Upload byte_trigram_proto_64_patch_2_v1/final_report.json with huggingface_hub
Browse files
byte_trigram_proto_64_patch_2_v1/final_report.json
ADDED
|
@@ -0,0 +1,3033 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"run_name": "byte_trigram_proto_64_patch_2_v1_64x64_h64_d1_lr0.001",
|
| 3 |
+
"config": {
|
| 4 |
+
"V": 32,
|
| 5 |
+
"D": 4,
|
| 6 |
+
"patch_size": 2,
|
| 7 |
+
"hidden": 64,
|
| 8 |
+
"depth": 1,
|
| 9 |
+
"n_cross_layers": 1,
|
| 10 |
+
"n_heads": 4,
|
| 11 |
+
"smooth_mid": 16,
|
| 12 |
+
"linear_readout": true,
|
| 13 |
+
"svd_mode": "none",
|
| 14 |
+
"match_params": true,
|
| 15 |
+
"dataset": "byte_trigram",
|
| 16 |
+
"img_size": 64,
|
| 17 |
+
"batch_size": 1024,
|
| 18 |
+
"lr": 0.001,
|
| 19 |
+
"epochs": 50,
|
| 20 |
+
"target_cv": 1.0,
|
| 21 |
+
"allowed_types": null,
|
| 22 |
+
"curriculum": null,
|
| 23 |
+
"tier_schedule": null,
|
| 24 |
+
"tree_depth": 4
|
| 25 |
+
},
|
| 26 |
+
"n_params": 52571,
|
| 27 |
+
"n_patches": 1024,
|
| 28 |
+
"best_test_mse": 3.5706435755855637e-07,
|
| 29 |
+
"history": [
|
| 30 |
+
{
|
| 31 |
+
"epoch": 1,
|
| 32 |
+
"global_batch": 500,
|
| 33 |
+
"train_recon": 0.0027674232192075577,
|
| 34 |
+
"test_mse": 0.0001354175910819322,
|
| 35 |
+
"S0": 2.44702410697937,
|
| 36 |
+
"SD": 1.7133033275604248,
|
| 37 |
+
"ratio": 1.4282491916835296,
|
| 38 |
+
"erank": 3.9609861373901367,
|
| 39 |
+
"row_cv": 0.8638597070617909,
|
| 40 |
+
"cv_in_band": true,
|
| 41 |
+
"s_delta": 0.024999180808663368,
|
| 42 |
+
"alpha_mean": 0.02386362850666046,
|
| 43 |
+
"alpha_std": 0.0003088540106546134,
|
| 44 |
+
"epoch_max_grad": 1.1047625925621887
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 1,
|
| 48 |
+
"global_batch": 976,
|
| 49 |
+
"epoch_test_mse": 5.946664094226435e-05,
|
| 50 |
+
"train_recon": 0.0014925856672845363,
|
| 51 |
+
"S0": 2.324779748916626,
|
| 52 |
+
"SD": 1.6715044975280762,
|
| 53 |
+
"ratio": 1.390830636024475,
|
| 54 |
+
"erank": 3.964594841003418,
|
| 55 |
+
"row_cv": 0.8575342387723678,
|
| 56 |
+
"cv_in_band": true,
|
| 57 |
+
"s_delta": 0.024340152740478516,
|
| 58 |
+
"alpha_mean": 0.023855727165937424,
|
| 59 |
+
"alpha_std": 0.0003300230309832841,
|
| 60 |
+
"epoch_max_grad": 1.1047625925621887,
|
| 61 |
+
"epoch_time_s": 264.79387044906616,
|
| 62 |
+
"tree_metrics": null,
|
| 63 |
+
"sp_metrics": null,
|
| 64 |
+
"bt_metrics": {
|
| 65 |
+
"per_byte_acc": 0.5010884404182434,
|
| 66 |
+
"per_byte_l1": 0.5989481806755066,
|
| 67 |
+
"trigram_exact_rate": 0.137298583984375,
|
| 68 |
+
"per_channel_acc": [
|
| 69 |
+
0.51763916015625,
|
| 70 |
+
0.480010986328125,
|
| 71 |
+
0.505615234375
|
| 72 |
+
]
|
| 73 |
+
}
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 2,
|
| 77 |
+
"global_batch": 1000,
|
| 78 |
+
"train_recon": 5.2550662682430506e-05,
|
| 79 |
+
"test_mse": 4.8699053877498955e-05,
|
| 80 |
+
"S0": 2.341148853302002,
|
| 81 |
+
"SD": 1.6827635765075684,
|
| 82 |
+
"ratio": 1.3912523851082705,
|
| 83 |
+
"erank": 3.964885950088501,
|
| 84 |
+
"row_cv": 0.8315765933288904,
|
| 85 |
+
"cv_in_band": true,
|
| 86 |
+
"s_delta": 0.024470359086990356,
|
| 87 |
+
"alpha_mean": 0.02385430969297886,
|
| 88 |
+
"alpha_std": 0.0003351231571286917,
|
| 89 |
+
"epoch_max_grad": 0.01792588568615288
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 2,
|
| 93 |
+
"global_batch": 1500,
|
| 94 |
+
"train_recon": 0.00016022638076976765,
|
| 95 |
+
"test_mse": 0.0005130392964929342,
|
| 96 |
+
"S0": 2.075920581817627,
|
| 97 |
+
"SD": 1.511354684829712,
|
| 98 |
+
"ratio": 1.373549563791527,
|
| 99 |
+
"erank": 3.9582903385162354,
|
| 100 |
+
"row_cv": 0.9225158136496108,
|
| 101 |
+
"cv_in_band": true,
|
| 102 |
+
"s_delta": 0.02280120924115181,
|
| 103 |
+
"alpha_mean": 0.023900160565972328,
|
| 104 |
+
"alpha_std": 0.00025629877927713096,
|
| 105 |
+
"epoch_max_grad": 0.5382357802479467
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 2,
|
| 109 |
+
"global_batch": 1952,
|
| 110 |
+
"epoch_test_mse": 2.9024911270244046e-05,
|
| 111 |
+
"train_recon": 0.00011511787850164093,
|
| 112 |
+
"S0": 2.190427780151367,
|
| 113 |
+
"SD": 1.5969853401184082,
|
| 114 |
+
"ratio": 1.3716017007827759,
|
| 115 |
+
"erank": 3.965315341949463,
|
| 116 |
+
"row_cv": 0.8495220970556727,
|
| 117 |
+
"cv_in_band": true,
|
| 118 |
+
"s_delta": 0.023823678493499756,
|
| 119 |
+
"alpha_mean": 0.023888517171144485,
|
| 120 |
+
"alpha_std": 0.00028586553526110947,
|
| 121 |
+
"epoch_max_grad": 0.5382357802479467,
|
| 122 |
+
"epoch_time_s": 265.0908205509186,
|
| 123 |
+
"tree_metrics": null,
|
| 124 |
+
"sp_metrics": null,
|
| 125 |
+
"bt_metrics": {
|
| 126 |
+
"per_byte_acc": 0.6690775752067566,
|
| 127 |
+
"per_byte_l1": 0.3673909604549408,
|
| 128 |
+
"trigram_exact_rate": 0.324981689453125,
|
| 129 |
+
"per_channel_acc": [
|
| 130 |
+
0.697174072265625,
|
| 131 |
+
0.682342529296875,
|
| 132 |
+
0.627716064453125
|
| 133 |
+
]
|
| 134 |
+
}
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3,
|
| 138 |
+
"global_batch": 2000,
|
| 139 |
+
"train_recon": 0.0007904732312150978,
|
| 140 |
+
"test_mse": 0.00010439680772833526,
|
| 141 |
+
"S0": 2.082271099090576,
|
| 142 |
+
"SD": 1.49592924118042,
|
| 143 |
+
"ratio": 1.3919582743953172,
|
| 144 |
+
"erank": 3.9550657272338867,
|
| 145 |
+
"row_cv": 0.9624760685825962,
|
| 146 |
+
"cv_in_band": true,
|
| 147 |
+
"s_delta": 0.023278411477804184,
|
| 148 |
+
"alpha_mean": 0.0239274762570858,
|
| 149 |
+
"alpha_std": 0.00018925382755696774,
|
| 150 |
+
"epoch_max_grad": 0.4983902426281303
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 3,
|
| 154 |
+
"global_batch": 2500,
|
| 155 |
+
"train_recon": 0.0001290029843960964,
|
| 156 |
+
"test_mse": 4.4659707782557234e-05,
|
| 157 |
+
"S0": 2.0748836994171143,
|
| 158 |
+
"SD": 1.4964163303375244,
|
| 159 |
+
"ratio": 1.3865684592491911,
|
| 160 |
+
"erank": 3.9589219093322754,
|
| 161 |
+
"row_cv": 0.8985843027417505,
|
| 162 |
+
"cv_in_band": true,
|
| 163 |
+
"s_delta": 0.02336796186864376,
|
| 164 |
+
"alpha_mean": 0.023930899798870087,
|
| 165 |
+
"alpha_std": 0.00018714838370215148,
|
| 166 |
+
"epoch_max_grad": 0.4983902426281303
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3,
|
| 170 |
+
"global_batch": 2928,
|
| 171 |
+
"epoch_test_mse": 2.196892148058396e-05,
|
| 172 |
+
"train_recon": 0.00010415576559994331,
|
| 173 |
+
"S0": 2.0407698154449463,
|
| 174 |
+
"SD": 1.5172103643417358,
|
| 175 |
+
"ratio": 1.3450803756713867,
|
| 176 |
+
"erank": 3.964437484741211,
|
| 177 |
+
"row_cv": 0.8683343555119709,
|
| 178 |
+
"cv_in_band": true,
|
| 179 |
+
"s_delta": 0.023511111736297607,
|
| 180 |
+
"alpha_mean": 0.0239417664706707,
|
| 181 |
+
"alpha_std": 0.00015769827587064356,
|
| 182 |
+
"epoch_max_grad": 0.4983902426281303,
|
| 183 |
+
"epoch_time_s": 264.87710905075073,
|
| 184 |
+
"tree_metrics": null,
|
| 185 |
+
"sp_metrics": null,
|
| 186 |
+
"bt_metrics": {
|
| 187 |
+
"per_byte_acc": 0.75244140625,
|
| 188 |
+
"per_byte_l1": 0.2728169858455658,
|
| 189 |
+
"trigram_exact_rate": 0.470458984375,
|
| 190 |
+
"per_channel_acc": [
|
| 191 |
+
0.749847412109375,
|
| 192 |
+
0.753326416015625,
|
| 193 |
+
0.754150390625
|
| 194 |
+
]
|
| 195 |
+
}
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 4,
|
| 199 |
+
"global_batch": 3000,
|
| 200 |
+
"train_recon": 3.0978066661191406e-05,
|
| 201 |
+
"test_mse": 1.958818393177353e-05,
|
| 202 |
+
"S0": 2.0512263774871826,
|
| 203 |
+
"SD": 1.525133728981018,
|
| 204 |
+
"ratio": 1.344948528158364,
|
| 205 |
+
"erank": 3.96504545211792,
|
| 206 |
+
"row_cv": 0.8311612095119938,
|
| 207 |
+
"cv_in_band": true,
|
| 208 |
+
"s_delta": 0.023666854947805405,
|
| 209 |
+
"alpha_mean": 0.023940090090036392,
|
| 210 |
+
"alpha_std": 0.0001607833692105487,
|
| 211 |
+
"epoch_max_grad": 0.05079713327976941
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 4,
|
| 215 |
+
"global_batch": 3500,
|
| 216 |
+
"train_recon": 5.774103999635747e-05,
|
| 217 |
+
"test_mse": 1.9385915948078036e-05,
|
| 218 |
+
"S0": 2.000978708267212,
|
| 219 |
+
"SD": 1.4696054458618164,
|
| 220 |
+
"ratio": 1.361575448897462,
|
| 221 |
+
"erank": 3.962959051132202,
|
| 222 |
+
"row_cv": 0.8659678711964218,
|
| 223 |
+
"cv_in_band": true,
|
| 224 |
+
"s_delta": 0.022808801382780075,
|
| 225 |
+
"alpha_mean": 0.023967966437339783,
|
| 226 |
+
"alpha_std": 0.00011928829189855605,
|
| 227 |
+
"epoch_max_grad": 0.238771018929549
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 4,
|
| 231 |
+
"global_batch": 3904,
|
| 232 |
+
"epoch_test_mse": 1.7933658557012677e-05,
|
| 233 |
+
"train_recon": 5.421655870312876e-05,
|
| 234 |
+
"S0": 1.9920097589492798,
|
| 235 |
+
"SD": 1.471780776977539,
|
| 236 |
+
"ratio": 1.3534690141677856,
|
| 237 |
+
"erank": 3.9629409313201904,
|
| 238 |
+
"row_cv": 0.80592387665684,
|
| 239 |
+
"cv_in_band": true,
|
| 240 |
+
"s_delta": 0.023068904876708984,
|
| 241 |
+
"alpha_mean": 0.02398916706442833,
|
| 242 |
+
"alpha_std": 9.748384763952345e-05,
|
| 243 |
+
"epoch_max_grad": 0.238771018929549,
|
| 244 |
+
"epoch_time_s": 264.9651896953583,
|
| 245 |
+
"tree_metrics": null,
|
| 246 |
+
"sp_metrics": null,
|
| 247 |
+
"bt_metrics": {
|
| 248 |
+
"per_byte_acc": 0.7701416015625,
|
| 249 |
+
"per_byte_l1": 0.2477213591337204,
|
| 250 |
+
"trigram_exact_rate": 0.49493408203125,
|
| 251 |
+
"per_channel_acc": [
|
| 252 |
+
0.7955322265625,
|
| 253 |
+
0.805419921875,
|
| 254 |
+
0.70947265625
|
| 255 |
+
]
|
| 256 |
+
}
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 5,
|
| 260 |
+
"global_batch": 4000,
|
| 261 |
+
"train_recon": 6.190030474802673e-05,
|
| 262 |
+
"test_mse": 1.834646172937937e-05,
|
| 263 |
+
"S0": 1.9759689569473267,
|
| 264 |
+
"SD": 1.458849310874939,
|
| 265 |
+
"ratio": 1.3544709029732058,
|
| 266 |
+
"erank": 3.962581157684326,
|
| 267 |
+
"row_cv": 0.9336950242027201,
|
| 268 |
+
"cv_in_band": true,
|
| 269 |
+
"s_delta": 0.02293865755200386,
|
| 270 |
+
"alpha_mean": 0.024000151082873344,
|
| 271 |
+
"alpha_std": 8.66135087562725e-05,
|
| 272 |
+
"epoch_max_grad": 0.14154122294806537
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 5,
|
| 276 |
+
"global_batch": 4500,
|
| 277 |
+
"train_recon": 4.3268709625725924e-05,
|
| 278 |
+
"test_mse": 1.1398780770832673e-05,
|
| 279 |
+
"S0": 1.9942256212234497,
|
| 280 |
+
"SD": 1.4810492992401123,
|
| 281 |
+
"ratio": 1.346495088841191,
|
| 282 |
+
"erank": 3.9642558097839355,
|
| 283 |
+
"row_cv": 0.8136725999615086,
|
| 284 |
+
"cv_in_band": true,
|
| 285 |
+
"s_delta": 0.023561954498291016,
|
| 286 |
+
"alpha_mean": 0.024013753980398178,
|
| 287 |
+
"alpha_std": 7.849843677831814e-05,
|
| 288 |
+
"epoch_max_grad": 0.18808397424852344
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5,
|
| 292 |
+
"global_batch": 4880,
|
| 293 |
+
"epoch_test_mse": 1.3505083444761113e-05,
|
| 294 |
+
"train_recon": 5.448421102133192e-05,
|
| 295 |
+
"S0": 1.937578558921814,
|
| 296 |
+
"SD": 1.4392330646514893,
|
| 297 |
+
"ratio": 1.3462576866149902,
|
| 298 |
+
"erank": 3.9626893997192383,
|
| 299 |
+
"row_cv": 0.8721368458631036,
|
| 300 |
+
"cv_in_band": true,
|
| 301 |
+
"s_delta": 0.023052960634231567,
|
| 302 |
+
"alpha_mean": 0.024062106385827065,
|
| 303 |
+
"alpha_std": 8.303476352011785e-05,
|
| 304 |
+
"epoch_max_grad": 0.21583379021392787,
|
| 305 |
+
"epoch_time_s": 264.83897280693054,
|
| 306 |
+
"tree_metrics": null,
|
| 307 |
+
"sp_metrics": null,
|
| 308 |
+
"bt_metrics": {
|
| 309 |
+
"per_byte_acc": 0.8482462763786316,
|
| 310 |
+
"per_byte_l1": 0.165252685546875,
|
| 311 |
+
"trigram_exact_rate": 0.6495361328125,
|
| 312 |
+
"per_channel_acc": [
|
| 313 |
+
0.849945068359375,
|
| 314 |
+
0.8519287109375,
|
| 315 |
+
0.842864990234375
|
| 316 |
+
]
|
| 317 |
+
}
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 6,
|
| 321 |
+
"global_batch": 5000,
|
| 322 |
+
"train_recon": 3.874563594005546e-05,
|
| 323 |
+
"test_mse": 1.181236257252749e-05,
|
| 324 |
+
"S0": 1.9423298835754395,
|
| 325 |
+
"SD": 1.4421780109405518,
|
| 326 |
+
"ratio": 1.3468031375964955,
|
| 327 |
+
"erank": 3.96311354637146,
|
| 328 |
+
"row_cv": 0.9263615767502885,
|
| 329 |
+
"cv_in_band": true,
|
| 330 |
+
"s_delta": 0.023250164464116096,
|
| 331 |
+
"alpha_mean": 0.02406703121960163,
|
| 332 |
+
"alpha_std": 8.694447024026886e-05,
|
| 333 |
+
"epoch_max_grad": 0.09491111069015706
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 6,
|
| 337 |
+
"global_batch": 5500,
|
| 338 |
+
"train_recon": 5.0061241208529666e-05,
|
| 339 |
+
"test_mse": 1.0687507710827049e-05,
|
| 340 |
+
"S0": 1.9165692329406738,
|
| 341 |
+
"SD": 1.4197666645050049,
|
| 342 |
+
"ratio": 1.3499184530506585,
|
| 343 |
+
"erank": 3.9630701541900635,
|
| 344 |
+
"row_cv": 0.8563269743368669,
|
| 345 |
+
"cv_in_band": true,
|
| 346 |
+
"s_delta": 0.022973759099841118,
|
| 347 |
+
"alpha_mean": 0.024105150252580643,
|
| 348 |
+
"alpha_std": 0.00014700098836328834,
|
| 349 |
+
"epoch_max_grad": 0.18749636857733912
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 6,
|
| 353 |
+
"global_batch": 5856,
|
| 354 |
+
"epoch_test_mse": 8.934552175924182e-05,
|
| 355 |
+
"train_recon": 4.6415729129279735e-05,
|
| 356 |
+
"S0": 1.8994265794754028,
|
| 357 |
+
"SD": 1.405021071434021,
|
| 358 |
+
