GPT-oss-MOF-Yearwise: pre 2017 training
This repository shows a series of AutoModelForCausalLM which are respectively trained on literature data with different publication year range.
We reports their forecasting performance on next-period published experimental data.
This 1999-2017 branch specifically trained GPT on MOF literatures prior than 2017.
Result
| Train set spans | Inference | Test size | Accuracy |
|---|---|---|---|
| ~2017 | Holdout | 1320P+ 1275N | 0.73295 |
| ~2017 | 22 questions | 11P + 11N | 17/22 |
| ~2017 | 2018-2020 | 2232P + 1922N | 0.66635 |
Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "GPT-oss-MOF-Yearwise"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
To go throught the quick start details for reprduction, see the official released MOFinder main checkpoint which is well-trained on full PN data.
Citation and license
When using these weights, cite MOFinder, identify StarLiu714/GPT-oss-MOF and the Hub revision used, and reference the GPT-oss-20B base model.
This release is licensed under the GNU General Public License v3.0; see LICENSE.
The underlying GPT-oss-20B model retains its upstream Apache-2.0 license, reproduced in LICENSE.base-model. This release includes MOF fine-tuning changes merged into the base model and a BF16 weight export.
MOFinder codebase is separately licensed under MIT.