Update README.md

README.md

@@ -1,3 +1,3 @@
-# Jam 350m
+# Jam
 
-Jam 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.  This is the 350m parameter version.  There is a 1b parameter version trained with the same data.
+Jam 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. This model is trained on jm52m only and trained for one epoch, which is ∼300,000 iterations.