"ratio": 1.3518847227096558,
|
| 359 |
+
"erank": 3.961536169052124,
|
| 360 |
+
"row_cv": 0.911617979147907,
|
| 361 |
+
"cv_in_band": true,
|
| 362 |
+
"s_delta": 0.022837847471237183,
|
| 363 |
+
"alpha_mean": 0.024124983698129654,
|
| 364 |
+
"alpha_std": 0.00017319612379651517,
|
| 365 |
+
"epoch_max_grad": 0.18749636857733912,
|
| 366 |
+
"epoch_time_s": 264.9289765357971,
|
| 367 |
+
"tree_metrics": null,
|
| 368 |
+
"sp_metrics": null,
|
| 369 |
+
"bt_metrics": {
|
| 370 |
+
"per_byte_acc": 0.3417154848575592,
|
| 371 |
+
"per_byte_l1": 0.9048054814338684,
|
| 372 |
+
"trigram_exact_rate": 0.034393310546875,
|
| 373 |
+
"per_channel_acc": [
|
| 374 |
+
0.339813232421875,
|
| 375 |
+
0.23065185546875,
|
| 376 |
+
0.454681396484375
|
| 377 |
+
]
|
| 378 |
+
}
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 7,
|
| 382 |
+
"global_batch": 6000,
|
| 383 |
+
"train_recon": 2.4281153148169527e-05,
|
| 384 |
+
"test_mse": 9.734287232276984e-06,
|
| 385 |
+
"S0": 1.9105498790740967,
|
| 386 |
+
"SD": 1.4157369136810303,
|
| 387 |
+
"ratio": 1.3495091122625473,
|
| 388 |
+
"erank": 3.9625093936920166,
|
| 389 |
+
"row_cv": 0.8837893982426422,
|
| 390 |
+
"cv_in_band": true,
|
| 391 |
+
"s_delta": 0.023214953020215034,
|
| 392 |
+
"alpha_mean": 0.024129364639520645,
|
| 393 |
+
"alpha_std": 0.000180811679456383,
|
| 394 |
+
"epoch_max_grad": 0.06305860274259956
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 7,
|
| 398 |
+
"global_batch": 6500,
|
| 399 |
+
"train_recon": 3.818676045936983e-05,
|
| 400 |
+
"test_mse": 2.218884401372634e-05,
|
| 401 |
+
"S0": 1.867276906967163,
|
| 402 |
+
"SD": 1.3968987464904785,
|
| 403 |
+
"ratio": 1.3367303094022698,
|
| 404 |
+
"erank": 3.9617011547088623,
|
| 405 |
+
"row_cv": 1.010573690071569,
|
| 406 |
+
"cv_in_band": true,
|
| 407 |
+
"s_delta": 0.022711945697665215,
|
| 408 |
+
"alpha_mean": 0.024173475801944733,
|
| 409 |
+
"alpha_std": 0.00025568436831235886,
|
| 410 |
+
"epoch_max_grad": 0.150227729276863
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 7,
|
| 414 |
+
"global_batch": 6832,
|
| 415 |
+
"epoch_test_mse": 3.1283317104680464e-05,
|
| 416 |
+
"train_recon": 3.359204554663236e-05,
|
| 417 |
+
"S0": 1.8790645599365234,
|
| 418 |
+
"SD": 1.39583158493042,
|
| 419 |
+
"ratio": 1.346197247505188,
|
| 420 |
+
"erank": 3.9621734619140625,
|
| 421 |
+
"row_cv": 0.9616172895231485,
|
| 422 |
+
"cv_in_band": true,
|
| 423 |
+
"s_delta": 0.022997379302978516,
|
| 424 |
+
"alpha_mean": 0.024180227890610695,
|
| 425 |
+
"alpha_std": 0.00026366757811047137,
|
| 426 |
+
"epoch_max_grad": 0.150227729276863,
|
| 427 |
+
"epoch_time_s": 264.85445737838745,
|
| 428 |
+
"tree_metrics": null,
|
| 429 |
+
"sp_metrics": null,
|
| 430 |
+
"bt_metrics": {
|
| 431 |
+
"per_byte_acc": 0.50091552734375,
|
| 432 |
+
"per_byte_l1": 0.5252583622932434,
|
| 433 |
+
"trigram_exact_rate": 0.1707763671875,
|
| 434 |
+
"per_channel_acc": [
|
| 435 |
+
0.464752197265625,
|
| 436 |
+
0.447021484375,
|
| 437 |
+
0.590972900390625
|
| 438 |
+
]
|
| 439 |
+
}
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 8,
|
| 443 |
+
"global_batch": 7000,
|
| 444 |
+
"train_recon": 1.7817291536846912e-05,
|
| 445 |
+
"test_mse": 8.326454917551018e-06,
|
| 446 |
+
"S0": 1.8864144086837769,
|
| 447 |
+
"SD": 1.403271198272705,
|
| 448 |
+
"ratio": 1.3442978075533778,
|
| 449 |
+
"erank": 3.9631450176239014,
|
| 450 |
+
"row_cv": 0.9006092031815097,
|
| 451 |
+
"cv_in_band": true,
|
| 452 |
+
"s_delta": 0.023146286606788635,
|
| 453 |
+
"alpha_mean": 0.024180501699447632,
|
| 454 |
+
"alpha_std": 0.0002652645925991237,
|
| 455 |
+
"epoch_max_grad": 0.04496039492997877
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 8,
|
| 459 |
+
"global_batch": 7500,
|
| 460 |
+
"train_recon": 3.706768047627819e-05,
|
| 461 |
+
"test_mse": 0.00030620256438851357,
|
| 462 |
+
"S0": 1.849372386932373,
|
| 463 |
+
"SD": 1.3763422966003418,
|
| 464 |
+
"ratio": 1.3436863620798285,
|
| 465 |
+
"erank": 3.9604697227478027,
|
| 466 |
+
"row_cv": 0.9079515205819144,
|
| 467 |
+
"cv_in_band": true,
|
| 468 |
+
"s_delta": 0.022692188620567322,
|
| 469 |
+
"alpha_mean": 0.024224502965807915,
|
| 470 |
+
"alpha_std": 0.00033049940248019993,
|
| 471 |
+
"epoch_max_grad": 0.16122116916917212
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 8,
|
| 475 |
+
"global_batch": 7808,
|
| 476 |
+
"epoch_test_mse": 7.779824745375663e-06,
|
| 477 |
+
"train_recon": 3.377334115513252e-05,
|
| 478 |
+
"S0": 1.871016263961792,
|
| 479 |
+
"SD": 1.3930304050445557,
|
| 480 |
+
"ratio": 1.343126654624939,
|
| 481 |
+
"erank": 3.9616332054138184,
|
| 482 |
+
"row_cv": 0.8795594088953732,
|
| 483 |
+
"cv_in_band": true,
|
| 484 |
+
"s_delta": 0.023200273513793945,
|
| 485 |
+
"alpha_mean": 0.024239055812358856,
|
| 486 |
+
"alpha_std": 0.0003578162577468902,
|
| 487 |
+
"epoch_max_grad": 0.16122116916917212,
|
| 488 |
+
"epoch_time_s": 264.94762659072876,
|
| 489 |
+
"tree_metrics": null,
|
| 490 |
+
"sp_metrics": null,
|
| 491 |
+
"bt_metrics": {
|
| 492 |
+
"per_byte_acc": 0.9170125126838684,
|
| 493 |
+
"per_byte_l1": 0.0891316756606102,
|
| 494 |
+
"trigram_exact_rate": 0.799407958984375,
|
| 495 |
+
"per_channel_acc": [
|
| 496 |
+
0.91796875,
|
| 497 |
+
0.923828125,
|
| 498 |
+
0.90924072265625
|
| 499 |
+
]
|
| 500 |
+
}
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 9,
|
| 504 |
+
"global_batch": 8000,
|
| 505 |
+
"train_recon": 2.5347588653801267e-05,
|
| 506 |
+
"test_mse": 1.930293728946708e-05,
|
| 507 |
+
"S0": 1.863385558128357,
|
| 508 |
+
"SD": 1.3860912322998047,
|
| 509 |
+
"ratio": 1.3443455244956493,
|
| 510 |
+
"erank": 3.9616453647613525,
|
| 511 |
+
"row_cv": 1.0138702694724622,
|
| 512 |
+
"cv_in_band": true,
|
| 513 |
+
"s_delta": 0.023062409833073616,
|
| 514 |
+
"alpha_mean": 0.024249248206615448,
|
| 515 |
+
"alpha_std": 0.00037720639375038445,
|
| 516 |
+
"epoch_max_grad": 0.05679120910607045
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 9,
|
| 520 |
+
"global_batch": 8500,
|
| 521 |
+
"train_recon": 2.3676046587862754e-05,
|
| 522 |
+
"test_mse": 5.6699309425312094e-06,
|
| 523 |
+
"S0": 1.8638752698898315,
|
| 524 |
+
"SD": 1.3832794427871704,
|
| 525 |
+
"ratio": 1.3474321953776647,
|
| 526 |
+
"erank": 3.961972951889038,
|
| 527 |
+
"row_cv": 0.8526614704512907,
|
| 528 |
+
"cv_in_band": true,
|
| 529 |
+
"s_delta": 0.023190360516309738,
|
| 530 |
+
"alpha_mean": 0.024271922186017036,
|
| 531 |
+
"alpha_std": 0.00041434483136981726,
|
| 532 |
+
"epoch_max_grad": 0.0847236980099905
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 9,
|
| 536 |
+
"global_batch": 8784,
|
| 537 |
+
"epoch_test_mse": 3.7823650712380186e-05,
|
| 538 |
+
"train_recon": 2.585673805566837e-05,
|
| 539 |
+
"S0": 1.8655860424041748,
|
| 540 |
+
"SD": 1.3758091926574707,
|
| 541 |
+
"ratio": 1.3559918403625488,
|
| 542 |
+
"erank": 3.9603919982910156,
|
| 543 |
+
"row_cv": 0.9104012158663509,
|
| 544 |
+
"cv_in_band": true,
|
| 545 |
+
"s_delta": 0.023150920867919922,
|
| 546 |
+
"alpha_mean": 0.024292536079883575,
|
| 547 |
+
"alpha_std": 0.0004429812834132463,
|
| 548 |
+
"epoch_max_grad": 0.11537839263113685,
|
| 549 |
+
"epoch_time_s": 264.8957371711731,
|
| 550 |
+
"tree_metrics": null,
|
| 551 |
+
"sp_metrics": null,
|
| 552 |
+
"bt_metrics": {
|
| 553 |
+
"per_byte_acc": 0.4053853452205658,
|
| 554 |
+
"per_byte_l1": 0.6337992548942566,
|
| 555 |
+
"trigram_exact_rate": 0.061676025390625,
|
| 556 |
+
"per_channel_acc": [
|
| 557 |
+
0.33319091796875,
|
| 558 |
+
0.565887451171875,
|
| 559 |
+
0.31707763671875
|
| 560 |
+
]
|
| 561 |
+
}
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 10,
|
| 565 |
+
"global_batch": 9000,
|
| 566 |
+
"train_recon": 2.898303484420871e-05,
|
| 567 |
+
"test_mse": 6.730435416102409e-06,
|
| 568 |
+
"S0": 1.8589872121810913,
|
| 569 |
+
"SD": 1.373025894165039,
|
| 570 |
+
"ratio": 1.3539345518113686,
|
| 571 |
+
"erank": 3.960310935974121,
|
| 572 |
+
"row_cv": 0.9441000601060248,
|
| 573 |
+
"cv_in_band": true,
|
| 574 |
+
"s_delta": 0.023119250312447548,
|
| 575 |
+
"alpha_mean": 0.024307236075401306,
|
| 576 |
+
"alpha_std": 0.0004712300142273307,
|
| 577 |
+
"epoch_max_grad": 0.07803582601914096
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 10,
|
| 581 |
+
"global_batch": 9500,
|
| 582 |
+
"train_recon": 2.1805106575684783e-05,
|
| 583 |
+
"test_mse": 5.273246642900631e-06,
|
| 584 |
+
"S0": 1.8565095663070679,
|
| 585 |
+
"SD": 1.365691900253296,
|
| 586 |
+
"ratio": 1.3593912011697717,
|
| 587 |
+
"erank": 3.960419178009033,
|
| 588 |
+
"row_cv": 0.8886652959533914,
|
| 589 |
+
"cv_in_band": true,
|
| 590 |
+
"s_delta": 0.02312673255801201,
|
| 591 |
+
"alpha_mean": 0.02432398870587349,
|
| 592 |
+
"alpha_std": 0.0005035317735746503,
|
| 593 |
+
"epoch_max_grad": 0.07803582601914096
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 10,
|
| 597 |
+
"global_batch": 9760,
|
| 598 |
+
"epoch_test_mse": 7.078125041880412e-06,
|
| 599 |
+
"train_recon": 2.5987710482714186e-05,
|
| 600 |
+
"S0": 1.8470972776412964,
|
| 601 |
+
"SD": 1.3530970811843872,
|
| 602 |
+
"ratio": 1.3650884628295898,
|
| 603 |
+
"erank": 3.9583864212036133,
|
| 604 |
+
"row_cv": 0.8460402120105749,
|
| 605 |
+
"cv_in_band": true,
|
| 606 |
+
"s_delta": 0.023035556077957153,
|
| 607 |
+
"alpha_mean": 0.024357806891202927,
|
| 608 |
+
"alpha_std": 0.0005525918095372617,
|
| 609 |
+
"epoch_max_grad": 0.09552203030222949,
|
| 610 |
+
"epoch_time_s": 264.86489939689636,
|
| 611 |
+
"tree_metrics": null,
|
| 612 |
+
"sp_metrics": null,
|
| 613 |
+
"bt_metrics": {
|
| 614 |
+
"per_byte_acc": 0.9215189814567566,
|
| 615 |
+
"per_byte_l1": 0.0839945450425148,
|
| 616 |
+
"trigram_exact_rate": 0.808837890625,
|
| 617 |
+
"per_channel_acc": [
|
| 618 |
+
0.919097900390625,
|
| 619 |
+
0.919769287109375,
|
| 620 |
+
0.925689697265625
|
| 621 |
+
]
|
| 622 |
+
}
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 11,
|
| 626 |
+
"global_batch": 10000,
|
| 627 |
+
"train_recon": 5.731945795635814e-06,
|
| 628 |
+
"test_mse": 4.601793989422731e-06,
|
| 629 |
+
"S0": 1.857172966003418,
|
| 630 |
+
"SD": 1.3553857803344727,
|
| 631 |
+
"ratio": 1.3702172320584223,
|
| 632 |
+
"erank": 3.9593253135681152,
|
| 633 |
+
"row_cv": 0.8029046316669253,
|
| 634 |
+
"cv_in_band": true,
|
| 635 |
+
"s_delta": 0.023208508267998695,
|
| 636 |
+
"alpha_mean": 0.024350006133317947,
|
| 637 |
+
"alpha_std": 0.0005434461054392159,
|
| 638 |
+
"epoch_max_grad": 0.008080845067884597
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 11,
|
| 642 |
+
"global_batch": 10500,
|
| 643 |
+
"train_recon": 2.3559292787615197e-05,
|
| 644 |
+
"test_mse": 5.2057903303648345e-06,
|
| 645 |
+
"S0": 1.8451991081237793,
|
| 646 |
+
"SD": 1.3561946153640747,
|
| 647 |
+
"ratio": 1.3605710224875947,
|
| 648 |
+
"erank": 3.9597861766815186,
|
| 649 |
+
"row_cv": 0.90357892807925,
|
| 650 |
+
"cv_in_band": true,
|
| 651 |
+
"s_delta": 0.02329949475824833,
|
| 652 |
+
"alpha_mean": 0.024388881400227547,
|
| 653 |
+
"alpha_std": 0.0006122316117398441,
|
| 654 |
+
"epoch_max_grad": 0.13666417813071224
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 11,
|
| 658 |
+
"global_batch": 10736,
|
| 659 |
+
"epoch_test_mse": 4.6204135236621365e-06,
|
| 660 |
+
"train_recon": 2.230211508004522e-05,
|
| 661 |
+
"S0": 1.8451106548309326,
|
| 662 |
+
"SD": 1.3541594743728638,
|
| 663 |
+
"ratio": 1.3625504970550537,
|
| 664 |
+
"erank": 3.959346294403076,
|
| 665 |
+
"row_cv": 0.8612691280041216,
|
| 666 |
+
"cv_in_band": true,
|
| 667 |
+
"s_delta": 0.023319482803344727,
|
| 668 |
+
"alpha_mean": 0.024397600442171097,
|
| 669 |
+
"alpha_std": 0.0006294787744991481,
|
| 670 |
+
"epoch_max_grad": 0.13666417813071224,
|
| 671 |
+
"epoch_time_s": 264.89744544029236,
|
| 672 |
+
"tree_metrics": null,
|
| 673 |
+
"sp_metrics": null,
|
| 674 |
+
"bt_metrics": {
|
| 675 |
+
"per_byte_acc": 0.9615885615348816,
|
| 676 |
+
"per_byte_l1": 0.0421142578125,
|
| 677 |
+
"trigram_exact_rate": 0.90850830078125,
|
| 678 |
+
"per_channel_acc": [
|
| 679 |
+
0.9603271484375,
|
| 680 |
+
0.9595947265625,
|
| 681 |
+
0.96484375
|
| 682 |
+
]
|
| 683 |
+
}
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 12,
|
| 687 |
+
"global_batch": 11000,
|
| 688 |
+
"train_recon": 2.2296813447362432e-05,
|
| 689 |
+
"test_mse": 0.0003127831732854247,
|
| 690 |
+
"S0": 1.8383612632751465,
|
| 691 |
+
"SD": 1.3491995334625244,
|
| 692 |
+
"ratio": 1.3625569858682725,
|
| 693 |
+
"erank": 3.9591455459594727,
|
| 694 |
+
"row_cv": 0.9658353629126867,
|
| 695 |
+
"cv_in_band": true,
|
| 696 |
+
"s_delta": 0.0232902429997921,
|
| 697 |
+
"alpha_mean": 0.02440892532467842,
|
| 698 |
+
"alpha_std": 0.0006490845116786659,
|
| 699 |
+
"epoch_max_grad": 0.11222803985203927
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 12,
|
| 703 |
+
"global_batch": 11500,
|
| 704 |
+
"train_recon": 1.8817835454102583e-05,
|
| 705 |
+
"test_mse": 0.00022196389909368008,
|
| 706 |
+
"S0": 1.8365789651870728,
|
| 707 |
+
"SD": 1.3308309316635132,
|
| 708 |
+
"ratio": 1.3800242447710032,
|
| 709 |
+
"erank": 3.9577579498291016,
|
| 710 |
+
"row_cv": 1.0607436135575226,
|
| 711 |
+
"cv_in_band": true,
|
| 712 |
+
"s_delta": 0.023097924888134003,
