Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
# Jam-sojm
|
| 2 |
|
| 3 |
-
Jam-sojm model is trained on so13m and then jm52m for one epoch each after resetting the learning rate and decay.
|
|
|
|
| 1 |
# Jam-sojm
|
| 2 |
|
| 3 |
+
Jam-sojm is a GPT2-like model for research in fine-grained Java analysis. It is intended for fine-grained analysis of Java source code at the level of methods, statements, and variables, as a foundation for downstream tasks like code completion, comment generation, and automated bug repair. Jam-sojm model is trained on so13m and then jm52m for one epoch each after resetting the learning rate and decay.
|