|
|
--- |
|
|
tags: |
|
|
- text-generation-inference |
|
|
- transformers |
|
|
- unsloth |
|
|
- olmo2 |
|
|
license: apache-2.0 |
|
|
language: |
|
|
- en |
|
|
datasets: |
|
|
- Pinkstack/roblox-luau-corpus-text |
|
|
- Roblox/luau_corpus |
|
|
- boatbomber/roblox-info-dump |
|
|
- wikimedia/wikipedia |
|
|
pipeline_tag: text-generation |
|
|
--- |
|
|
|
|
|
 |
|
|
|
|
|
# print("Before we start") |
|
|
We are not related to Roblox in any way, any mention of Roblox is purely to help people understand what the model is about. |
|
|
As per the [Roblox website](https://create.roblox.com/docs/assistant/guide), they use Meta's Llama 3 (we assume 70B) for their AI assistant. This model, while powerful, cannot come close to the performance of a 70B model. |
|
|
|
|
|
# print("Stages of pre-training") |
|
|
|
|
|
This model was pre-trained in 3 stages. |
|
|
|
|
|
- Stage 1: Pre-training on the Pinkstack/roblox-luau-corpus-text & Roblox/luau_corpus on 4096 tokens (the maximum olmo 2 can usually reach) |
|
|
|
|
|
- Stage 2: Pre-training on the boatbomber/roblox-info-dump with rope scaling set to 4, so stage 2 was for expanding the context of the model to **16384**. |
|
|
|
|
|
|
|
|
!stage 3 and onwards were with added layers. the model started with 16 layers, then we merged another 20 to make the model bigger and deeper! |
|
|
- Stage 3: Training on a mix of Pinkstack/roblox-luau-corpus-text & Roblox/luau_corpus + wikimedia/wikipedia with rope scaling set to 8, aka **32768** tokens of context. We mixed the wikimedia/wikipedia to hopefully improve the general text and knowledge of the model. |
|
|
|
|
|
This repo contains the stage 3 pre-trained/base model. |