|
| 713 |
+
"alpha_mean": 0.024427657946944237,
|
| 714 |
+
"alpha_std": 0.0006872243247926235,
|
| 715 |
+
"epoch_max_grad": 0.11222803985203927
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 12,
|
| 719 |
+
"global_batch": 11712,
|
| 720 |
+
"epoch_test_mse": 4.8089221719419585e-05,
|
| 721 |
+
"train_recon": 2.0316920018767184e-05,
|
| 722 |
+
"S0": 1.8371564149856567,
|
| 723 |
+
"SD": 1.3296172618865967,
|
| 724 |
+
"ratio": 1.3817182779312134,
|
| 725 |
+
"erank": 3.957249164581299,
|
| 726 |
+
"row_cv": 0.9653443213848032,
|
| 727 |
+
"cv_in_band": true,
|
| 728 |
+
"s_delta": 0.02316001057624817,
|
| 729 |
+
"alpha_mean": 0.024447325617074966,
|
| 730 |
+
"alpha_std": 0.000717672985047102,
|
| 731 |
+
"epoch_max_grad": 0.11222803985203927,
|
| 732 |
+
"epoch_time_s": 264.9489116668701,
|
| 733 |
+
"tree_metrics": null,
|
| 734 |
+
"sp_metrics": null,
|
| 735 |
+
"bt_metrics": {
|
| 736 |
+
"per_byte_acc": 0.4530741274356842,
|
| 737 |
+
"per_byte_l1": 0.65081787109375,
|
| 738 |
+
"trigram_exact_rate": 0.071533203125,
|
| 739 |
+
"per_channel_acc": [
|
| 740 |
+
0.4356689453125,
|
| 741 |
+
0.41998291015625,
|
| 742 |
+
0.503570556640625
|
| 743 |
+
]
|
| 744 |
+
}
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 13,
|
| 748 |
+
"global_batch": 12000,
|
| 749 |
+
"train_recon": 1.0637479550249887e-05,
|
| 750 |
+
"test_mse": 3.7256772884575184e-06,
|
| 751 |
+
"S0": 1.8492767810821533,
|
| 752 |
+
"SD": 1.3363832235336304,
|
| 753 |
+
"ratio": 1.3837922645829244,
|
| 754 |
+
"erank": 3.957552433013916,
|
| 755 |
+
"row_cv": 0.8674563104781582,
|
| 756 |
+
"cv_in_band": true,
|
| 757 |
+
"s_delta": 0.02337268553674221,
|
| 758 |
+
"alpha_mean": 0.02445177547633648,
|
| 759 |
+
"alpha_std": 0.000727581325918436,
|
| 760 |
+
"epoch_max_grad": 0.039296803470778026
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 13,
|
| 764 |
+
"global_batch": 12500,
|
| 765 |
+
"train_recon": 1.5155626144364137e-05,
|
| 766 |
+
"test_mse": 4.891643584414851e-06,
|
| 767 |
+
"S0": 1.8457717895507812,
|
| 768 |
+
"SD": 1.3365625143051147,
|
| 769 |
+
"ratio": 1.3809842457691284,
|
| 770 |
+
"erank": 3.9576780796051025,
|
| 771 |
+
"row_cv": 0.9093181838470336,
|
| 772 |
+
"cv_in_band": true,
|
| 773 |
+
"s_delta": 0.0234711691737175,
|
| 774 |
+
"alpha_mean": 0.024478310719132423,
|
| 775 |
+
"alpha_std": 0.0007698611007072031,
|
| 776 |
+
"epoch_max_grad": 0.07600619345630438
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 13,
|
| 780 |
+
"global_batch": 12688,
|
| 781 |
+
"epoch_test_mse": 4.819459557620576e-06,
|
| 782 |
+
"train_recon": 1.6856996553803477e-05,
|
| 783 |
+
"S0": 1.8498460054397583,
|
| 784 |
+
"SD": 1.3305994272232056,
|
| 785 |
+
"ratio": 1.3902350664138794,
|
| 786 |
+
"erank": 3.9562833309173584,
|
| 787 |
+
"row_cv": 0.9360308767181038,
|
| 788 |
+
"cv_in_band": true,
|
| 789 |
+
"s_delta": 0.023469507694244385,
|
| 790 |
+
"alpha_mean": 0.02449227310717106,
|
| 791 |
+
"alpha_std": 0.0007953800959512591,
|
| 792 |
+
"epoch_max_grad": 0.07600619345630438,
|
| 793 |
+
"epoch_time_s": 264.903990983963,
|
| 794 |
+
"tree_metrics": null,
|
| 795 |
+
"sp_metrics": null,
|
| 796 |
+
"bt_metrics": {
|
| 797 |
+
"per_byte_acc": 0.954925537109375,
|
| 798 |
+
"per_byte_l1": 0.047943115234375,
|
| 799 |
+
"trigram_exact_rate": 0.888214111328125,
|
| 800 |
+
"per_channel_acc": [
|
| 801 |
+
0.964630126953125,
|
| 802 |
+
0.931793212890625,
|
| 803 |
+
0.968353271484375
|
| 804 |
+
]
|
| 805 |
+
}
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 14,
|
| 809 |
+
"global_batch": 13000,
|
| 810 |
+
"train_recon": 9.88568507919715e-06,
|
| 811 |
+
"test_mse": 3.992892743553966e-06,
|
| 812 |
+
"S0": 1.8515442609786987,
|
| 813 |
+
"SD": 1.3293683528900146,
|
| 814 |
+
"ratio": 1.392800003870625,
|
| 815 |
+
"erank": 3.956265926361084,
|
| 816 |
+
"row_cv": 0.8749815083453858,
|
| 817 |
+
"cv_in_band": true,
|
| 818 |
+
"s_delta": 0.023500435054302216,
|
| 819 |
+
"alpha_mean": 0.02449673041701317,
|
| 820 |
+
"alpha_std": 0.0008049673051573336,
|
| 821 |
+
"epoch_max_grad": 0.0349853385959494
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 14,
|
| 825 |
+
"global_batch": 13500,
|
| 826 |
+
"train_recon": 1.5075184218663448e-05,
|
| 827 |
+
"test_mse": 8.909645111998543e-05,
|
| 828 |
+
"S0": 1.8651354312896729,
|
| 829 |
+
"SD": 1.3346917629241943,
|
| 830 |
+
"ratio": 1.397427832497478,
|
| 831 |
+
"erank": 3.954793691635132,
|
| 832 |
+
"row_cv": 0.9502741133137687,
|
| 833 |
+
"cv_in_band": true,
|
| 834 |
+
"s_delta": 0.023693060502409935,
|
| 835 |
+
"alpha_mean": 0.024518676102161407,
|
| 836 |
+
"alpha_std": 0.000853365461807698,
|
| 837 |
+
"epoch_max_grad": 0.07213190028709508
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 14,
|
| 841 |
+
"global_batch": 13664,
|
| 842 |
+
"epoch_test_mse": 3.640959196491167e-06,
|
| 843 |
+
"train_recon": 1.4261653339205071e-05,
|
| 844 |
+
"S0": 1.855076551437378,
|
| 845 |
+
"SD": 1.3203482627868652,
|
| 846 |
+
"ratio": 1.404990315437317,
|
| 847 |
+
"erank": 3.9543099403381348,
|
| 848 |
+
"row_cv": 0.7872690126587416,
|
| 849 |
+
"cv_in_band": false,
|
| 850 |
+
"s_delta": 0.023485839366912842,
|
| 851 |
+
"alpha_mean": 0.024524005129933357,
|
| 852 |
+
"alpha_std": 0.0008624386391602457,
|
| 853 |
+
"epoch_max_grad": 0.07213190028709508,
|
| 854 |
+
"epoch_time_s": 264.9932293891907,
|
| 855 |
+
"tree_metrics": null,
|
| 856 |
+
"sp_metrics": null,
|
| 857 |
+
"bt_metrics": {
|
| 858 |
+
"per_byte_acc": 0.9734293818473816,
|
| 859 |
+
"per_byte_l1": 0.02913411520421505,
|
| 860 |
+
"trigram_exact_rate": 0.937713623046875,
|
| 861 |
+
"per_channel_acc": [
|
| 862 |
+
0.9693603515625,
|
| 863 |
+
0.973846435546875,
|
| 864 |
+
0.977081298828125
|
| 865 |
+
]
|
| 866 |
+
}
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 15,
|
| 870 |
+
"global_batch": 14000,
|
| 871 |
+
"train_recon": 1.3662750062405307e-05,
|
| 872 |
+
"test_mse": 3.602173592298641e-06,
|
| 873 |
+
"S0": 1.855586051940918,
|
| 874 |
+
"SD": 1.323643445968628,
|
| 875 |
+
"ratio": 1.4018775551479763,
|
| 876 |
+
"erank": 3.9542250633239746,
|
| 877 |
+
"row_cv": 0.8276929344914649,
|
| 878 |
+
"cv_in_band": true,
|
| 879 |
+
"s_delta": 0.023541931062936783,
|
| 880 |
+
"alpha_mean": 0.024540338665246964,
|
| 881 |
+
"alpha_std": 0.0008879150846041739,
|
| 882 |
+
"epoch_max_grad": 0.05301195512255295
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 15,
|
| 886 |
+
"global_batch": 14500,
|
| 887 |
+
"train_recon": 1.3209962772410644e-05,
|
| 888 |
+
"test_mse": 5.860197234142106e-06,
|
| 889 |
+
"S0": 1.8600354194641113,
|
| 890 |
+
"SD": 1.325706958770752,
|
| 891 |
+
"ratio": 1.4030517024352749,
|
| 892 |
+
"erank": 3.9539589881896973,
|
| 893 |
+
"row_cv": 0.9561914717959945,
|
| 894 |
+
"cv_in_band": true,
|
| 895 |
+
"s_delta": 0.023643940687179565,
|
| 896 |
+
"alpha_mean": 0.024561237543821335,
|
| 897 |
+
"alpha_std": 0.0009197407634928823,
|
| 898 |
+
"epoch_max_grad": 0.05301195512255295
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 15,
|
| 902 |
+
"global_batch": 14640,
|
| 903 |
+
"epoch_test_mse": 1.918262607359793e-05,
|
| 904 |
+
"train_recon": 1.2814545847677958e-05,
|
| 905 |
+
"S0": 1.8623089790344238,
|
| 906 |
+
"SD": 1.3224735260009766,
|
| 907 |
+
"ratio": 1.4082013368606567,
|
| 908 |
+
"erank": 3.953677177429199,
|
| 909 |
+
"row_cv": 0.9636423462972895,
|
| 910 |
+
"cv_in_band": true,
|
| 911 |
+
"s_delta": 0.02364325523376465,
|
| 912 |
+
"alpha_mean": 0.024563707411289215,
|
| 913 |
+
"alpha_std": 0.0009250469738617539,
|
| 914 |
+
"epoch_max_grad": 0.05301195512255295,
|
| 915 |
+
"epoch_time_s": 264.9079806804657,
|
| 916 |
+
"tree_metrics": null,
|
| 917 |
+
"sp_metrics": null,
|
| 918 |
+
"bt_metrics": {
|
| 919 |
+
"per_byte_acc": 0.64532470703125,
|
| 920 |
+
"per_byte_l1": 0.3618570864200592,
|
| 921 |
+
"trigram_exact_rate": 0.229766845703125,
|
| 922 |
+
"per_channel_acc": [
|
| 923 |
+
0.415863037109375,
|
| 924 |
+
0.7127685546875,
|
| 925 |
+
0.807342529296875
|
| 926 |
+
]
|
| 927 |
+
}
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 16,
|
| 931 |
+
"global_batch": 15000,
|
| 932 |
+
"train_recon": 1.0087842372917699e-05,
|
| 933 |
+
"test_mse": 4.082298983121291e-05,
|
| 934 |
+
"S0": 1.862236738204956,
|
| 935 |
+
"SD": 1.3193862438201904,
|
| 936 |
+
"ratio": 1.4114416705593076,
|
| 937 |
+
"erank": 3.953678607940674,
|
| 938 |
+
"row_cv": 0.9472048207138313,
|
| 939 |
+
"cv_in_band": true,
|
| 940 |
+
"s_delta": 0.023626556620001793,
|
| 941 |
+
"alpha_mean": 0.02457253634929657,
|
| 942 |
+
"alpha_std": 0.0009433587547391653,
|
| 943 |
+
"epoch_max_grad": 0.03360610211129723
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 16,
|
| 947 |
+
"global_batch": 15500,
|
| 948 |
+
"train_recon": 1.0968404909669706e-05,
|
| 949 |
+
"test_mse": 4.504205207922496e-06,
|
| 950 |
+
"S0": 1.8633925914764404,
|
| 951 |
+
"SD": 1.313594937324524,
|
| 952 |
+
"ratio": 1.4185442744521222,
|
| 953 |
+
"erank": 3.9526610374450684,
|
| 954 |
+
"row_cv": 1.0599272130445612,
|
| 955 |
+
"cv_in_band": true,
|
| 956 |
+
"s_delta": 0.02363010123372078,
|
| 957 |
+
"alpha_mean": 0.024590596556663513,
|
| 958 |
+
"alpha_std": 0.0009782082634046674,
|
| 959 |
+
"epoch_max_grad": 0.035077086279493096
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 16,
|
| 963 |
+
"global_batch": 15616,
|
| 964 |
+
"epoch_test_mse": 3.9244455991138234e-06,
|
| 965 |
+
"train_recon": 1.0248088137748963e-05,
|
| 966 |
+
"S0": 1.8596898317337036,
|
| 967 |
+
"SD": 1.3099844455718994,
|
| 968 |
+
"ratio": 1.4196274280548096,
|
| 969 |
+
"erank": 3.952789068222046,
|
| 970 |
+
"row_cv": 0.8786962449037597,
|
| 971 |
+
"cv_in_band": true,
|
| 972 |
+
"s_delta": 0.023593008518218994,
|
| 973 |
+
"alpha_mean": 0.024590328335762024,
|
| 974 |
+
"alpha_std": 0.0009786387672647834,
|
| 975 |
+
"epoch_max_grad": 0.035077086279493096,
|
| 976 |
+
"epoch_time_s": 264.895117521286,
|
| 977 |
+
"tree_metrics": null,
|
| 978 |
+
"sp_metrics": null,
|
| 979 |
+
"bt_metrics": {
|
| 980 |
+
"per_byte_acc": 0.9686686396598816,
|
| 981 |
+
"per_byte_l1": 0.0332234688103199,
|
| 982 |
+
"trigram_exact_rate": 0.921966552734375,
|
| 983 |
+
"per_channel_acc": [
|
| 984 |
+
0.96380615234375,
|
| 985 |
+
0.97564697265625,
|
| 986 |
+
0.966552734375
|
| 987 |
+
]
|
| 988 |
+
}
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 17,
|
| 992 |
+
"global_batch": 16000,
|
| 993 |
+
"train_recon": 1.4977885254433923e-05,
|
| 994 |
+
"test_mse": 5.926080120843835e-06,
|
| 995 |
+
"S0": 1.8676445484161377,
|
| 996 |
+
"SD": 1.3081820011138916,
|
| 997 |
+
"ratio": 1.427664141953668,
|
| 998 |
+
"erank": 3.9516451358795166,
|
| 999 |
+
"row_cv": 0.9809350785600964,
|
| 1000 |
+
"cv_in_band": true,
|
| 1001 |
+
"s_delta": 0.023672420531511307,
|
| 1002 |
+
"alpha_mean": 0.02461244724690914,
|
| 1003 |
+
"alpha_std": 0.0010140499798581004,
|
| 1004 |
+
"epoch_max_grad": 0.054045017385208456
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 17,
|
| 1008 |
+
"global_batch": 16500,
|
| 1009 |
+
"train_recon": 1.2643991188658767e-05,
|
| 1010 |
+
"test_mse": 3.0452235932898475e-06,
|
| 1011 |
+
"S0": 1.8687657117843628,
|
| 1012 |
+
"SD": 1.309316635131836,
|
| 1013 |
+
"ratio": 1.427283246365661,
|
| 1014 |
+
"erank": 3.951298236846924,
|
| 1015 |
+
"row_cv": 1.0539028650075373,
|
| 1016 |
+
"cv_in_band": true,
|
| 1017 |
+
"s_delta": 0.02373437210917473,
|
| 1018 |
+
"alpha_mean": 0.024631358683109283,
|
| 1019 |
+
"alpha_std": 0.0010460406774654984,
|
| 1020 |
+
"epoch_max_grad": 0.054045017385208456
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 17,
|
| 1024 |
+
"global_batch": 16592,
|
| 1025 |
+
"epoch_test_mse": 1.930016307742335e-05,
|
| 1026 |
+
"train_recon": 1.2538094923041201e-05,
|
| 1027 |
+
"S0": 1.8689783811569214,
|
| 1028 |
+
"SD": 1.306301474571228,
|
| 1029 |
+
"ratio": 1.430740475654602,
|
| 1030 |
+
"erank": 3.9507575035095215,
|
| 1031 |
+
"row_cv": 1.0467304319937745,
|
| 1032 |
+
"cv_in_band": true,
|
| 1033 |
+
"s_delta": 0.02369856834411621,
|
| 1034 |
+
"alpha_mean": 0.024634912610054016,
|
| 1035 |
+
"alpha_std": 0.001051040249876678,
|
| 1036 |
+
"epoch_max_grad": 0.054045017385208456,
|
| 1037 |
+
"epoch_time_s": 264.9195246696472,
|
| 1038 |
+
"tree_metrics": null,
|
| 1039 |
+
"sp_metrics": null,
|
| 1040 |
+
"bt_metrics": {
|
| 1041 |
+
"per_byte_acc": 0.5840250849723816,
|
| 1042 |
+
"per_byte_l1": 0.42108154296875,
|
| 1043 |
+
"trigram_exact_rate": 0.145599365234375,
|
| 1044 |
+
"per_channel_acc": [
|
| 1045 |
+
0.615997314453125,
|
| 1046 |
+
0.54840087890625,
|
| 1047 |
+
0.587677001953125
|
| 1048 |
+
]
|
| 1049 |
+
}
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 18,
|
| 1053 |
+
"global_batch": 17000,
|
| 1054 |
+
"train_recon": 8.509423365159605e-06,
|
| 1055 |
+
"test_mse": 2.9721086320932955e-06,
|
| 1056 |
+
"S0": 1.8700101375579834,
|
| 1057 |
+
"SD": 1.3031258583068848,
|
| 1058 |
+
"ratio": 1.4350188136374238,
|
| 1059 |
+
"erank": 3.950408935546875,
|
| 1060 |
+
"row_cv": 0.8826772459602501,
|
| 1061 |
+
"cv_in_band": true,
|
| 1062 |
+
"s_delta": 0.023708632215857506,
|
| 1063 |
+
"alpha_mean": 0.024642694741487503,
|
| 1064 |
+
"alpha_std": 0.0010696150129660964,
