Skip to content

ahxt/LiteLlama-460M-1T

A smaller open LLaMa 2 derived model

Visit official site
Category AI models

What ahxt/LiteLlama-460M-1T is

The model is described as an open source reduction of LLaMa 2 with 460M parameters, trained on 1T tokens.

What you can do with it

  • Generate text from the model
  • Try question answering tasks
  • Evaluate text generation on MMLU

Model notes

  • GPT2Tokenizer handles text tokenisation
  • HuggingFace Transformers is used alongside it
  • MMLU is listed as an evaluation task

Pros & cons

✓ What we like

  • Works with the HuggingFace Transformers library
  • A smaller model size is stated

! What to watch out for

  • It is described as a scaled down LLaMa 2 variant
  • Training is stated at 1T tokens

FAQ

How large is the model?

The material states 460M parameters and 1T training tokens.

What library is named?

HuggingFace Transformers is named for use with the model.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review