This is an intermediate training checkpoint from the TinyLlama project, which aims to pretrain a 1.1B parameter model on 3 trillion tokens. The model follows the architecture and tokenizer of Llama but is significantly smaller. It serves as a resource for researchers and developers interested in small language models.
In the Quantised & converted weights space, TinyLlama 1.1B Intermediate Step 1431k 3T takes a focused approach. It focuses on creating a compact, efficient language model that can run on resource-constrained devices while maintaining reasonable performance. TinyLlama 1.1B Intermediate Step 1431k 3T is an open-source project aimed at developers. TinyLlama 1.1B Intermediate Step 1431k 3T is open source under the Apache-2.0 license. TinyLlama 1.1B Intermediate Step 1431k 3T is available on the web, the command line, and API.
TinyLlama builds and maintains TinyLlama 1.1B Intermediate Step 1431k 3T, and it first shipped in 2023. The GitHub repository has been archived. Among its 3 catalogued features are text generation, small language model, and pretraining checkpoint.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do