|
| 1065 |
+
"epoch_max_grad": 0.04395828563134251
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 18,
|
| 1069 |
+
"global_batch": 17500,
|
| 1070 |
+
"train_recon": 8.701313100224743e-06,
|
| 1071 |
+
"test_mse": 6.11140558248735e-06,
|
| 1072 |
+
"S0": 1.8732001781463623,
|
| 1073 |
+
"SD": 1.3018431663513184,
|
| 1074 |
+
"ratio": 1.4388831252289456,
|
| 1075 |
+
"erank": 3.9499473571777344,
|
| 1076 |
+
"row_cv": 0.9242769957223593,
|
| 1077 |
+
"cv_in_band": true,
|
| 1078 |
+
"s_delta": 0.023763004690408707,
|
| 1079 |
+
"alpha_mean": 0.02465570904314518,
|
| 1080 |
+
"alpha_std": 0.0010949858697131276,
|
| 1081 |
+
"epoch_max_grad": 0.04395828563134251
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 18,
|
| 1085 |
+
"global_batch": 17568,
|
| 1086 |
+
"epoch_test_mse": 3.8991663903289005e-06,
|
| 1087 |
+
"train_recon": 8.533190589425848e-06,
|
| 1088 |
+
"S0": 1.870848536491394,
|
| 1089 |
+
"SD": 1.2997833490371704,
|
| 1090 |
+
"ratio": 1.4393541812896729,
|
| 1091 |
+
"erank": 3.950000524520874,
|
| 1092 |
+
"row_cv": 0.9068477236213712,
|
| 1093 |
+
"cv_in_band": true,
|
| 1094 |
+
"s_delta": 0.02373790740966797,
|
| 1095 |
+
"alpha_mean": 0.024656642228364944,
|
| 1096 |
+
"alpha_std": 0.001096972730010748,
|
| 1097 |
+
"epoch_max_grad": 0.04395828563134251,
|
| 1098 |
+
"epoch_time_s": 264.92607593536377,
|
| 1099 |
+
"tree_metrics": null,
|
| 1100 |
+
"sp_metrics": null,
|
| 1101 |
+
"bt_metrics": {
|
| 1102 |
+
"per_byte_acc": 0.964599609375,
|
| 1103 |
+
"per_byte_l1": 0.036895751953125,
|
| 1104 |
+
"trigram_exact_rate": 0.908538818359375,
|
| 1105 |
+
"per_channel_acc": [
|
| 1106 |
+
0.93402099609375,
|
| 1107 |
+
0.975677490234375,
|
| 1108 |
+
0.984100341796875
|
| 1109 |
+
]
|
| 1110 |
+
}
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 19,
|
| 1114 |
+
"global_batch": 18000,
|
| 1115 |
+
"train_recon": 1.0037494464898845e-05,
|
| 1116 |
+
"test_mse": 7.482302862626966e-06,
|
| 1117 |
+
"S0": 1.8752045631408691,
|
| 1118 |
+
"SD": 1.2986069917678833,
|
| 1119 |
+
"ratio": 1.4440123613903388,
|
| 1120 |
+
"erank": 3.948704242706299,
|
| 1121 |
+
"row_cv": 0.9586050351917162,
|
| 1122 |
+
"cv_in_band": true,
|
| 1123 |
+
"s_delta": 0.023753076791763306,
|
| 1124 |
+
"alpha_mean": 0.024673432111740112,
|
| 1125 |
+
"alpha_std": 0.001122961868532002,
|
| 1126 |
+
"epoch_max_grad": 0.04430315572900294
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 19,
|
| 1130 |
+
"global_batch": 18500,
|
| 1131 |
+
"train_recon": 9.165165217075892e-06,
|
| 1132 |
+
"test_mse": 2.782782303256681e-06,
|
| 1133 |
+
"S0": 1.8744301795959473,
|
| 1134 |
+
"SD": 1.2963218688964844,
|
| 1135 |
+
"ratio": 1.4459604594435969,
|
| 1136 |
+
"erank": 3.9486470222473145,
|
| 1137 |
+
"row_cv": 0.9187378581167366,
|
| 1138 |
+
"cv_in_band": true,
|
| 1139 |
+
"s_delta": 0.023753218352794647,
|
| 1140 |
+
"alpha_mean": 0.024685390293598175,
|
| 1141 |
+
"alpha_std": 0.0011491116601973772,
|
| 1142 |
+
"epoch_max_grad": 0.04430315572900294
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 19,
|
| 1146 |
+
"global_batch": 18544,
|
| 1147 |
+
"epoch_test_mse": 6.031918941152981e-06,
|
| 1148 |
+
"train_recon": 9.307905113156847e-06,
|
| 1149 |
+
"S0": 1.8758759498596191,
|
| 1150 |
+
"SD": 1.295669674873352,
|
| 1151 |
+
"ratio": 1.4478042125701904,
|
| 1152 |
+
"erank": 3.94842267036438,
|
| 1153 |
+
"row_cv": 0.9371729158654708,
|
| 1154 |
+
"cv_in_band": true,
|
| 1155 |
+
"s_delta": 0.023759514093399048,
|
| 1156 |
+
"alpha_mean": 0.02468731254339218,
|
| 1157 |
+
"alpha_std": 0.0011517866514623165,
|
| 1158 |
+
"epoch_max_grad": 0.04430315572900294,
|
| 1159 |
+
"epoch_time_s": 264.9407160282135,
|
| 1160 |
+
"tree_metrics": null,
|
| 1161 |
+
"sp_metrics": null,
|
| 1162 |
+
"bt_metrics": {
|
| 1163 |
+
"per_byte_acc": 0.9275614619255066,
|
| 1164 |
+
"per_byte_l1": 0.074188232421875,
|
| 1165 |
+
"trigram_exact_rate": 0.809295654296875,
|
| 1166 |
+
"per_channel_acc": [
|
| 1167 |
+
0.933319091796875,
|
| 1168 |
+
0.92138671875,
|
| 1169 |
+
0.927978515625
|
| 1170 |
+
]
|
| 1171 |
+
}
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 20,
|
| 1175 |
+
"global_batch": 19000,
|
| 1176 |
+
"train_recon": 8.999634754321992e-06,
|
| 1177 |
+
"test_mse": 3.3996270758507308e-06,
|
| 1178 |
+
"S0": 1.8776209354400635,
|
| 1179 |
+
"SD": 1.2958753108978271,
|
| 1180 |
+
"ratio": 1.4489209765482554,
|
| 1181 |
+
"erank": 3.9474596977233887,
|
| 1182 |
+
"row_cv": 0.9299753657676948,
|
| 1183 |
+
"cv_in_band": true,
|
| 1184 |
+
"s_delta": 0.023802880197763443,
|
| 1185 |
+
"alpha_mean": 0.024699609726667404,
|
| 1186 |
+
"alpha_std": 0.0011784255038946867,
|
| 1187 |
+
"epoch_max_grad": 0.0547859312751362
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 20,
|
| 1191 |
+
"global_batch": 19500,
|
| 1192 |
+
"train_recon": 8.145814777323329e-06,
|
| 1193 |
+
"test_mse": 2.2938884285395034e-06,
|
| 1194 |
+
"S0": 1.8800435066223145,
|
| 1195 |
+
"SD": 1.2898519039154053,
|
| 1196 |
+
"ratio": 1.4575653889719444,
|
| 1197 |
+
"erank": 3.946781873703003,
|
| 1198 |
+
"row_cv": 0.8976492881391739,
|
| 1199 |
+
"cv_in_band": true,
|
| 1200 |
+
"s_delta": 0.02379540726542473,
|
| 1201 |
+
"alpha_mean": 0.024710137397050858,
|
| 1202 |
+
"alpha_std": 0.0011991638457402587,
|
| 1203 |
+
"epoch_max_grad": 0.0547859312751362
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 20,
|
| 1207 |
+
"global_batch": 19520,
|
| 1208 |
+
"epoch_test_mse": 3.1082244699064175e-06,
|
| 1209 |
+
"train_recon": 8.056532156949109e-06,
|
| 1210 |
+
"S0": 1.8790555000305176,
|
| 1211 |
+
"SD": 1.2892391681671143,
|
| 1212 |
+
"ratio": 1.4574917554855347,
|
| 1213 |
+
"erank": 3.9468722343444824,
|
| 1214 |
+
"row_cv": 0.8976492881391739,
|
| 1215 |
+
"cv_in_band": true,
|
| 1216 |
+
"s_delta": 0.023787736892700195,
|
| 1217 |
+
"alpha_mean": 0.024709969758987427,
|
| 1218 |
+
"alpha_std": 0.0011990320635959506,
|
| 1219 |
+
"epoch_max_grad": 0.0547859312751362,
|
| 1220 |
+
"epoch_time_s": 264.9473309516907,
|
| 1221 |
+
"tree_metrics": null,
|
| 1222 |
+
"sp_metrics": null,
|
| 1223 |
+
"bt_metrics": {
|
| 1224 |
+
"per_byte_acc": 0.9798075556755066,
|
| 1225 |
+
"per_byte_l1": 0.02144368551671505,
|
| 1226 |
+
"trigram_exact_rate": 0.949615478515625,
|
| 1227 |
+
"per_channel_acc": [
|
| 1228 |
+
0.98480224609375,
|
| 1229 |
+
0.97265625,
|
| 1230 |
+
0.981964111328125
|
| 1231 |
+
]
|
| 1232 |
+
}
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 21,
|
| 1236 |
+
"global_batch": 20000,
|
| 1237 |
+
"train_recon": 7.21766799595495e-06,
|
| 1238 |
+
"test_mse": 3.4119173051294638e-06,
|
| 1239 |
+
"S0": 1.8819869756698608,
|
| 1240 |
+
"SD": 1.2887976169586182,
|
| 1241 |
+
"ratio": 1.4602657052613346,
|
| 1242 |
+
"erank": 3.9464282989501953,
|
| 1243 |
+
"row_cv": 0.9283857649016127,
|
| 1244 |
+
"cv_in_band": true,
|
| 1245 |
+
"s_delta": 0.0238087996840477,
|
| 1246 |
+
"alpha_mean": 0.024722110480070114,
|
| 1247 |
+
"alpha_std": 0.0012208169791847467,
|
| 1248 |
+
"epoch_max_grad": 0.027135417859868675
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 21,
|
| 1252 |
+
"global_batch": 20496,
|
| 1253 |
+
"epoch_test_mse": 3.5418808503891343e-06,
|
| 1254 |
+
"train_recon": 8.514564468510635e-06,
|
| 1255 |
+
"S0": 1.8842278718948364,
|
| 1256 |
+
"SD": 1.281954288482666,
|
| 1257 |
+
"ratio": 1.4698089361190796,
|
| 1258 |
+
"erank": 3.9441633224487305,
|
| 1259 |
+
"row_cv": 0.9666951218213806,
|
| 1260 |
+
"cv_in_band": true,
|
| 1261 |
+
"s_delta": 0.023718446493148804,
|
| 1262 |
+
"alpha_mean": 0.024737326428294182,
|
| 1263 |
+
"alpha_std": 0.0012542594922706485,
|
| 1264 |
+
"epoch_max_grad": 0.06165194631497287,
|
| 1265 |
+
"epoch_time_s": 264.5668349266052,
|
| 1266 |
+
"tree_metrics": null,
|
| 1267 |
+
"sp_metrics": null,
|
| 1268 |
+
"bt_metrics": {
|
| 1269 |
+
"per_byte_acc": 0.9682820439338684,
|
| 1270 |
+
"per_byte_l1": 0.0329691581428051,
|
| 1271 |
+
"trigram_exact_rate": 0.920440673828125,
|
| 1272 |
+
"per_channel_acc": [
|
| 1273 |
+
0.9827880859375,
|
| 1274 |
+
0.945281982421875,
|
| 1275 |
+
0.976776123046875
|
| 1276 |
+
]
|
| 1277 |
+
}
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 22,
|
| 1281 |
+
"global_batch": 20500,
|
| 1282 |
+
"train_recon": 2.700222012208542e-06,
|
| 1283 |
+
"test_mse": 3.2569546419836115e-06,
|
| 1284 |
+
"S0": 1.8844757080078125,
|
| 1285 |
+
"SD": 1.2816991806030273,
|
| 1286 |
+
"ratio": 1.4702948412733137,
|
| 1287 |
+
"erank": 3.9440693855285645,
|
| 1288 |
+
"row_cv": 0.8990554371313216,
|
| 1289 |
+
"cv_in_band": true,
|
| 1290 |
+
"s_delta": 0.023714708164334297,
|
| 1291 |
+
"alpha_mean": 0.024737508967518806,
|
| 1292 |
+
"alpha_std": 0.001254502683877945,
|
| 1293 |
+
"epoch_max_grad": 0.004384793066512839
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 22,
|
| 1297 |
+
"global_batch": 21000,
|
| 1298 |
+
"train_recon": 4.232956856847681e-06,
|
| 1299 |
+
"test_mse": 4.644461569114355e-06,
|
| 1300 |
+
"S0": 1.883184552192688,
|
| 1301 |
+
"SD": 1.2798237800598145,
|
| 1302 |
+
"ratio": 1.471440495808157,
|
| 1303 |
+
"erank": 3.9444022178649902,
|
| 1304 |
+
"row_cv": 0.9442300907774921,
|
| 1305 |
+
"cv_in_band": true,
|
| 1306 |
+
"s_delta": 0.02374250814318657,
|
| 1307 |
+
"alpha_mean": 0.024740368127822876,
|
| 1308 |
+
"alpha_std": 0.0012600423069670796,
|
| 1309 |
+
"epoch_max_grad": 0.02352434524739567
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 22,
|
| 1313 |
+
"global_batch": 21472,
|
| 1314 |
+
"epoch_test_mse": 1.9418056526774306e-06,
|
| 1315 |
+
"train_recon": 5.535641366991438e-06,
|
| 1316 |
+
"S0": 1.8863468170166016,
|
| 1317 |
+
"SD": 1.276906132698059,
|
| 1318 |
+
"ratio": 1.4772791862487793,
|
| 1319 |
+
"erank": 3.9436099529266357,
|
| 1320 |
+
"row_cv": 0.8796811880568727,
|
| 1321 |
+
"cv_in_band": true,
|
| 1322 |
+
"s_delta": 0.023750245571136475,
|
| 1323 |
+
"alpha_mean": 0.024751922115683556,
|
| 1324 |
+
"alpha_std": 0.0012820905540138483,
|
| 1325 |
+
"epoch_max_grad": 0.02830064850978595,
|
| 1326 |
+
"epoch_time_s": 264.82367753982544,
|
| 1327 |
+
"tree_metrics": null,
|
| 1328 |
+
"sp_metrics": null,
|
| 1329 |
+
"bt_metrics": {
|
| 1330 |
+
"per_byte_acc": 0.9899088740348816,
|
| 1331 |
+
"per_byte_l1": 0.011077880859375,
|
| 1332 |
+
"trigram_exact_rate": 0.977508544921875,
|
| 1333 |
+
"per_channel_acc": [
|
| 1334 |
+
0.98980712890625,
|
| 1335 |
+
0.99053955078125,
|
| 1336 |
+
0.9893798828125
|
| 1337 |
+
]
|
| 1338 |
+
}
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 23,
|
| 1342 |
+
"global_batch": 21500,
|
| 1343 |
+
"train_recon": 1.799134094783637e-06,
|
| 1344 |
+
"test_mse": 1.7122205235864385e-06,
|
| 1345 |
+
"S0": 1.8872833251953125,
|
| 1346 |
+
"SD": 1.2778854370117188,
|
| 1347 |
+
"ratio": 1.4768798953056748,
|
| 1348 |
+
"erank": 3.943668842315674,
|
| 1349 |
+
"row_cv": 0.9007543660111829,
|
| 1350 |
+
"cv_in_band": true,
|
| 1351 |
+
"s_delta": 0.023767396807670593,
|
| 1352 |
+
"alpha_mean": 0.02475159429013729,
|
| 1353 |
+
"alpha_std": 0.0012818154646083713,
|
| 1354 |
+
"epoch_max_grad": 0.0021126709256431763
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 23,
|
| 1358 |
+
"global_batch": 22000,
|
| 1359 |
+
"train_recon": 5.933829981395113e-06,
|
| 1360 |
+
"test_mse": 1.7642066723055905e-06,
|
| 1361 |
+
"S0": 1.8886682987213135,
|
| 1362 |
+
"SD": 1.275449514389038,
|
| 1363 |
+
"ratio": 1.4807863914693273,
|
| 1364 |
+
"erank": 3.9432995319366455,
|
| 1365 |
+
"row_cv": 0.8454809307945155,
|
| 1366 |
+
"cv_in_band": true,
|
| 1367 |
+
"s_delta": 0.0237728264182806,
|
| 1368 |
+
"alpha_mean": 0.02476242557168007,
|
| 1369 |
+
"alpha_std": 0.0013015256263315678,
|
| 1370 |
+
"epoch_max_grad": 0.034700720846555164
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 23,
|
| 1374 |
+
"global_batch": 22448,
|
| 1375 |
+
"epoch_test_mse": 5.441129588871263e-06,
|
| 1376 |
+
"train_recon": 5.817537552657161e-06,
|
| 1377 |
+
"S0": 1.8872625827789307,
|
| 1378 |
+
"SD": 1.2725155353546143,
|
| 1379 |
+
"ratio": 1.4830958843231201,
|
| 1380 |
+
"erank": 3.942934513092041,
|
| 1381 |
+
"row_cv": 0.942917916101737,
|
| 1382 |
+
"cv_in_band": true,
|
| 1383 |
+
"s_delta": 0.02374887466430664,
|
| 1384 |
+
"alpha_mean": 0.0247687716037035,
|
| 1385 |
+
"alpha_std": 0.0013206156436353922,
|
| 1386 |
+
"epoch_max_grad": 0.034700720846555164,
|
| 1387 |
+
"epoch_time_s": 265.0069115161896,
|
| 1388 |
+
"tree_metrics": null,
|
| 1389 |
+
"sp_metrics": null,
|
| 1390 |
+
"bt_metrics": {
|
| 1391 |
+
"per_byte_acc": 0.9243571162223816,
|
| 1392 |
+
"per_byte_l1": 0.0767415389418602,
|
| 1393 |
+
"trigram_exact_rate": 0.791748046875,
|
| 1394 |
+
"per_channel_acc": [
|
| 1395 |
+
0.921844482421875,
|
| 1396 |
+
0.915008544921875,
|
| 1397 |
+
0.93621826171875
|
| 1398 |
+
]
|
| 1399 |
+
}
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 24,
|
| 1403 |
+
"global_batch": 22500,
|
| 1404 |
+
"train_recon": 2.173661753291666e-06,
|
| 1405 |
+
"test_mse": 1.9297019662189996e-06,
|
| 1406 |
+
"S0": 1.8903157711029053,
|
| 1407 |
+
"SD": 1.2740178108215332,
|
| 1408 |
+
"ratio": 1.4837435867921853,
|
| 1409 |
+
"erank": 3.9429006576538086,
|
| 1410 |
+
"row_cv": 0.9992104554711861,
|
| 1411 |
+
"cv_in_band": true,
