benchmark concept

Best Practice concept

Chatbot concept

Circuit Breaker concept

Collaborative Filtering concept

Computer Vision concept

Continuous Improvement concept

Deep Learning concept

Design Thinking concept

Edge Computing concept

EU AI Act concept

Explainability concept

Feedback Loop concept

Fine-Tuning concept

Generative KI concept

Grafikspeicher concept

Hugging organization

LangChain tool

Machine Learning concept

Mixedbread organization

MVP concept

Neural Network concept

NumPy tool

Schnittstellen-Design concept

Speech-to-Text concept

BERT concept

Chunk als Textsegment für Embeddings concept

Cosine-Similarity metric

Decoder (Architektur) concept

deepset-mxbai-embed-de-large-v1 model

Einbettungsmodell concept

Embedding concept

Embedding-layer concept

Embedding-tabelle concept

Encoder (Architektur) concept

German_Semantic_V3 model

Hard Negative Precision metric

jina-embeddings-v3 model

Kontextverarbeitung concept

MTEB (Massive Text Embedding Benchmark) metric

multilingual-e5-large-instruct model

mxbai-embed-large-v1 model

Normalisierung concept

PCA (Hauptkomponentenanalyse) process

Qwen3-Embedding-0.6B model

Self-Retrieval metric

Textfenster concept

Token-embeddings concept

Treffsicherheit metric

Vektor concept

XLM-RoBERTa concept

Zahlenvektoren concept

Wissensbasis concept

Attention concept

Attention-heads concept

Attention-mechanismen concept

Attention-mechanismus concept

Basis-Modell concept

Bit-Quantisierung method

Decoder-only concept

Flash concept

Fp16 concept

Gpt-2 model

Gpt-2-modell model

Gpt-3 model

Int4 concept

Int8 concept

Llms concept

Mini-LLM concept

Multi-head concept

Precision concept

Prompt concept

Qlora method

Quantisierung method

Rope concept

Rotary concept

Self-attention concept

Tokenisieren process

Tokenisierung process

Tokenizer tool

Tokenizer-training process

Tokenizer-vergleich process

Tokenizers-library tool

Tokens concept

Transformer concept

Transformer-architektur concept

Transformer-block concept

Transformer-layer concept

Transformer-modell concept

Biases concept

Dropout method

Feed-forward concept

Feed-forward-networks concept

Gelu concept

Layer concept

Parametern concept

Parameters concept

Residual concept

RNN concept

Weight concept

NLP concept

Performance-Tracking process

Chunking-strategie method

Chunks concept

Retrieval Augmented Generation method

Rag-lookup process

Rag-system concept

Retrieval concept

Retrieval-qualität metric

Adamw concept

Backpropagation method

Backpropagation-schritt process

Batch-Size concept

Checkpoint concept

Checkpoint-konfiguration concept

Cosine concept

Cosine-scheduler concept

Epoch concept

Fine-tunen process

Gradienten concept

Gradients concept

Hyperparameter concept

Loss metric

Loss-kurven metric

Loss-verlauf metric

Loss-wert metric

Optimizer concept

Optimizer-state concept

Optimizer-wahl concept

Scheduler concept

Tensorboard tool

Training from Scratch method

Trainingsdaten concept

Wandb tool

Weight_decay concept

Hybrid Search method

Vector-Datenbank concept

Vektor-Datenbank tool

Vektordatenbank tool