|
--- |
|
language: |
|
- en |
|
tags: |
|
- causal-lm |
|
library_name: transformers |
|
license: apache-2.0 |
|
datasets: |
|
- allenai/dolma |
|
--- |
|
|
|
# Gemstone-512x12_lr_ablation |
|
Gemstone-512x12_lr_ablation is part of the [Gemstone Suite of Models](https://huggingface.co/collections/tomg-group-umd/gemstone-models-679408ee3f19f1d4d00e8b10). A set of models trained with varying widths and depths. This particular version, denoted by the `_lr_ablation` postfix, corresponds to an ablation detailed in the paper where we train the same suite of models but with a learning rate that is half of the original. |
|
|
|
## Training |
|
We train using [litgpt](https://github.com/Lightning-AI/litgpt) and [AxoNN](https://github.com/axonn-ai/litgpt) using AMD MI250X GPUs on [Frontier](https://www.olcf.ornl.gov/olcf-resources/compute-systems/frontier/) at Oak Ridge National Laboratory with a global batch size of 2048. |
|
|
|
## Data |
|
Train and validation data is taken from non-overlapping subsets of [dolma](https://huggingface.co/datasets/allenai/dolma). As such it is _not_ an instruction model. |
|
This model is trained for 100 billion tokens (in contrast to the main suite, which is trained to 350 billion tokens), we upload checkpoints every 2 billion tokens (477 steps). |
|
|
|
## Using Gemstone-512x12_lr_ablation |
|
The Gemstones are based on the [gemma-2b](https://huggingface.co/google/gemma-2b) architecture and use [modeling_gemma.py](https://github.com/huggingface/transformers/blob/main/src/transformers/models/gemma/modeling_gemma.py) to run using the transformers library. |
|
|
|
## Licence |
|
This model is released under the [apache-2.0](https://choosealicense.com/licenses/apache-2.0/) licence. |
|
|
|
## Contact |
|
Please, feel free to contact us with any questions, or open a discussion thread. |
|
|
|
# Citation |
|
``` |
|
@article{mcleish2024gemstones |
|
title={Gemstones: A Model Suite for Multi-Faceted Scaling Laws}, |
|
author={Sean McLeish and John Kirchenbauer and David Yu Miller and Siddharth Singh and Abhinav Bhatele and Micah Goldblum and Ashwinee Panda and Tom Goldstein}, |
|
journal={arXiv preprint arXiv:2502.}, |
|
year={2025} |
|
} |
|
``` |
