Péter Harang – Medium

Péter Harang

How efficient are embeddings on Hungarian text?

Let’s check their performance as RAG systems’ components!

Jan 10

How efficient are embeddings on Hungarian text?

Jan 10

Mennyire tudnak magyarul az embedding-ek?

Vizsgáljuk meg, mennyire tudják ellátni a feladatukat RAG rendszerekben!

Jan 9

Mennyire tudnak magyarul az embedding-ek?

Jan 9

Investigating the embedding layer implementation of Tensorflow.JS

Following the data in your model is the best way to learn more about what’s actually happening inside your model.

Jan 5

Investigating the embedding layer implementation of Tensorflow.JS

Jan 5

Building a Transformer in Excel

Pico-scale reference implementation of everyone’s favourite LLM architecture, for demostration purposes

Dec 21, 2024

Building a Transformer in Excel

Dec 21, 2024

Setting up Google AI services for API-based text inference

Let’s do a speed-run of the third big AI provider, and see how it fares with the competition.

Jun 19, 2024

Setting up Google AI services for API-based text inference

Jun 19, 2024

How efficient are LLM’s tokenizers on Hungarian text?

Additional analysis from the Hungarian language-proficiency tests’ data

Jun 12, 2024

How efficient are LLM’s tokenizers on Hungarian text?

Jun 12, 2024

Mennyire hatékony az LLM-ek tokenizálója magyar szövegen?

Járulékos elemzés a magyar nyelvi tesztelés adataiból

Jun 11, 2024

Mennyire hatékony az LLM-ek tokenizálója magyar szövegen?

Jun 11, 2024

To what extent can LLMs speak Hungarian? Pt.2

More models, HuWNLI test, some thoughts on controllability

Jun 2, 2024

To what extent can LLMs speak Hungarian? Pt.2

Jun 2, 2024

Mennyire tudnak magyarul az LLM-ek? Pt.2

Még több modell, HuWNLI teszt, gondolatok a vezérelhetőségről

Jun 2, 2024

Mennyire tudnak magyarul az LLM-ek? Pt.2

Jun 2, 2024

Setting up AWS Bedrock for API-based text inference

Last time I struggled with WatsonX, now let’s check whether AWS is better or worse in this regard. Let’s do a speed-run!

May 29, 2024

Setting up AWS Bedrock for API-based text inference

May 29, 2024

Péter Harang

Péter Harang

I design and build complex, heavily integrated IT ecosystems for the banking sector, focusing on e-channels and AI

Following

Help
Status
About
Careers
Press
Blog
Privacy
Rules
Terms
Text to speech