YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Own A checkpoint

Private Hugging Face Hub copy for CSCI 680 Assignment 1, uploaded 2026-10-19 from the hash-verified local staging folder.

52,249,088-parameter randomly initialized T5 architecture; own corpus-only 16k byte BPE. Condition: nopt. Maximum/stop/validation-selected steps: 10000/10000/10000. Actual GPU reservation hours: 3.692889.

Final exact match: 0/5,000 on both tasks. Revised-code CodeBLEU: 0.0910875535 (0-1). Comment BLEU-4: 0.1863250007 percent (0-100 scale). CodeBLEU dataflow hash-order sensitivity and the separate fixed-seed recomputation are disclosed in the companion analysis.

Research/coursework artifact; poor measured reviewer usefulness. Preserve the supplied-data CC BY 4.0 attribution (Tufano et al., ICSE 2022; https://doi.org/10.5281/zenodo.5387856) and source MIT notices. No third-party trained tokenizer, weights or assembled corpus were used. Licensing, hosting destination/access and student personal review remain pending. Original local paths in run-provenance.json describe the actual run; relocate explicitly when restoring. The weights/config/tokenizer support restoration; optimizer/RNG resume files remain in the private local run directory.

Downloads last month
20
Safetensors
Model size
52.2M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support