mirrors/stability-ai--stablelm
1
0
Fork 0
Code Issues Pull Requests Actions Deployments Models Agent Notes Packages Projects Releases Wiki Activity

fix(README): clean up links and clarifications

Browse Source
This commit is contained in:
Jonathan Tow
2023-09-30 17:39:40 -04:00
committed by GitHub
parent 92f9c7a010
commit 54b3684c29
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+3 -3
View File
@@ -29,7 +29,7 @@ This repository contains Stability AI's ongoing development of the StableLM seri
### StableLM-3B-4E1T
> Technical Report: [StableLM-3B-4E1T](https://stability.wandb.io/stability-llm/stable-lm/reports/StableLM-3B-4E1T--VmlldzoyMjU4?accessToken=u3zujipenkx5g7rtcj9qojjgxpconyjktjkli2po09nffrffdhhchq045vp0wyfo)
> Technical Report: [StableLM-3B-4E1T](https://stability.wandb.io/stability-llm/stable-lm/reports/StableLM-3B-4E1T--VmlldzoyMjU4)
StableLM-3B-4E1T is a 3 billion (3B) parameter language model pre-trained under the multi-epoch regime to study the impact of repeated tokens on downstream performance. Given prior success in this area ([Tay et al., 2023](https://arxiv.org/pdf/2205.05131.pdf) and [Taylor et al., 2022](https://galactica.org/static/paper.pdf)), we train on 1 trillion (1T) tokens for 4 epochs following the observations of [Muennighoff et al. (2023)](https://arxiv.org/abs/2305.16264) in "Scaling Data-Constrained Language Models" in which they find "training with up to 4 epochs of repeated data yields negligible changes to loss compared to having unique data." Further inspiration for the token count is taken from "Go smol or go home" ([De Vries, 2023](https://www.harmdevries.com/post/model-size-vs-compute-overhead/)), which suggests a 2.96B model trained for 2.85 trillion tokens achieves a similar loss to a Chinchilla compute-optimal 9.87B language model ($k_n = 0.3$).
@@ -51,13 +51,13 @@ The model is a decoder-only transformer similar to the LLaMA ([Touvron et al., 2
#### Training Data
The dataset is comprised of a filtered mixture of open-source large-scale datasets available on the [HuggingFace Hub](https://huggingface.co/datasets): Falcon RefinedWeb extract ([Penedo et al., 2023](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)), RedPajama-Data ([Together Computer., 2023](https://github.com/togethercomputer/RedPajama-Data)) and The Pile ([Gao et al., 2020](https://arxiv.org/abs/2101.00027)) both without the *Books3* subset, and StarCoder ([Li et al., 2023](https://arxiv.org/abs/2305.06161)).
The dataset is comprised of a filtered mixture of open-source large-scale datasets available on the [HuggingFace Hub](https://huggingface.co/datasets): Falcon RefinedWeb extract ([Penedo et al., 2023](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)), and RedPajama-Data ([Together Computer., 2023](https://github.com/togethercomputer/RedPajama-Data)) and The Pile ([Gao et al., 2020](https://arxiv.org/abs/2101.00027)) both without *Books3* and other subsets, and StarCoder ([Li et al., 2023](https://arxiv.org/abs/2305.06161)).
> Given the large amount of web data, we recommend fine-tuning the base StableLM-3B-4E1T for your downstream tasks.
#### Training Details
Please refer to the provided YAML configuration file [`stablelm-3b-4e1t.yml`](./configs/stablelm-3b-4e1t.yml) for complete hyperparameter settings and the [technical report](https://stability.wandb.io/stability-llm/stable-lm/reports/StableLM-3B-4E1T--VmlldzoyMjU4?accessToken=u3zujipenkx5g7rtcj9qojjgxpconyjktjkli2po09nffrffdhhchq045vp0wyfo) for further details.
Please refer to the provided YAML configuration file [`stablelm-3b-4e1t.yml`](./configs/stablelm-3b-4e1t.yml) for complete hyperparameter settings and the [technical report](https://stability.wandb.io/stability-llm/stable-lm/reports/StableLM-3B-4E1T--VmlldzoyMjU4) for further details.
#### Downstream Results