Foundry-LLM-1.2B-800B

A 1.2B parameter language model pretrained on 800B tokens, part of the VLA Foundry model collection.

Model Description

• Architecture: Transformer (24 layers, 2048 hidden dim, 16 heads, SwiGLU FFN, RoPE, QK-norm)
• Parameters: 1.2B (non-embedding)
• Tokenizer: SmolVLM2 (vocab size 49,280)
• Training data: 800B tokens from DCLM-Baseline-1.0
• LR schedule: Warmup + constant (no decay)
• Sequence length: 2048

Earlier checkpoint of the Foundry LLM, used as the language backbone for the downstream VLM and VLA models.

Evaluation Results

Multiple-choice reasoning benchmarks:

HellaSwag	MMLU	ARC-e	ARC-c	PIQA	WinoGrande	OpenBookQA	BoolQ
64.3	26.0	70.3	37.0	75.8	60.9	40.0	63.2

Usage

git clone https://github.com/TRI-ML/vla_foundry.git
cd vla_foundry
pip install -e .

from vla_foundry.models.base_model import BaseModel
model = BaseModel.from_pretrained("TRI-ML/Foundry-LLM-1.2B-800B")

Links

• Project page: tri-ml.github.io/vla_foundry
• Paper: VLA Foundry (arXiv 2604.19728)
• Code: github.com/TRI-ML/vla_foundry
• Collection: VLA Foundry collection