“A clean, from-scratch implementation of the OLMo architecture with KV caching, RoPE, and an efficient autoregressive inference pipeline. Designed as a minimal yet extensible foundation for post-training research, including RLHF, preference optimization, and reasoning-focused systems.”
OLMo-7B QLoRA fine-tuned on ESOL solubility regression. RDKit-Augmented Prompts + Regression Head on 4096-dim hidden states. Best RMSE: 0.8582. DeepChem GSoC 2026.
Empirical analysis quantifying the 0.009 ROC-AUC gap between PyTorch GNNs (2D graphs) and OLMo-7B (1D SMILES) on Tox21. Exposes the Random Split illusion and provides the baseline justification for Topological State Machine (TSM) integration.
Fine-tuning OLMo-7B with QLoRA & DeepChem for Molecular Toxicity Prediction on the Tox21 dataset.
Add a description, image, and links to the olmo-7b topic page so that developers can more easily learn about it.
To associate your repository with the olmo-7b topic, visit your repo's landing page and select "manage topics."