MatthewsO3
/

GraphCode-CErl-base

Model card Files Files and versions

MatthewsO3 commited on 15 days ago

Commit

944b470

·

verified ·

1 Parent(s): 1710164

Update README.md

Files changed (1) hide show

README.md +6 -0

README.md CHANGED Viewed

@@ -185,6 +185,12 @@ Final evaluation metrics (epoch 6):
 C++ and Erlang (the two training languages) achieve the strongest results. The model shows solid zero-shot transfer to Java, and reasonable transfer to Python and JavaScript despite not being trained on those languages.
 #### Summary
 The model converges steadily across all 6 epochs. C++ and Erlang show the sharpest perplexity improvements in the first two epochs (from ~5.1 → ~2.1 and ~10.5 → ~2.1 respectively), then plateau. Java, Python, and JavaScript perplexity curves are flatter throughout, consistent with zero-shot generalization rather than direct training signal.

 C++ and Erlang (the two training languages) achieve the strongest results. The model shows solid zero-shot transfer to Java, and reasonable transfer to Python and JavaScript despite not being trained on those languages.
+![plot_top1_accuracy](https://cdn-uploads.huggingface.co/production/uploads/68f92922b614bb0f8762a6c3/mj_5l0_qwx28LdD2t00kc.png)
+![plot_top5_accuracy](https://cdn-uploads.huggingface.co/production/uploads/68f92922b614bb0f8762a6c3/peNpeFzKnpJKl3qCUCOaL.png)
+![plot_perplexity](https://cdn-uploads.huggingface.co/production/uploads/68f92922b614bb0f8762a6c3/bjYgFamL6FhPodhmWYyAb.png)
 #### Summary
 The model converges steadily across all 6 epochs. C++ and Erlang show the sharpest perplexity improvements in the first two epochs (from ~5.1 → ~2.1 and ~10.5 → ~2.1 respectively), then plateau. Java, Python, and JavaScript perplexity curves are flatter throughout, consistent with zero-shot generalization rather than direct training signal.