|
| 1412 |
+
"s_delta": 0.023779941722750664,
|
| 1413 |
+
"alpha_mean": 0.02476825751364231,
|
| 1414 |
+
"alpha_std": 0.0013205315917730331,
|
| 1415 |
+
"epoch_max_grad": 0.006976437433085034
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 24,
|
| 1419 |
+
"global_batch": 23000,
|
| 1420 |
+
"train_recon": 5.0985521746181055e-06,
|
| 1421 |
+
"test_mse": 2.821057933033444e-06,
|
| 1422 |
+
"S0": 1.8900506496429443,
|
| 1423 |
+
"SD": 1.2767555713653564,
|
| 1424 |
+
"ratio": 1.4803543271937243,
|
| 1425 |
+
"erank": 3.9433481693267822,
|
| 1426 |
+
"row_cv": 0.9743718690720167,
|
| 1427 |
+
"cv_in_band": true,
|
| 1428 |
+
"s_delta": 0.02382512018084526,
|
| 1429 |
+
"alpha_mean": 0.024776402860879898,
|
| 1430 |
+
"alpha_std": 0.0013376015704125166,
|
| 1431 |
+
"epoch_max_grad": 0.03848813960088124
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 24,
|
| 1435 |
+
"global_batch": 23424,
|
| 1436 |
+
"epoch_test_mse": 5.792163890146185e-06,
|
| 1437 |
+
"train_recon": 5.225492475649836e-06,
|
| 1438 |
+
"S0": 1.892228364944458,
|
| 1439 |
+
"SD": 1.2730449438095093,
|
| 1440 |
+
"ratio": 1.486379861831665,
|
| 1441 |
+
"erank": 3.942685842514038,
|
| 1442 |
+
"row_cv": 1.0108861155724824,
|
| 1443 |
+
"cv_in_band": true,
|
| 1444 |
+
"s_delta": 0.023803293704986572,
|
| 1445 |
+
"alpha_mean": 0.02478092536330223,
|
| 1446 |
+
"alpha_std": 0.0013533023884519935,
|
| 1447 |
+
"epoch_max_grad": 0.03848813960088124,
|
| 1448 |
+
"epoch_time_s": 264.91101932525635,
|
| 1449 |
+
"tree_metrics": null,
|
| 1450 |
+
"sp_metrics": null,
|
| 1451 |
+
"bt_metrics": {
|
| 1452 |
+
"per_byte_acc": 0.8973185420036316,
|
| 1453 |
+
"per_byte_l1": 0.10369873046875,
|
| 1454 |
+
"trigram_exact_rate": 0.727386474609375,
|
| 1455 |
+
"per_channel_acc": [
|
| 1456 |
+
0.8538818359375,
|
| 1457 |
+
0.87542724609375,
|
| 1458 |
+
0.962646484375
|
| 1459 |
+
]
|
| 1460 |
+
}
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 25,
|
| 1464 |
+
"global_batch": 23500,
|
| 1465 |
+
"train_recon": 2.019269734438766e-06,
|
| 1466 |
+
"test_mse": 1.614036023056542e-06,
|
| 1467 |
+
"S0": 1.8910906314849854,
|
| 1468 |
+
"SD": 1.2724519968032837,
|
| 1469 |
+
"ratio": 1.4861783559412005,
|
| 1470 |
+
"erank": 3.9426217079162598,
|
| 1471 |
+
"row_cv": 0.9567940081964768,
|
| 1472 |
+
"cv_in_band": true,
|
| 1473 |
+
"s_delta": 0.02378915809094906,
|
| 1474 |
+
"alpha_mean": 0.0247811246663332,
|
| 1475 |
+
"alpha_std": 0.0013547716662287712,
|
| 1476 |
+
"epoch_max_grad": 0.006782252583430606
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 25,
|
| 1480 |
+
"global_batch": 24000,
|
| 1481 |
+
"train_recon": 3.919130914119427e-06,
|
| 1482 |
+
"test_mse": 2.8019308047078084e-06,
|
| 1483 |
+
"S0": 1.890474796295166,
|
| 1484 |
+
"SD": 1.2698471546173096,
|
| 1485 |
+
"ratio": 1.4887419911394562,
|
| 1486 |
+
"erank": 3.942373275756836,
|
| 1487 |
+
"row_cv": 0.9207410396159802,
|
| 1488 |
+
"cv_in_band": true,
|
| 1489 |
+
"s_delta": 0.02376268431544304,
|
| 1490 |
+
"alpha_mean": 0.024784808978438377,
|
| 1491 |
+
"alpha_std": 0.0013680196134373546,
|
| 1492 |
+
"epoch_max_grad": 0.022249594031117717
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 25,
|
| 1496 |
+
"global_batch": 24400,
|
| 1497 |
+
"epoch_test_mse": 1.6711522854166104e-05,
|
| 1498 |
+
"train_recon": 4.5000072119461546e-06,
|
| 1499 |
+
"S0": 1.8971973657608032,
|
| 1500 |
+
"SD": 1.271343469619751,
|
| 1501 |
+
"ratio": 1.4922776222229004,
|
| 1502 |
+
"erank": 3.941650629043579,
|
| 1503 |
+
"row_cv": 0.9793402516613149,
|
| 1504 |
+
"cv_in_band": true,
|
| 1505 |
+
"s_delta": 0.023812711238861084,
|
| 1506 |
+
"alpha_mean": 0.024788983166217804,
|
| 1507 |
+
"alpha_std": 0.0013826903887093067,
|
| 1508 |
+
"epoch_max_grad": 0.029564742571099686,
|
| 1509 |
+
"epoch_time_s": 264.8708622455597,
|
| 1510 |
+
"tree_metrics": null,
|
| 1511 |
+
"sp_metrics": null,
|
| 1512 |
+
"bt_metrics": {
|
| 1513 |
+
"per_byte_acc": 0.6824238896369934,
|
| 1514 |
+
"per_byte_l1": 0.3255411684513092,
|
| 1515 |
+
"trigram_exact_rate": 0.21295166015625,
|
| 1516 |
+
"per_channel_acc": [
|
| 1517 |
+
0.496307373046875,
|
| 1518 |
+
0.697967529296875,
|
| 1519 |
+
0.852996826171875
|
| 1520 |
+
]
|
| 1521 |
+
}
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 26,
|
| 1525 |
+
"global_batch": 24500,
|
| 1526 |
+
"train_recon": 2.486756271764534e-06,
|
| 1527 |
+
"test_mse": 1.839945298343082e-06,
|
| 1528 |
+
"S0": 1.8933954238891602,
|
| 1529 |
+
"SD": 1.2695002555847168,
|
| 1530 |
+
"ratio": 1.4914494114084211,
|
| 1531 |
+
"erank": 3.9419116973876953,
|
| 1532 |
+
"row_cv": 1.0269778729610608,
|
| 1533 |
+
"cv_in_band": true,
|
| 1534 |
+
"s_delta": 0.023779774084687233,
|
| 1535 |
+
"alpha_mean": 0.02478945441544056,
|
| 1536 |
+
"alpha_std": 0.0013860133476555347,
|
| 1537 |
+
"epoch_max_grad": 0.012837596780953399
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 26,
|
| 1541 |
+
"global_batch": 25000,
|
| 1542 |
+
"train_recon": 3.948219066955972e-06,
|
| 1543 |
+
"test_mse": 1.2849135373471654e-06,
|
| 1544 |
+
"S0": 1.896077275276184,
|
| 1545 |
+
"SD": 1.268405556678772,
|
| 1546 |
+
"ratio": 1.4948509570491124,
|
| 1547 |
+
"erank": 3.941680431365967,
|
| 1548 |
+
"row_cv": 0.9155247682156913,
|
| 1549 |
+
"cv_in_band": true,
|
| 1550 |
+
"s_delta": 0.023800596594810486,
|
| 1551 |
+
"alpha_mean": 0.02479332499206066,
|
| 1552 |
+
"alpha_std": 0.0014012217288836837,
|
| 1553 |
+
"epoch_max_grad": 0.03169386960661586
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 26,
|
| 1557 |
+
"global_batch": 25376,
|
| 1558 |
+
"epoch_test_mse": 1.2103043244860601e-05,
|
| 1559 |
+
"train_recon": 4.066635316316151e-06,
|
| 1560 |
+
"S0": 1.8928967714309692,
|
| 1561 |
+
"SD": 1.2646763324737549,
|
| 1562 |
+
"ratio": 1.4967440366744995,
|
| 1563 |
+
"erank": 3.9413371086120605,
|
| 1564 |
+
"row_cv": 0.9680603401509859,
|
| 1565 |
+
"cv_in_band": true,
|
| 1566 |
+
"s_delta": 0.02374407649040222,
|
| 1567 |
+
"alpha_mean": 0.0247971098870039,
|
| 1568 |
+
"alpha_std": 0.0014121626736596227,
|
| 1569 |
+
"epoch_max_grad": 0.03169386960661586,
|
| 1570 |
+
"epoch_time_s": 264.87934017181396,
|
| 1571 |
+
"tree_metrics": null,
|
| 1572 |
+
"sp_metrics": null,
|
| 1573 |
+
"bt_metrics": {
|
| 1574 |
+
"per_byte_acc": 0.7305195927619934,
|
| 1575 |
+
"per_byte_l1": 0.2709452211856842,
|
| 1576 |
+
"trigram_exact_rate": 0.236968994140625,
|
| 1577 |
+
"per_channel_acc": [
|
| 1578 |
+
0.484100341796875,
|
| 1579 |
+
0.7225341796875,
|
| 1580 |
+
0.98492431640625
|
| 1581 |
+
]
|
| 1582 |
+
}
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 27,
|
| 1586 |
+
"global_batch": 25500,
|
| 1587 |
+
"train_recon": 1.9161199051517716e-06,
|
| 1588 |
+
"test_mse": 1.7177912923216354e-06,
|
| 1589 |
+
"S0": 1.8955004215240479,
|
| 1590 |
+
"SD": 1.2666597366333008,
|
| 1591 |
+
"ratio": 1.496455876617351,
|
| 1592 |
+
"erank": 3.94144344329834,
|
| 1593 |
+
"row_cv": 0.9402496056446257,
|
| 1594 |
+
"cv_in_band": true,
|
| 1595 |
+
"s_delta": 0.02378179132938385,
|
| 1596 |
+
"alpha_mean": 0.024796638637781143,
|
| 1597 |
+
"alpha_std": 0.0014129763003438711,
|
| 1598 |
+
"epoch_max_grad": 0.010592848201574794
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 27,
|
| 1602 |
+
"global_batch": 26000,
|
| 1603 |
+
"train_recon": 3.414806283166318e-06,
|
| 1604 |
+
"test_mse": 8.687167792231776e-06,
|
| 1605 |
+
"S0": 1.8939745426177979,
|
| 1606 |
+
"SD": 1.2630103826522827,
|
| 1607 |
+
"ratio": 1.4995716216083623,
|
| 1608 |
+
"erank": 3.9410252571105957,
|
| 1609 |
+
"row_cv": 0.9735757970741528,
|
| 1610 |
+
"cv_in_band": true,
|
| 1611 |
+
"s_delta": 0.023737266659736633,
|
| 1612 |
+
"alpha_mean": 0.024798842146992683,
|
| 1613 |
+
"alpha_std": 0.001427465700544417,
|
| 1614 |
+
"epoch_max_grad": 0.018845988258516636
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 27,
|
| 1618 |
+
"global_batch": 26352,
|
| 1619 |
+
"epoch_test_mse": 8.883682284795213e-06,
|
| 1620 |
+
"train_recon": 3.6563369918928296e-06,
|
| 1621 |
+
"S0": 1.896807074546814,
|
| 1622 |
+
"SD": 1.2648762464523315,
|
| 1623 |
+
"ratio": 1.4995989799499512,
|
| 1624 |
+
"erank": 3.9408254623413086,
|
| 1625 |
+
"row_cv": 0.9593717861834775,
|
| 1626 |
+
"cv_in_band": true,
|
| 1627 |
+
"s_delta": 0.023766905069351196,
|
| 1628 |
+
"alpha_mean": 0.024798832833766937,
|
| 1629 |
+
"alpha_std": 0.001442883163690567,
|
| 1630 |
+
"epoch_max_grad": 0.024303449560340244,
|
| 1631 |
+
"epoch_time_s": 264.90487909317017,
|
| 1632 |
+
"tree_metrics": null,
|
| 1633 |
+
"sp_metrics": null,
|
| 1634 |
+
"bt_metrics": {
|
| 1635 |
+
"per_byte_acc": 0.8019612431526184,
|
| 1636 |
+
"per_byte_l1": 0.1988932341337204,
|
| 1637 |
+
"trigram_exact_rate": 0.5279541015625,
|
| 1638 |
+
"per_channel_acc": [
|
| 1639 |
+
0.90411376953125,
|
| 1640 |
+
0.591094970703125,
|
| 1641 |
+
0.910675048828125
|
| 1642 |
+
]
|
| 1643 |
+
}
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 28,
|
| 1647 |
+
"global_batch": 26500,
|
| 1648 |
+
"train_recon": 1.7460134583072087e-06,
|
| 1649 |
+
"test_mse": 1.100670829146111e-06,
|
| 1650 |
+
"S0": 1.8955347537994385,
|
| 1651 |
+
"SD": 1.2647769451141357,
|
| 1652 |
+
"ratio": 1.4987106984633367,
|
| 1653 |
+
"erank": 3.941080331802368,
|
| 1654 |
+
"row_cv": 0.8363554295664984,
|
| 1655 |
+
"cv_in_band": true,
|
| 1656 |
+
"s_delta": 0.02377304993569851,
|
| 1657 |
+
"alpha_mean": 0.024797767400741577,
|
| 1658 |
+
"alpha_std": 0.001448134076781571,
|
| 1659 |
+
"epoch_max_grad": 0.008920319637661314
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 28,
|
| 1663 |
+
"global_batch": 27000,
|
| 1664 |
+
"train_recon": 2.9179128488028065e-06,
|
| 1665 |
+
"test_mse": 2.0533093447738793e-06,
|
| 1666 |
+
"S0": 1.8956741094589233,
|
| 1667 |
+
"SD": 1.262389898300171,
|
| 1668 |
+
"ratio": 1.5016549934334316,
|
| 1669 |
+
"erank": 3.9407615661621094,
|
| 1670 |
+
"row_cv": 0.9796424599975287,
|
| 1671 |
+
"cv_in_band": true,
|
| 1672 |
+
"s_delta": 0.023793790489435196,
|
| 1673 |
+
"alpha_mean": 0.024798868224024773,
|
| 1674 |
+
"alpha_std": 0.0014613403473049402,
|
| 1675 |
+
"epoch_max_grad": 0.014916167176054531
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 28,
|
| 1679 |
+
"global_batch": 27328,
|
| 1680 |
+
"epoch_test_mse": 1.2711435250821523e-06,
|
| 1681 |
+
"train_recon": 2.983934587739739e-06,
|
| 1682 |
+
"S0": 1.8949766159057617,
|
| 1683 |
+
"SD": 1.2606444358825684,
|
| 1684 |
+
"ratio": 1.5031808614730835,
|
| 1685 |
+
"erank": 3.9406230449676514,
|
| 1686 |
+
"row_cv": 0.8834887883469416,
|
| 1687 |
+
"cv_in_band": true,
|
| 1688 |
+
"s_delta": 0.023813515901565552,
|
| 1689 |
+
"alpha_mean": 0.024798493832349777,
|
| 1690 |
+
"alpha_std": 0.0014714642893522978,
|
| 1691 |
+
"epoch_max_grad": 0.01513129311603079,
|
| 1692 |
+
"epoch_time_s": 264.92989206314087,
|
| 1693 |
+
"tree_metrics": null,
|
| 1694 |
+
"sp_metrics": null,
|
| 1695 |
+
"bt_metrics": {
|
| 1696 |
+
"per_byte_acc": 0.99432373046875,
|
| 1697 |
+
"per_byte_l1": 0.0062052407301962376,
|
| 1698 |
+
"trigram_exact_rate": 0.98773193359375,
|
| 1699 |
+
"per_channel_acc": [
|
| 1700 |
+
0.993560791015625,
|
| 1701 |
+
0.994781494140625,
|
| 1702 |
+
0.99462890625
|
| 1703 |
+
]
|
| 1704 |
+
}
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 29,
|
| 1708 |
+
"global_batch": 27500,
|
| 1709 |
+
"train_recon": 2.467163086045497e-06,
|
| 1710 |
+
"test_mse": 4.646417892217869e-06,
|
| 1711 |
+
"S0": 1.897601842880249,
|
| 1712 |
+
"SD": 1.2613435983657837,
|
| 1713 |
+
"ratio": 1.5044289520274428,
|
| 1714 |
+
"erank": 3.940566062927246,
|
| 1715 |
+
"row_cv": 0.9608164198454155,
|
| 1716 |
+
"cv_in_band": true,
|
| 1717 |
+
"s_delta": 0.023829646408557892,
|
| 1718 |
+
"alpha_mean": 0.024797463789582253,
|
| 1719 |
+
"alpha_std": 0.0014723323984071612,
|
| 1720 |
+
"epoch_max_grad": 0.02147062414392727
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 29,
|
| 1724 |
+
"global_batch": 28000,
|
| 1725 |
+
"train_recon": 2.814228489271608e-06,
|
| 1726 |
+
"test_mse": 9.26052325667115e-06,
|
| 1727 |
+
"S0": 1.8961085081100464,
|
| 1728 |
+
"SD": 1.2613221406936646,
|
| 1729 |
+
"ratio": 1.5032706014615544,
|
| 1730 |
+
"erank": 3.940647602081299,
|
| 1731 |
+
"row_cv": 0.9672879373082286,
|
| 1732 |
+
"cv_in_band": true,
|
| 1733 |
+
"s_delta": 0.02389451302587986,
|
| 1734 |
+
"alpha_mean": 0.024796534329652786,
|
| 1735 |
+
"alpha_std": 0.0014903489500284195,
|
| 1736 |
+
"epoch_max_grad": 0.02147062414392727
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 29,
|
| 1740 |
+
"global_batch": 28304,
|
| 1741 |
+
"epoch_test_mse": 1.0996999197232072e-06,
|
| 1742 |
+
"train_recon": 2.77898686841231e-06,
|
| 1743 |
+
"S0": 1.8953030109405518,
|
| 1744 |
+
"SD": 1.2580885887145996,
|
| 1745 |
+
"ratio": 1.5064940452575684,
|
| 1746 |
+
"erank": 3.9404067993164062,
|
| 1747 |
+
"row_cv": 0.8682890382628892,
|
| 1748 |
+
"cv_in_band": true,
|
| 1749 |
+
"s_delta": 0.0238858163356781,
|
| 1750 |
+
"alpha_mean": 0.024796565994620323,
|
| 1751 |
+
"alpha_std": 0.0014979669358581305,
|
| 1752 |
+
"epoch_max_grad": 0.02147062414392727,
|
| 1753 |
+
"epoch_time_s": 264.8245906829834,
|
| 1754 |
+
"tree_metrics": null,
|
| 1755 |
