Lunfus
/

nlp-hw2-codegen-finetune-20k

encoder-decoder

Model card Files Files and versions

NLP HW2 code-generation Transformer — 20k fine-tune

Fine-tuned from the final 8x8 RoPE checkpoint for up to 20,000 optimizer steps.

model.pt: final full checkpoint for continued training.
best.pt: full checkpoint with the best validation execution accuracy.
inference-best.pt: lightweight version of the best checkpoint.
latest.pt: recovery checkpoint, uploaded if the run did not finish normally.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support