Pythia-1.4b is a 1.4 billion parameter autoregressive language model hosted on Hugging Face. It belongs to the Pythia suite released by EleutherAI and carries the pipeline tag for text generation.
The model uses the GPT-NeoX architecture implemented in PyTorch with safetensors weights. Its tags indicate support for causal language modeling and English text. It is provided as an open repository containing the model files along with tokenizer configuration that defines beginning-of-string, end-of-string, and unknown tokens as variants of the special token <|endoftext|.
Delivery occurs through the Hugging Face ecosystem, where the model can be loaded via the transformers library. The repository remains public and ungated. No pricing information appears because the model is distributed at no cost under open terms typical of such releases.
Researchers and developers use it for text-generation tasks and to examine training dynamics of decoder-only transformer models. The page records over 2.7 million lifetime downloads and 89,000 recent downloads, reflecting community interest in the artifact.
In the Foundation models & chat space, Pythia 1.4b takes a focused approach. It focuses on providing transparent, reproducible large language model training data and checkpoints for research. Pythia 1.4b is an open-source project aimed at AI researchers. The project is open source (Apache-2.0). The product ships for the web, the command line, and API.
Behind Pythia 1.4b is EleutherAI, based in the United States, and the product first shipped in 2021. The project is developed in the open on GitHub with 2.9k stars. Among its 3 catalogued features are Causal Language Modeling, Training Suite, and Public Checkpoints.
Latest indexed changes and source events
EleutherAI/pythia-1.4b verified by the PulseGate indexer
Other apps tracked under the same category.