+
"sp_metrics": null,
|
| 1756 |
+
"bt_metrics": {
|
| 1757 |
+
"per_byte_acc": 0.9949137568473816,
|
| 1758 |
+
"per_byte_l1": 0.00555419921875,
|
| 1759 |
+
"trigram_exact_rate": 0.989044189453125,
|
| 1760 |
+
"per_channel_acc": [
|
| 1761 |
+
0.99462890625,
|
| 1762 |
+
0.995025634765625,
|
| 1763 |
+
0.995086669921875
|
| 1764 |
+
]
|
| 1765 |
+
}
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 30,
|
| 1769 |
+
"global_batch": 28500,
|
| 1770 |
+
"train_recon": 1.3135870758838903e-06,
|
| 1771 |
+
"test_mse": 1.2069241392964614e-06,
|
| 1772 |
+
"S0": 1.8953742980957031,
|
| 1773 |
+
"SD": 1.257088541984558,
|
| 1774 |
+
"ratio": 1.5077492314312204,
|
| 1775 |
+
"erank": 3.940396308898926,
|
| 1776 |
+
"row_cv": 1.1296194153051187,
|
| 1777 |
+
"cv_in_band": true,
|
| 1778 |
+
"s_delta": 0.02386978641152382,
|
| 1779 |
+
"alpha_mean": 0.024795234203338623,
|
| 1780 |
+
"alpha_std": 0.0014984392328187823,
|
| 1781 |
+
"epoch_max_grad": 0.004620218610720772
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 30,
|
| 1785 |
+
"global_batch": 29000,
|
| 1786 |
+
"train_recon": 2.42882746685306e-06,
|
| 1787 |
+
"test_mse": 4.753424946102314e-06,
|
| 1788 |
+
"S0": 1.896955966949463,
|
| 1789 |
+
"SD": 1.2577550411224365,
|
| 1790 |
+
"ratio": 1.5082077907431939,
|
| 1791 |
+
"erank": 3.940188407897949,
|
| 1792 |
+
"row_cv": 1.0243478373886923,
|
| 1793 |
+
"cv_in_band": true,
|
| 1794 |
+
"s_delta": 0.023943761363625526,
|
| 1795 |
+
"alpha_mean": 0.024796951562166214,
|
| 1796 |
+
"alpha_std": 0.0015085841296240687,
|
| 1797 |
+
"epoch_max_grad": 0.01313438497674956
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 30,
|
| 1801 |
+
"global_batch": 29280,
|
| 1802 |
+
"epoch_test_mse": 1.0770738485007314e-06,
|
| 1803 |
+
"train_recon": 2.53200523152149e-06,
|
| 1804 |
+
"S0": 1.8961541652679443,
|
| 1805 |
+
"SD": 1.2558772563934326,
|
| 1806 |
+
"ratio": 1.5098243951797485,
|
| 1807 |
+
"erank": 3.9400408267974854,
|
| 1808 |
+
"row_cv": 0.8587111589500542,
|
| 1809 |
+
"cv_in_band": true,
|
| 1810 |
+
"s_delta": 0.02395474910736084,
|
| 1811 |
+
"alpha_mean": 0.024798797443509102,
|
| 1812 |
+
"alpha_std": 0.0015138944145292044,
|
| 1813 |
+
"epoch_max_grad": 0.01313438497674956,
|
| 1814 |
+
"epoch_time_s": 265.00290989875793,
|
| 1815 |
+
"tree_metrics": null,
|
| 1816 |
+
"sp_metrics": null,
|
| 1817 |
+
"bt_metrics": {
|
| 1818 |
+
"per_byte_acc": 0.9951578974723816,
|
| 1819 |
+
"per_byte_l1": 0.0052591958083212376,
|
| 1820 |
+
"trigram_exact_rate": 0.989715576171875,
|
| 1821 |
+
"per_channel_acc": [
|
| 1822 |
+
0.99493408203125,
|
| 1823 |
+
0.995147705078125,
|
| 1824 |
+
0.995391845703125
|
| 1825 |
+
]
|
| 1826 |
+
}
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 31,
|
| 1830 |
+
"global_batch": 29500,
|
| 1831 |
+
"train_recon": 1.262390597648019e-06,
|
| 1832 |
+
"test_mse": 3.1016693355923053e-06,
|
| 1833 |
+
"S0": 1.8954792022705078,
|
| 1834 |
+
"SD": 1.2541604042053223,
|
| 1835 |
+
"ratio": 1.5113530779645494,
|
| 1836 |
+
"erank": 3.94000244140625,
|
| 1837 |
+
"row_cv": 0.9232080788801957,
|
| 1838 |
+
"cv_in_band": true,
|
| 1839 |
+
"s_delta": 0.023927289992570877,
|
| 1840 |
+
"alpha_mean": 0.02479725331068039,
|
| 1841 |
+
"alpha_std": 0.0015148746315389872,
|
| 1842 |
+
"epoch_max_grad": 0.0053949301379513604
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 31,
|
| 1846 |
+
"global_batch": 30000,
|
| 1847 |
+
"train_recon": 2.160274267263655e-06,
|
| 1848 |
+
"test_mse": 6.201492851687362e-06,
|
| 1849 |
+
"S0": 1.8958641290664673,
|
| 1850 |
+
"SD": 1.2529094219207764,
|
| 1851 |
+
"ratio": 1.513169332726634,
|
| 1852 |
+
"erank": 3.9396228790283203,
|
| 1853 |
+
"row_cv": 1.0179497875200039,
|
| 1854 |
+
"cv_in_band": true,
|
| 1855 |
+
"s_delta": 0.02395879663527012,
|
| 1856 |
+
"alpha_mean": 0.024797867983579636,
|
| 1857 |
+
"alpha_std": 0.0015246589900925756,
|
| 1858 |
+
"epoch_max_grad": 0.016739567118401594
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 31,
|
| 1862 |
+
"global_batch": 30256,
|
| 1863 |
+
"epoch_test_mse": 2.3198323491669724e-06,
|
| 1864 |
+
"train_recon": 2.192164837997664e-06,
|
| 1865 |
+
"S0": 1.894954800605774,
|
| 1866 |
+
"SD": 1.2529220581054688,
|
| 1867 |
+
"ratio": 1.5124282836914062,
|
| 1868 |
+
"erank": 3.939846992492676,
|
| 1869 |
+
"row_cv": 1.0013522671967876,
|
| 1870 |
+
"cv_in_band": true,
|
| 1871 |
+
"s_delta": 0.023984044790267944,
|
| 1872 |
+
"alpha_mean": 0.024797948077321053,
|
| 1873 |
+
"alpha_std": 0.0015292898751795292,
|
| 1874 |
+
"epoch_max_grad": 0.016739567118401594,
|
| 1875 |
+
"epoch_time_s": 264.97965955734253,
|
| 1876 |
+
"tree_metrics": null,
|
| 1877 |
+
"sp_metrics": null,
|
| 1878 |
+
"bt_metrics": {
|
| 1879 |
+
"per_byte_acc": 0.9901631474494934,
|
| 1880 |
+
"per_byte_l1": 0.010203043930232525,
|
| 1881 |
+
"trigram_exact_rate": 0.97552490234375,
|
| 1882 |
+
"per_channel_acc": [
|
| 1883 |
+
0.99151611328125,
|
| 1884 |
+
0.984405517578125,
|
| 1885 |
+
0.99456787109375
|
| 1886 |
+
]
|
| 1887 |
+
}
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 32,
|
| 1891 |
+
"global_batch": 30500,
|
| 1892 |
+
"train_recon": 1.2479918560719866e-06,
|
| 1893 |
+
"test_mse": 1.957933136509382e-06,
|
| 1894 |
+
"S0": 1.8969004154205322,
|
| 1895 |
+
"SD": 1.2532347440719604,
|
| 1896 |
+
"ratio": 1.5136034244639314,
|
| 1897 |
+
"erank": 3.9398574829101562,
|
| 1898 |
+
"row_cv": 0.9487942649520559,
|
| 1899 |
+
"cv_in_band": true,
|
| 1900 |
+
"s_delta": 0.023993534967303276,
|
| 1901 |
+
"alpha_mean": 0.024796046316623688,
|
| 1902 |
+
"alpha_std": 0.0015306933782994747,
|
| 1903 |
+
"epoch_max_grad": 0.0052875430007463696
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 32,
|
| 1907 |
+
"global_batch": 31000,
|
| 1908 |
+
"train_recon": 1.8300526448605503e-06,
|
| 1909 |
+
"test_mse": 1.7692004803393502e-06,
|
| 1910 |
+
"S0": 1.895942211151123,
|
| 1911 |
+
"SD": 1.2518470287322998,
|
| 1912 |
+
"ratio": 1.514515873337908,
|
| 1913 |
+
"erank": 3.9397237300872803,
|
| 1914 |
+
"row_cv": 0.974408638117066,
|
| 1915 |
+
"cv_in_band": true,
|
| 1916 |
+
"s_delta": 0.024006947875022888,
|
| 1917 |
+
"alpha_mean": 0.024798039346933365,
|
| 1918 |
+
"alpha_std": 0.0015350276371464133,
|
| 1919 |
+
"epoch_max_grad": 0.011372175135638266
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 32,
|
| 1923 |
+
"global_batch": 31232,
|
| 1924 |
+
"epoch_test_mse": 9.132497524660721e-07,
|
| 1925 |
+
"train_recon": 1.8720234590027123e-06,
|
| 1926 |
+
"S0": 1.8958433866500854,
|
| 1927 |
+
"SD": 1.2510356903076172,
|
| 1928 |
+
"ratio": 1.5154191255569458,
|
| 1929 |
+
"erank": 3.9395792484283447,
|
| 1930 |
+
"row_cv": 0.837228984964454,
|
| 1931 |
+
"cv_in_band": true,
|
| 1932 |
+
"s_delta": 0.02401164174079895,
|
| 1933 |
+
"alpha_mean": 0.02479933574795723,
|
| 1934 |
+
"alpha_std": 0.0015363013371825218,
|
| 1935 |
+
"epoch_max_grad": 0.01153441601267975,
|
| 1936 |
+
"epoch_time_s": 264.86951303482056,
|
| 1937 |
+
"tree_metrics": null,
|
| 1938 |
+
"sp_metrics": null,
|
| 1939 |
+
"bt_metrics": {
|
| 1940 |
+
"per_byte_acc": 0.9959818720817566,
|
| 1941 |
+
"per_byte_l1": 0.00433349609375,
|
| 1942 |
+
"trigram_exact_rate": 0.99139404296875,
|
| 1943 |
+
"per_channel_acc": [
|
| 1944 |
+
0.99566650390625,
|
| 1945 |
+
0.99609375,
|
| 1946 |
+
0.996185302734375
|
| 1947 |
+
]
|
| 1948 |
+
}
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 33,
|
| 1952 |
+
"global_batch": 31500,
|
| 1953 |
+
"train_recon": 1.425849136458666e-06,
|
| 1954 |
+
"test_mse": 1.026169798024057e-06,
|
| 1955 |
+
"S0": 1.8965543508529663,
|
| 1956 |
+
"SD": 1.250990629196167,
|
| 1957 |
+
"ratio": 1.5160420001796424,
|
| 1958 |
+
"erank": 3.939680337905884,
|
| 1959 |
+
"row_cv": 1.0644049508828173,
|
| 1960 |
+
"cv_in_band": true,
|
| 1961 |
+
"s_delta": 0.02401936613023281,
|
| 1962 |
+
"alpha_mean": 0.024798579514026642,
|
| 1963 |
+
"alpha_std": 0.0015375803923234344,
|
| 1964 |
+
"epoch_max_grad": 0.005775834801578209
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 33,
|
| 1968 |
+
"global_batch": 32000,
|
| 1969 |
+
"train_recon": 1.6292036978763729e-06,
|
| 1970 |
+
"test_mse": 9.765697086550063e-07,
|
| 1971 |
+
"S0": 1.896059274673462,
|
| 1972 |
+
"SD": 1.2501869201660156,
|
| 1973 |
+
"ratio": 1.5166206180236417,
|
| 1974 |
+
"erank": 3.939704656600952,
|
| 1975 |
+
"row_cv": 0.9174513308652641,
|
| 1976 |
+
"cv_in_band": true,
|
| 1977 |
+
"s_delta": 0.02403256669640541,
|
| 1978 |
+
"alpha_mean": 0.024798575788736343,
|
| 1979 |
+
"alpha_std": 0.0015416238456964493,
|
| 1980 |
+
"epoch_max_grad": 0.006660416592077838
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 33,
|
| 1984 |
+
"global_batch": 32208,
|
| 1985 |
+
"epoch_test_mse": 1.5181622933596373e-06,
|
| 1986 |
+
"train_recon": 1.7001872266774388e-06,
|
| 1987 |
+
"S0": 1.8958736658096313,
|
| 1988 |
+
"SD": 1.2491717338562012,
|
| 1989 |
+
"ratio": 1.5177046060562134,
|
| 1990 |
+
"erank": 3.939544916152954,
|
| 1991 |
+
"row_cv": 1.0257192001763413,
|
| 1992 |
+
"cv_in_band": true,
|
| 1993 |
+
"s_delta": 0.024031370878219604,
|
| 1994 |
+
"alpha_mean": 0.024799268692731857,
|
| 1995 |
+
"alpha_std": 0.0015432307263836265,
|
| 1996 |
+
"epoch_max_grad": 0.010189931284622469,
|
| 1997 |
+
"epoch_time_s": 264.9012019634247,
|
| 1998 |
+
"tree_metrics": null,
|
| 1999 |
+
"sp_metrics": null,
|
| 2000 |
+
"bt_metrics": {
|
| 2001 |
+
"per_byte_acc": 0.9943034052848816,
|
| 2002 |
+
"per_byte_l1": 0.0060323080979287624,
|
| 2003 |
+
"trigram_exact_rate": 0.98651123046875,
|
| 2004 |
+
"per_channel_acc": [
|
| 2005 |
+
0.99560546875,
|
| 2006 |
+
0.991424560546875,
|
| 2007 |
+
0.995880126953125
|
| 2008 |
+
]
|
| 2009 |
+
}
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 34,
|
| 2013 |
+
"global_batch": 32500,
|
| 2014 |
+
"train_recon": 1.0184558213763486e-06,
|
| 2015 |
+
"test_mse": 7.801633614690218e-07,
|
| 2016 |
+
"S0": 1.8959306478500366,
|
| 2017 |
+
"SD": 1.249167799949646,
|
| 2018 |
+
"ratio": 1.5177549667457904,
|
| 2019 |
+
"erank": 3.939739227294922,
|
| 2020 |
+
"row_cv": 1.1196634544300423,
|
| 2021 |
+
"cv_in_band": true,
|
| 2022 |
+
"s_delta": 0.024030685424804688,
|
| 2023 |
+
"alpha_mean": 0.02479802444577217,
|
| 2024 |
+
"alpha_std": 0.0015438341069966555,
|
| 2025 |
+
"epoch_max_grad": 0.0038894231565229087
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 34,
|
| 2029 |
+
"global_batch": 33000,
|
| 2030 |
+
"train_recon": 1.4515293787582383e-06,
|
| 2031 |
+
"test_mse": 1.4625393305323087e-06,
|
| 2032 |
+
"S0": 1.8957910537719727,
|
| 2033 |
+
"SD": 1.2482762336730957,
|
| 2034 |
+
"ratio": 1.5187271754796379,
|
| 2035 |
+
"erank": 3.939640522003174,
|
| 2036 |
+
"row_cv": 1.0480135620373134,
|
| 2037 |
+
"cv_in_band": true,
|
| 2038 |
+
"s_delta": 0.02404320053756237,
|
| 2039 |
+
"alpha_mean": 0.024799661710858345,
|
| 2040 |
+
"alpha_std": 0.0015460792928934097,
|
| 2041 |
+
"epoch_max_grad": 0.014764529465965346
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 34,
|
| 2045 |
+
"global_batch": 33184,
|
| 2046 |
+
"epoch_test_mse": 3.839639249417814e-06,
|
| 2047 |
+
"train_recon": 1.5292458323045356e-06,
|
| 2048 |
+
"S0": 1.8940974473953247,
|
| 2049 |
+
"SD": 1.2458407878875732,
|
| 2050 |
+
"ratio": 1.520336627960205,
|
| 2051 |
+
"erank": 3.9394288063049316,
|
| 2052 |
+
"row_cv": 1.0272354796595595,
|
| 2053 |
+
"cv_in_band": true,
|
| 2054 |
+
"s_delta": 0.024018704891204834,
|
| 2055 |
+
"alpha_mean": 0.024799451231956482,
|
| 2056 |
+
"alpha_std": 0.0015483071329072118,
|
| 2057 |
+
"epoch_max_grad": 0.015142576003637678,
|
| 2058 |
+
"epoch_time_s": 265.02515959739685,
|
| 2059 |
+
"tree_metrics": null,
|
| 2060 |
+
"sp_metrics": null,
|
| 2061 |
+
"bt_metrics": {
|
| 2062 |
+
"per_byte_acc": 0.9381103515625,
|
| 2063 |
+
"per_byte_l1": 0.0622049979865551,
|
| 2064 |
+
"trigram_exact_rate": 0.819244384765625,
|
| 2065 |
+
"per_channel_acc": [
|
| 2066 |
+
0.96905517578125,
|
| 2067 |
+
0.856842041015625,
|
| 2068 |
+
0.988433837890625
|
| 2069 |
+
]
|
| 2070 |
+
}
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 35,
|
| 2074 |
+
"global_batch": 33500,
|
| 2075 |
+
"train_recon": 9.253616862587237e-07,
|
| 2076 |
+
"test_mse": 9.376826142215577e-07,
|
| 2077 |
+
"S0": 1.8959720134735107,
|
| 2078 |
+
"SD": 1.2475343942642212,
|
| 2079 |
+
"ratio": 1.5197753320412268,
|
| 2080 |
+
"erank": 3.939718246459961,
|
| 2081 |
+
"row_cv": 1.0058872762609983,
|
| 2082 |
+
"cv_in_band": true,
|
| 2083 |
+
"s_delta": 0.024043714627623558,
|
| 2084 |
+
"alpha_mean": 0.024797676131129265,
|
| 2085 |
+
"alpha_std": 0.0015495780389755964,
|
| 2086 |
+
"epoch_max_grad": 0.006614061278693776
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 35,
|
| 2090 |
+
"global_batch": 34000,
|
| 2091 |
+
"train_recon": 1.182576642415089e-06,
|
| 2092 |
+
"test_mse": 1.1673351991703385e-06,
|
| 2093 |
+
"S0": 1.8960168361663818,
|
| 2094 |
+
"SD": 1.2470295429229736,
|
| 2095 |
+
"ratio": 1.520426546205112,
|
| 2096 |
+
"erank": 3.9397292137145996,
|
| 2097 |
+
"row_cv": 1.0806812653139561,
|
| 2098 |
+
"cv_in_band": true,
|
| 2099 |
+
"s_delta": 0.024047359824180603,
|
| 2100 |
+
"alpha_mean": 0.02479812689125538,
|
| 2101 |
+
