Downloads · 30 days
0
evenfarther/gpt-oss-20b-chemical-synthesis-adapter
gpt-oss-20b-chemical-synthesis-adapter is a machine learning model from evenfarther. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
--- license: apache-2.0 basemodel: unsloth/gpt-oss-20b-unsloth-bnb-4bit tags: - chemistry - synthesis - classification - gpt-oss - qlora - unsloth - mxfp4 metrics: - TPR - TNR
Downloads · 30 days
0
Access
Public
Updated Oct 6, 2025
Repo size
227 MB
Likes
0
Public
Click a slice to open those files.
.safetensors127 MB · 56%
From the Hugging Face model README
license: apache-2.0 base_model: unsloth/gpt-oss-20b-unsloth-bnb-4bit tags:
language:
This repository contains the LoRA adapter for the GPT-OSS 20B chemical synthesis classifier fine-tuned with focal loss. Training prompts follow the template:
As chief synthesis scientist, judge {compound} critically. Avoid bias. P=synthesizable, U=not:
The base checkpoint remains in Unsloth's MXFP4 4-bit format (unsloth/gpt-oss-20b-unsloth-bnb-4bit). Attaching this adapter restores the best validation performance obtained during training.
train_pu_hem.jsonl) / 79,114 validation (validate_pu_hem.jsonl) samples (11% P / 89% U)| Metric | Value |
|---|---|
| TPR (P Recall) | 0.9043 |
| TNR (U Specificity) | 0.9366 |
The training and validation splits combine multiple public sources and internal curation:
After de-duplication across all sources, approximately 2,560 unique compositions were appended to the base dataset. The combined corpus contains 316,442 training samples and 79,114 validation samples with 11 percent P and 89 percent U class balance.
unsloth>=2025.9.6, transformers>=4.56.0, peft>=0.17.1, bitsandbytes>=0.47.0.