"alpha_std": 0.001550282584503293,
|
| 2102 |
+
"epoch_max_grad": 0.006963930976309947
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 35,
|
| 2106 |
+
"global_batch": 34160,
|
| 2107 |
+
"epoch_test_mse": 1.0280411430358073e-06,
|
| 2108 |
+
"train_recon": 1.235591091212565e-06,
|
| 2109 |
+
"S0": 1.895575761795044,
|
| 2110 |
+
"SD": 1.2453856468200684,
|
| 2111 |
+
"ratio": 1.522079348564148,
|
| 2112 |
+
"erank": 3.9395174980163574,
|
| 2113 |
+
"row_cv": 0.9495997880988205,
|
| 2114 |
+
"cv_in_band": true,
|
| 2115 |
+
"s_delta": 0.024030059576034546,
|
| 2116 |
+
"alpha_mean": 0.024798400700092316,
|
| 2117 |
+
"alpha_std": 0.0015510342782363296,
|
| 2118 |
+
"epoch_max_grad": 0.00908947212905859,
|
| 2119 |
+
"epoch_time_s": 264.89005184173584,
|
| 2120 |
+
"tree_metrics": null,
|
| 2121 |
+
"sp_metrics": null,
|
| 2122 |
+
"bt_metrics": {
|
| 2123 |
+
"per_byte_acc": 0.99639892578125,
|
| 2124 |
+
"per_byte_l1": 0.0038248698692768812,
|
| 2125 |
+
"trigram_exact_rate": 0.99200439453125,
|
| 2126 |
+
"per_channel_acc": [
|
| 2127 |
+
0.996429443359375,
|
| 2128 |
+
0.9967041015625,
|
| 2129 |
+
0.996063232421875
|
| 2130 |
+
]
|
| 2131 |
+
}
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 36,
|
| 2135 |
+
"global_batch": 34500,
|
| 2136 |
+
"train_recon": 9.326842714204651e-07,
|
| 2137 |
+
"test_mse": 9.168727501673857e-07,
|
| 2138 |
+
"S0": 1.8964378833770752,
|
| 2139 |
+
"SD": 1.2461364269256592,
|
| 2140 |
+
"ratio": 1.521854130239361,
|
| 2141 |
+
"erank": 3.9396660327911377,
|
| 2142 |
+
"row_cv": 0.9377363438388081,
|
| 2143 |
+
"cv_in_band": true,
|
| 2144 |
+
"s_delta": 0.024038998410105705,
|
| 2145 |
+
"alpha_mean": 0.024796899408102036,
|
| 2146 |
+
"alpha_std": 0.0015519872540608048,
|
| 2147 |
+
"epoch_max_grad": 0.005767503007511208
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 36,
|
| 2151 |
+
"global_batch": 35000,
|
| 2152 |
+
"train_recon": 1.1003253866820594e-06,
|
| 2153 |
+
"test_mse": 1.3633568869408919e-06,
|
| 2154 |
+
"S0": 1.8956432342529297,
|
| 2155 |
+
"SD": 1.2451478242874146,
|
| 2156 |
+
"ratio": 1.522424231125766,
|
| 2157 |
+
"erank": 3.939668655395508,
|
| 2158 |
+
"row_cv": 0.9512282120433999,
|
| 2159 |
+
"cv_in_band": true,
|
| 2160 |
+
"s_delta": 0.024033628404140472,
|
| 2161 |
+
"alpha_mean": 0.024796675890684128,
|
| 2162 |
+
"alpha_std": 0.0015534936683252454,
|
| 2163 |
+
"epoch_max_grad": 0.007246246222578345
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 36,
|
| 2167 |
+
"global_batch": 35136,
|
| 2168 |
+
"epoch_test_mse": 1.1180632613104534e-06,
|
| 2169 |
+
"train_recon": 1.1309467088728682e-06,
|
| 2170 |
+
"S0": 1.8941624164581299,
|
| 2171 |
+
"SD": 1.2439231872558594,
|
| 2172 |
+
"ratio": 1.5227326154708862,
|
| 2173 |
+
"erank": 3.9396979808807373,
|
| 2174 |
+
"row_cv": 1.053280538768312,
|
| 2175 |
+
"cv_in_band": true,
|
| 2176 |
+
"s_delta": 0.0240136981010437,
|
| 2177 |
+
"alpha_mean": 0.02479678764939308,
|
| 2178 |
+
"alpha_std": 0.0015539340674877167,
|
| 2179 |
+
"epoch_max_grad": 0.007246246222578345,
|
| 2180 |
+
"epoch_time_s": 264.9323420524597,
|
| 2181 |
+
"tree_metrics": null,
|
| 2182 |
+
"sp_metrics": null,
|
| 2183 |
+
"bt_metrics": {
|
| 2184 |
+
"per_byte_acc": 0.996551513671875,
|
| 2185 |
+
"per_byte_l1": 0.0036722819786518812,
|
| 2186 |
+
"trigram_exact_rate": 0.992156982421875,
|
| 2187 |
+
"per_channel_acc": [
|
| 2188 |
+
0.996307373046875,
|
| 2189 |
+
0.996490478515625,
|
| 2190 |
+
0.996856689453125
|
| 2191 |
+
]
|
| 2192 |
+
}
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 37,
|
| 2196 |
+
"global_batch": 35500,
|
| 2197 |
+
"train_recon": 7.423775645502946e-07,
|
| 2198 |
+
"test_mse": 9.360755939269438e-07,
|
| 2199 |
+
"S0": 1.8958840370178223,
|
| 2200 |
+
"SD": 1.2446895837783813,
|
| 2201 |
+
"ratio": 1.5231781855447786,
|
| 2202 |
+
"erank": 3.939791679382324,
|
| 2203 |
+
"row_cv": 0.9222670645958204,
|
| 2204 |
+
"cv_in_band": true,
|
| 2205 |
+
"s_delta": 0.024024110287427902,
|
| 2206 |
+
"alpha_mean": 0.024794921278953552,
|
| 2207 |
+
"alpha_std": 0.0015543897170573473,
|
| 2208 |
+
"epoch_max_grad": 0.003361170528442165
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 37,
|
| 2212 |
+
"global_batch": 36000,
|
| 2213 |
+
"train_recon": 9.42365045322849e-07,
|
| 2214 |
+
"test_mse": 1.0382881328041549e-06,
|
| 2215 |
+
"S0": 1.8950453996658325,
|
| 2216 |
+
"SD": 1.2430622577667236,
|
| 2217 |
+
"ratio": 1.5244975644465961,
|
| 2218 |
+
"erank": 3.939742088317871,
|
| 2219 |
+
"row_cv": 1.0837528425026566,
|
| 2220 |
+
"cv_in_band": true,
|
| 2221 |
+
"s_delta": 0.024003898724913597,
|
| 2222 |
+
"alpha_mean": 0.02479434199631214,
|
| 2223 |
+
"alpha_std": 0.001555952592752874,
|
| 2224 |
+
"epoch_max_grad": 0.0064433968222469456
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 37,
|
| 2228 |
+
"global_batch": 36112,
|
| 2229 |
+
"epoch_test_mse": 2.0274479615181915e-06,
|
| 2230 |
+
"train_recon": 9.52776739544598e-07,
|
| 2231 |
+
"S0": 1.893844723701477,
|
| 2232 |
+
"SD": 1.2416751384735107,
|
| 2233 |
+
"ratio": 1.5252336263656616,
|
| 2234 |
+
"erank": 3.9396884441375732,
|
| 2235 |
+
"row_cv": 0.9533951165142684,
|
| 2236 |
+
"cv_in_band": true,
|
| 2237 |
+
"s_delta": 0.023980259895324707,
|
| 2238 |
+
"alpha_mean": 0.02479412965476513,
|
| 2239 |
+
"alpha_std": 0.0015562993939965963,
|
| 2240 |
+
"epoch_max_grad": 0.0064433968222469456,
|
| 2241 |
+
"epoch_time_s": 264.84733867645264,
|
| 2242 |
+
"tree_metrics": null,
|
| 2243 |
+
"sp_metrics": null,
|
| 2244 |
+
"bt_metrics": {
|
| 2245 |
+
"per_byte_acc": 0.9951069951057434,
|
| 2246 |
+
"per_byte_l1": 0.0051066079176962376,
|
| 2247 |
+
"trigram_exact_rate": 0.988128662109375,
|
| 2248 |
+
"per_channel_acc": [
|
| 2249 |
+
0.99639892578125,
|
| 2250 |
+
0.99591064453125,
|
| 2251 |
+
0.993011474609375
|
| 2252 |
+
]
|
| 2253 |
+
}
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 38,
|
| 2257 |
+
"global_batch": 36500,
|
| 2258 |
+
"train_recon": 7.348885816034303e-07,
|
| 2259 |
+
"test_mse": 1.2694746374108945e-06,
|
| 2260 |
+
"S0": 1.8951281309127808,
|
| 2261 |
+
"SD": 1.242309808731079,
|
| 2262 |
+
"ratio": 1.525487525204062,
|
| 2263 |
+
"erank": 3.9398112297058105,
|
| 2264 |
+
"row_cv": 1.032111891741522,
|
| 2265 |
+
"cv_in_band": true,
|
| 2266 |
+
"s_delta": 0.02398829348385334,
|
| 2267 |
+
"alpha_mean": 0.02479243278503418,
|
| 2268 |
+
"alpha_std": 0.001556779257953167,
|
| 2269 |
+
"epoch_max_grad": 0.0051739199325145785
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 38,
|
| 2273 |
+
"global_batch": 37000,
|
| 2274 |
+
"train_recon": 8.40482396665106e-07,
|
| 2275 |
+
"test_mse": 6.125526397227077e-07,
|
| 2276 |
+
"S0": 1.8957537412643433,
|
| 2277 |
+
"SD": 1.242445468902588,
|
| 2278 |
+
"ratio": 1.5258244916620418,
|
| 2279 |
+
"erank": 3.9398515224456787,
|
| 2280 |
+
"row_cv": 1.0981429420141648,
|
| 2281 |
+
"cv_in_band": true,
|
| 2282 |
+
"s_delta": 0.023994550108909607,
|
| 2283 |
+
"alpha_mean": 0.024791397154331207,
|
| 2284 |
+
"alpha_std": 0.0015577414305880666,
|
| 2285 |
+
"epoch_max_grad": 0.006125053308450744
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 38,
|
| 2289 |
+
"global_batch": 37088,
|
| 2290 |
+
"epoch_test_mse": 9.499508967564907e-07,
|
| 2291 |
+
"train_recon": 8.547780649849778e-07,
|
| 2292 |
+
"S0": 1.8945488929748535,
|
| 2293 |
+
"SD": 1.2410422563552856,
|
| 2294 |
+
"ratio": 1.5265789031982422,
|
| 2295 |
+
"erank": 3.939797878265381,
|
| 2296 |
+
"row_cv": 0.9928262357068691,
|
| 2297 |
+
"cv_in_band": true,
|
| 2298 |
+
"s_delta": 0.023972541093826294,
|
| 2299 |
+
"alpha_mean": 0.02479124814271927,
|
| 2300 |
+
"alpha_std": 0.0015581370098516345,
|
| 2301 |
+
"epoch_max_grad": 0.006125053308450744,
|
| 2302 |
+
"epoch_time_s": 264.92857694625854,
|
| 2303 |
+
"tree_metrics": null,
|
| 2304 |
+
"sp_metrics": null,
|
| 2305 |
+
"bt_metrics": {
|
| 2306 |
+
"per_byte_acc": 0.9968159794807434,
|
| 2307 |
+
"per_byte_l1": 0.0033772785682231188,
|
| 2308 |
+
"trigram_exact_rate": 0.992584228515625,
|
| 2309 |
+
"per_channel_acc": [
|
| 2310 |
+
0.996124267578125,
|
| 2311 |
+
0.997222900390625,
|
| 2312 |
+
0.997100830078125
|
| 2313 |
+
]
|
| 2314 |
+
}
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 39,
|
| 2318 |
+
"global_batch": 37500,
|
| 2319 |
+
"train_recon": 6.579768566025361e-07,
|
| 2320 |
+
"test_mse": 6.263949217100162e-07,
|
| 2321 |
+
"S0": 1.8953913450241089,
|
| 2322 |
+
"SD": 1.2423917055130005,
|
| 2323 |
+
"ratio": 1.5255988279360637,
|
| 2324 |
+
"erank": 3.940107822418213,
|
| 2325 |
+
"row_cv": 0.9477659138275691,
|
| 2326 |
+
"cv_in_band": true,
|
| 2327 |
+
"s_delta": 0.023981932550668716,
|
| 2328 |
+
"alpha_mean": 0.02478928305208683,
|
| 2329 |
+
"alpha_std": 0.0015585095388814807,
|
| 2330 |
+
"epoch_max_grad": 0.003100381771148292
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 39,
|
| 2334 |
+
"global_batch": 38000,
|
| 2335 |
+
"train_recon": 7.283967477053864e-07,
|
| 2336 |
+
"test_mse": 5.539642984331294e-07,
|
| 2337 |
+
"S0": 1.8955450057983398,
|
| 2338 |
+
"SD": 1.2425537109375,
|
| 2339 |
+
"ratio": 1.5255235840975645,
|
| 2340 |
+
"erank": 3.940225839614868,
|
| 2341 |
+
"row_cv": 1.1596566652410472,
|
| 2342 |
+
"cv_in_band": true,
|
| 2343 |
+
"s_delta": 0.023983078077435493,
|
| 2344 |
+
"alpha_mean": 0.02478773146867752,
|
| 2345 |
+
"alpha_std": 0.0015594593714922667,
|
| 2346 |
+
"epoch_max_grad": 0.007122447125222315
|
| 2347 |
+
},
|
| 2348 |
+
{
|
| 2349 |
+
"epoch": 39,
|
| 2350 |
+
"global_batch": 38064,
|
| 2351 |
+
"epoch_test_mse": 7.835114470253756e-07,
|
| 2352 |
+
"train_recon": 7.224054664776923e-07,
|
| 2353 |
+
"S0": 1.8947134017944336,
|
| 2354 |
+
"SD": 1.2413544654846191,
|
| 2355 |
+
"ratio": 1.5263274908065796,
|
| 2356 |
+
"erank": 3.940152883529663,
|
| 2357 |
+
"row_cv": 1.144498834765762,
|
| 2358 |
+
"cv_in_band": true,
|
| 2359 |
+
"s_delta": 0.023960918188095093,
|
| 2360 |
+
"alpha_mean": 0.02478746324777603,
|
| 2361 |
+
"alpha_std": 0.0015595478471368551,
|
| 2362 |
+
"epoch_max_grad": 0.007122447125222315,
|
| 2363 |
+
"epoch_time_s": 264.98999857902527,
|
| 2364 |
+
"tree_metrics": null,
|
| 2365 |
+
"sp_metrics": null,
|
| 2366 |
+
"bt_metrics": {
|
| 2367 |
+
"per_byte_acc": 0.9973042607307434,
|
| 2368 |
+
"per_byte_l1": 0.00286865234375,
|
| 2369 |
+
"trigram_exact_rate": 0.9940185546875,
|
| 2370 |
+
"per_channel_acc": [
|
| 2371 |
+
0.997222900390625,
|
| 2372 |
+
0.997100830078125,
|
| 2373 |
+
0.997589111328125
|
| 2374 |
+
]
|
| 2375 |
+
}
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 40,
|
| 2379 |
+
"global_batch": 38500,
|
| 2380 |
+
"train_recon": 6.268673515361625e-07,
|
| 2381 |
+
"test_mse": 6.52356220598449e-07,
|
| 2382 |
+
"S0": 1.895932674407959,
|
| 2383 |
+
"SD": 1.2422738075256348,
|
| 2384 |
+
"ratio": 1.5261793717783445,
|
| 2385 |
+
"erank": 3.9403154850006104,
|
| 2386 |
+
"row_cv": 0.9594313731491599,
|
| 2387 |
+
"cv_in_band": true,
|
| 2388 |
+
"s_delta": 0.02397080510854721,
|
| 2389 |
+
"alpha_mean": 0.024785753339529037,
|
| 2390 |
+
"alpha_std": 0.0015599109465256333,
|
| 2391 |
+
"epoch_max_grad": 0.0032534925688592117
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 40,
|
| 2395 |
+
"global_batch": 39000,
|
| 2396 |
+
"train_recon": 6.77257249438485e-07,
|
| 2397 |
+
"test_mse": 5.789814849777031e-07,
|
| 2398 |
+
"S0": 1.895606279373169,
|
| 2399 |
+
"SD": 1.2417895793914795,
|
| 2400 |
+
"ratio": 1.526511653477529,
|
| 2401 |
+
"erank": 3.9404170513153076,
|
| 2402 |
+
"row_cv": 0.9751296722394397,
|
| 2403 |
+
"cv_in_band": true,
|
| 2404 |
+
"s_delta": 0.023959863930940628,
|
| 2405 |
+
"alpha_mean": 0.02478431537747383,
|
| 2406 |
+
"alpha_std": 0.0015606239903718233,
|
| 2407 |
+
"epoch_max_grad": 0.005710301150263215
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 40,
|
| 2411 |
+
"global_batch": 39040,
|
| 2412 |
+
"epoch_test_mse": 6.301781322690658e-07,
|
| 2413 |
+
"train_recon": 6.776033987025175e-07,
|
| 2414 |
+
"S0": 1.8946826457977295,
|
| 2415 |
+
"SD": 1.2405413389205933,
|
| 2416 |
+
"ratio": 1.5273030996322632,
|
| 2417 |
+
"erank": 3.940324306488037,
|
| 2418 |
+
"row_cv": 1.018604917082609,
|
| 2419 |
+
"cv_in_band": true,
|
| 2420 |
+
"s_delta": 0.02393844723701477,
|
| 2421 |
+
"alpha_mean": 0.02478417381644249,
|
| 2422 |
+
"alpha_std": 0.001560706994496286,
|
| 2423 |
+
"epoch_max_grad": 0.005710301150263215,
|
| 2424 |
+
"epoch_time_s": 264.8514950275421,
|
| 2425 |
+
"tree_metrics": null,
|
| 2426 |
+
"sp_metrics": null,
|
| 2427 |
+
"bt_metrics": {
|
| 2428 |
+
"per_byte_acc": 0.9974365234375,
|
| 2429 |
+
"per_byte_l1": 0.0027262370567768812,
|
| 2430 |
+
"trigram_exact_rate": 0.994110107421875,
|
| 2431 |
+
"per_channel_acc": [
|
| 2432 |
+
0.997344970703125,
|
| 2433 |
+
0.997467041015625,
|
| 2434 |
+
0.99749755859375
|
| 2435 |
+
]
|
| 2436 |
+
}
|
| 2437 |
+
},
|
| 2438 |
+
{
|
| 2439 |
+
"epoch": 41,
|
| 2440 |
+
"global_batch": 39500,
|
| 2441 |
+
"train_recon": 5.409520196424281e-07,
|
| 2442 |
+
"test_mse": 5.769349513684574e-07,
|
| 2443 |
+
"S0": 1.8957310914993286,
|
| 2444 |
+
"SD": 1.2419471740722656,
|
| 2445 |
+
"ratio": 1.5264184466229451,
|
| 2446 |
+
"erank": 3.9405922889709473,
|
| 2447 |
+
"row_cv": 0.9765204506452732,
|
| 2448 |
+
"cv_in_band": true,
|
| 2449 |
+
"s_delta": 0.023949813097715378,
|
| 2450 |
+
"alpha_mean": 0.02478225901722908,
|
| 2451 |
+
"alpha_std": 0.0015610185218974948,
|
| 2452 |
+
"epoch_max_grad": 0.0036954612116782914
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 41,
|
| 2456 |
+
"global_batch": 40000,
|
| 2457 |
+
"train_recon": 5.63868825883181e-07,
|
| 2458 |
+
"test_mse": 5.971387508907355e-07,
|
| 2459 |
+
"S0": 1.896195888519287,
|
| 2460 |
+
"SD": 1.2425912618637085,
|
| 2461 |
+
"ratio": 1.5260012937925,
|
| 2462 |
+
"erank": 3.9407691955566406,
|
| 2463 |
+
"row_cv": 1.047061895024705,
|
| 2464 |
+
"cv_in_band": true,
|
| 2465 |
+
"s_delta": 0.023949207738041878,
|
| 2466 |
+
"alpha_mean": 0.02478024736046791,
|
| 2467 |
+
"alpha_std": 0.0015615924494341016,
|
| 2468 |
+
"epoch_max_grad": 0.0036954612116782914
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 41,
|
| 2472 |
+
"global_batch": 40016,
|
| 2473 |
+
"epoch_test_mse": 6.672651497865445e-07,
|
| 2474 |
+
"train_recon": 5.676049629873068e-07,
|
| 2475 |
+
"S0": 1.8948991298675537,
|
| 2476 |
+
"SD": 1.2411723136901855,
|
| 2477 |
+
"ratio": 1.5267010927200317,
|
| 2478 |
+
"erank": 3.9406790733337402,
|
| 2479 |
+
"row_cv": 1.047061895024705,
|
| 2480 |
+
"cv_in_band": true,
|
| 2481 |
+
"s_delta": 0.023924916982650757,
|
| 2482 |
+
"alpha_mean": 0.024780310690402985,
|
| 2483 |
+
"alpha_std": 0.0015616294695064425,
|
| 2484 |
+
"epoch_max_grad": 0.0036954612116782914,
|
| 2485 |
+
"epoch_time_s": 264.9250771999359,
|
| 2486 |
+
"tree_metrics": null,
|
| 2487 |
+
"sp_metrics": null,
|
| 2488 |
+
"bt_metrics": {
|
| 2489 |
+
"per_byte_acc": 0.9974873661994934,
|
| 2490 |
+
"per_byte_l1": 0.0026957194786518812,
|
| 2491 |
+
"trigram_exact_rate": 0.994354248046875,
|
| 2492 |
+
"per_channel_acc": [
|
| 2493 |
+
0.997344970703125,
|
| 2494 |
+
0.99749755859375,
|
| 2495 |
+
0.99761962890625
|
| 2496 |
+
]
|
| 2497 |
+
}
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 42,
|
| 2501 |
+
"global_batch": 40500,
|
| 2502 |
+
"train_recon": 5.102005259997842e-07,
|
| 2503 |
+
"test_mse": 4.643132740511646e-07,
|
| 2504 |
+
"S0": 1.8961210250854492,
|
| 2505 |
+
"SD": 1.2429113388061523,
|
| 2506 |
+
"ratio": 1.5255480826582857,
|
| 2507 |
+
"erank": 3.940950393676758,
|
| 2508 |
+
"row_cv": 1.1747868726640276,
|
| 2509 |
+
"cv_in_band": true,
|
| 2510 |
+
"s_delta": 0.02394365519285202,
|
| 2511 |
+
"alpha_mean": 0.024778395891189575,
|
| 2512 |
+
"alpha_std": 0.0015617702156305313,
|
| 2513 |
+
"epoch_max_grad": 0.0020217889364730437
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 42,
|
| 2517 |
+
"global_batch": 40992,
|
| 2518 |
+
"epoch_test_mse": 5.613590440589178e-07,
|
| 2519 |
+
"train_recon": 5.342791493303669e-07,
|
| 2520 |
+
"S0": 1.8953204154968262,
|
| 2521 |
+
"SD": 1.2422767877578735,
|
| 2522 |
+
"ratio": 1.5256829261779785,
|
| 2523 |
+
"erank": 3.941063404083252,
|
| 2524 |
+
"row_cv": 1.0288416797728244,
|
| 2525 |
+
"cv_in_band": true,
|
| 2526 |
+
"s_delta": 0.023922383785247803,
|
| 2527 |
+
"alpha_mean": 0.02477671578526497,
|
| 2528 |
+
"alpha_std": 0.0015620269114151597,
|
| 2529 |
+
"epoch_max_grad": 0.003411140722819328,
|
| 2530 |
+
"epoch_time_s": 264.57036876678467,
|
| 2531 |
+
"tree_metrics": null,
|
| 2532 |
+
"sp_metrics": null,
|
| 2533 |
+
"bt_metrics": {
|
| 2534 |
+
"per_byte_acc": 0.9977824091911316,
|
| 2535 |
+
"per_byte_l1": 0.0023701984900981188,
|
| 2536 |
+
"trigram_exact_rate": 0.994903564453125,
|
| 2537 |
+
"per_channel_acc": [
|
| 2538 |
+
0.99761962890625,
|
| 2539 |
+
0.997833251953125,
|
| 2540 |
+
0.997894287109375
|
| 2541 |
+
]
|
| 2542 |
+
}
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 43,
|
| 2546 |
+
"global_batch": 41000,
|
| 2547 |
+
"train_recon": 4.637408430596679e-07,
|
| 2548 |
+
"test_mse": 4.378129574433842e-07,
|
| 2549 |
+
"S0": 1.896363615989685,
|
| 2550 |
+
"SD": 1.2430862188339233,
|
| 2551 |
+
"ratio": 1.5255286174061862,
|
| 2552 |
+
"erank": 3.941073179244995,
|
| 2553 |
+
"row_cv": 1.0130194701006545,
|
| 2554 |
+
"cv_in_band": true,
|
| 2555 |
+
"s_delta": 0.02393827959895134,
|
| 2556 |
+
"alpha_mean": 0.02477671578526497,
|
| 2557 |
+
"alpha_std": 0.0015620582271367311,
|
| 2558 |
+
"epoch_max_grad": 0.0015099841816341379
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 43,
|
| 2562 |
+
"global_batch": 41500,
|
| 2563 |
+
"train_recon": 4.657336019834217e-07,
|
| 2564 |
+
"test_mse": 4.3164499174963566e-07,
|
| 2565 |
+
"S0": 1.896611213684082,
|
| 2566 |
+
"SD": 1.2438461780548096,
|
| 2567 |
+
"ratio": 1.5247956153245121,
|
| 2568 |
+
"erank": 3.9413132667541504,
|
| 2569 |
+
"row_cv": 0.9293501147182849,
|
| 2570 |
+
"cv_in_band": true,
|
| 2571 |
+
"s_delta": 0.02394014783203602,
|
| 2572 |
+
"alpha_mean": 0.024774864315986633,
|
| 2573 |
+
"alpha_std": 0.0015621468191966414,
|
| 2574 |
+
"epoch_max_grad": 0.0016650023794090698
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 43,
|
| 2578 |
+
"global_batch": 41968,
|
| 2579 |
+
"epoch_test_mse": 4.3538424483813287e-07,
|
| 2580 |
+
"train_recon": 4.764573017945022e-07,
|
| 2581 |
+
"S0": 1.895654559135437,
|
| 2582 |
+
"SD": 1.2431647777557373,
|
| 2583 |
+
"ratio": 1.5248618125915527,
|
| 2584 |
+
"erank": 3.941436290740967,
|
| 2585 |
+
"row_cv": 1.1462304734020725,
|
| 2586 |
+
"cv_in_band": true,
|
| 2587 |
+
"s_delta": 0.02391532063484192,
|
| 2588 |
+
"alpha_mean": 0.024773165583610535,
|
| 2589 |
+
"alpha_std": 0.0015623716171830893,
|
| 2590 |
+
"epoch_max_grad": 0.0022224960980140773,
|
| 2591 |
+
"epoch_time_s": 264.90074944496155,
|
| 2592 |
+
"tree_metrics": null,
|
| 2593 |
+
"sp_metrics": null,
|
| 2594 |
+
"bt_metrics": {
|
| 2595 |
+
"per_byte_acc": 0.997955322265625,
|
| 2596 |
+
"per_byte_l1": 0.0021769206505268812,
|
| 2597 |
+
"trigram_exact_rate": 0.995147705078125,
|
| 2598 |
+
"per_channel_acc": [
|
| 2599 |
+
0.99786376953125,
|
| 2600 |
+
0.997955322265625,
|
| 2601 |
+
0.998046875
|
| 2602 |
+
]
|
| 2603 |
+
}
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 44,
|
| 2607 |
+
"global_batch": 42000,
|
| 2608 |
+
"train_recon": 4.313557200319451e-07,
|
| 2609 |
+
"test_mse": 4.217298510411638e-07,
|
| 2610 |
+
"S0": 1.8966248035430908,
|
| 2611 |
+
"SD": 1.2440450191497803,
|
| 2612 |
+
"ratio": 1.5245628245782263,
|
| 2613 |
+
"erank": 3.941469192504883,
|
| 2614 |
+
"row_cv": 0.913453495824278,
|
| 2615 |
+
"cv_in_band": true,
|
| 2616 |
+
"s_delta": 0.02393151819705963,
|
| 2617 |
+
"alpha_mean": 0.024773089215159416,
|
| 2618 |
+
"alpha_std": 0.0015623903600499034,
|
| 2619 |
+
"epoch_max_grad": 0.001022965781318407
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 44,
|
| 2623 |
+
"global_batch": 42500,
|
| 2624 |
+
"train_recon": 4.75286999224501e-07,
|
| 2625 |
+
"test_mse": 5.007020149605523e-07,
|
| 2626 |
+
"S0": 1.8966865539550781,
|
| 2627 |
+
"SD": 1.2444164752960205,
|
| 2628 |
+
"ratio": 1.5241573672209079,
|
| 2629 |
+
"erank": 3.941603660583496,
|
| 2630 |
+
"row_cv": 0.9836276492174126,
|
| 2631 |
+
"cv_in_band": true,
|
| 2632 |
+
"s_delta": 0.02392706274986267,
|
| 2633 |
+
"alpha_mean": 0.02477162517607212,
|
| 2634 |
+
"alpha_std": 0.0015625733649358153,
|
| 2635 |
+
"epoch_max_grad": 0.0020650071108003374
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 44,
|
| 2639 |
+
"global_batch": 42944,
|
| 2640 |
+
"epoch_test_mse": 4.6381704419218294e-07,
|
| 2641 |
+
"train_recon": 4.6990319986301514e-07,
|
| 2642 |
+
"S0": 1.8963221311569214,
|
| 2643 |
+
"SD": 1.2441587448120117,
|
| 2644 |
+
"ratio": 1.5241801738739014,
|
| 2645 |
+
"erank": 3.941697835922241,
|
| 2646 |
+
"row_cv": 0.9494167808386084,
|
| 2647 |
+
"cv_in_band": true,
|
| 2648 |
+
"s_delta": 0.023911625146865845,
|
| 2649 |
+
"alpha_mean": 0.02477026730775833,
|
| 2650 |
+
"alpha_std": 0.0015627137618139386,
|
| 2651 |
+
"epoch_max_grad": 0.0025432571245806355,
|
| 2652 |
+
"epoch_time_s": 264.79927277565,
|
| 2653 |
+
"tree_metrics": null,
|
| 2654 |
+
"sp_metrics": null,
|
| 2655 |
+
"bt_metrics": {
|
| 2656 |
+
"per_byte_acc": 0.9980265498161316,
|
| 2657 |
+
"per_byte_l1": 0.0021260578650981188,
|
| 2658 |
+
"trigram_exact_rate": 0.995361328125,
|
| 2659 |
+
"per_channel_acc": [
|
| 2660 |
+
0.997955322265625,
|
| 2661 |
+
0.998016357421875,
|
| 2662 |
+
0.99810791015625
|
| 2663 |
+
]
|
| 2664 |
+
}
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 45,
|
| 2668 |
+
"global_batch": 43000,
|
| 2669 |
+
"train_recon": 4.112804415764718e-07,
|
| 2670 |
+
"test_mse": 3.870932232530322e-07,
|
| 2671 |
+
"S0": 1.8970893621444702,
|
| 2672 |
+
"SD": 1.2448463439941406,
|
| 2673 |
+
"ratio": 1.523954627860363,
|
| 2674 |
+
"erank": 3.9417290687561035,
|
| 2675 |
+
"row_cv": 1.0702307823642123,
|
| 2676 |
+
"cv_in_band": true,
|
| 2677 |
+
"s_delta": 0.0239240899682045,
|
| 2678 |
+
"alpha_mean": 0.024770177900791168,
|
| 2679 |
+
"alpha_std": 0.0015627399552613497,
|
| 2680 |
+
"epoch_max_grad": 0.0011740093680829083
|
| 2681 |
+
},
|
| 2682 |
+
{
|
| 2683 |
+
"epoch": 45,
|
| 2684 |
+
"global_batch": 43500,
|
| 2685 |
+
"train_recon": 4.0940149352804803e-07,
|
| 2686 |
+
"test_mse": 4.13154168654728e-07,
|
| 2687 |
+
"S0": 1.8973259925842285,
|
| 2688 |
+
"SD": 1.2455343008041382,
|
| 2689 |
+
"ratio": 1.523302871808712,
|
| 2690 |
+
"erank": 3.941901683807373,
|
| 2691 |
+
"row_cv": 1.0074267302302278,
|
| 2692 |
+
"cv_in_band": true,
|
| 2693 |
+
"s_delta": 0.023928314447402954,
|
| 2694 |
+
"alpha_mean": 0.02476874738931656,
|
| 2695 |
+
"alpha_std": 0.0015627288958057761,
|
| 2696 |
+
"epoch_max_grad": 0.0011740093680829083
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 45,
|
| 2700 |
+
"global_batch": 43920,
|
| 2701 |
+
"epoch_test_mse": 3.9127582799665104e-07,
|
| 2702 |
+
"train_recon": 4.0948349201048436e-07,
|
| 2703 |
+
"S0": 1.8965778350830078,
|
| 2704 |
+
"SD": 1.2449404001235962,
|
| 2705 |
+
"ratio": 1.5234286785125732,
|
| 2706 |
+
"erank": 3.9419775009155273,
|
| 2707 |
+
"row_cv": 0.9721624292649762,
|
| 2708 |
+
"cv_in_band": true,
|
| 2709 |
+
"s_delta": 0.02390819787979126,
|
| 2710 |
+
"alpha_mean": 0.024767432361841202,
|
| 2711 |
+
"alpha_std": 0.001562700723297894,
|
| 2712 |
+
"epoch_max_grad": 0.0011798721069649846,
|
| 2713 |
+
"epoch_time_s": 264.8442163467407,
|
| 2714 |
+
"tree_metrics": null,
|
| 2715 |
+
"sp_metrics": null,
|
| 2716 |
+
"bt_metrics": {
|
| 2717 |
+
"per_byte_acc": 0.997955322265625,
|
| 2718 |
+
"per_byte_l1": 0.0021870930213481188,
|
| 2719 |
+
"trigram_exact_rate": 0.995269775390625,
|
| 2720 |
+
"per_channel_acc": [
|
| 2721 |
+
0.997894287109375,
|
| 2722 |
+
0.997955322265625,
|
| 2723 |
+
0.998016357421875
|
| 2724 |
+
]
|
| 2725 |
+
}
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 46,
|
| 2729 |
+
"global_batch": 44000,
|
| 2730 |
+
"train_recon": 3.8844059702114464e-07,
|
| 2731 |
+
"test_mse": 3.81506850999358e-07,
|
| 2732 |
+
"S0": 1.8974614143371582,
|
| 2733 |
+
"SD": 1.2459874153137207,
|
| 2734 |
+
"ratio": 1.5228575953400223,
|
| 2735 |
+
"erank": 3.9420552253723145,
|
| 2736 |
+
"row_cv": 1.041730799083086,
|
| 2737 |
+
"cv_in_band": true,
|
| 2738 |
+
"s_delta": 0.023923834785819054,
|
| 2739 |
+
"alpha_mean": 0.024767247959971428,
|
| 2740 |
+
"alpha_std": 0.0015626851236447692,
|
| 2741 |
+
"epoch_max_grad": 0.0005254425301370308
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 46,
|
| 2745 |
+
"global_batch": 44500,
|
| 2746 |
+
"train_recon": 4.0011650947564765e-07,
|
| 2747 |
+
"test_mse": 4.1803090766734385e-07,
|
| 2748 |
+
"S0": 1.8978290557861328,
|
| 2749 |
+
"SD": 1.2465229034423828,
|
| 2750 |
+
"ratio": 1.5224983314146314,
|
| 2751 |
+
"erank": 3.94217586517334,
|
| 2752 |
+
"row_cv": 0.9923070813837668,
|
| 2753 |
+
"cv_in_band": true,
|
| 2754 |
+
"s_delta": 0.02392435260117054,
|
| 2755 |
+
"alpha_mean": 0.024765990674495697,
|
| 2756 |
+
"alpha_std": 0.0015626095701009035,
|
| 2757 |
+
"epoch_max_grad": 0.001158068613891701
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 46,
|
| 2761 |
+
"global_batch": 44896,
|
| 2762 |
+
"epoch_test_mse": 3.8533617912435147e-07,
|
| 2763 |
+
"train_recon": 4.009965631968457e-07,
|
| 2764 |
+
"S0": 1.8968398571014404,
|
| 2765 |
+
"SD": 1.2459301948547363,
|
| 2766 |
+
"ratio": 1.5224286317825317,
|
| 2767 |
+
"erank": 3.9422521591186523,
|
| 2768 |
+
"row_cv": 1.0326723473969448,
|
| 2769 |
+
"cv_in_band": true,
|
| 2770 |
+
"s_delta": 0.023905783891677856,
|
| 2771 |
+
"alpha_mean": 0.02476504258811474,
|
| 2772 |
+
"alpha_std": 0.0015625194646418095,
|
| 2773 |
+
"epoch_max_grad": 0.001158068613891701,
|
| 2774 |
+
"epoch_time_s": 264.8767294883728,
|
| 2775 |
+
"tree_metrics": null,
|
| 2776 |
+
"sp_metrics": null,
|
| 2777 |
+
"bt_metrics": {
|
| 2778 |
+
"per_byte_acc": 0.9980265498161316,
|
| 2779 |
+
"per_byte_l1": 0.0020955402869731188,
|
| 2780 |
+
"trigram_exact_rate": 0.995330810546875,
|
| 2781 |
+
"per_channel_acc": [
|
| 2782 |
+
0.998077392578125,
|
| 2783 |
+
0.99798583984375,
|
| 2784 |
+
0.998016357421875
|
| 2785 |
+
]
|
| 2786 |
+
}
|
| 2787 |
+
},
|
| 2788 |
+
{
|
| 2789 |
+
"epoch": 47,
|
| 2790 |
+
"global_batch": 45000,
|
| 2791 |
+
"train_recon": 3.805560174677899e-07,
|
| 2792 |
+
"test_mse": 3.6502450484476867e-07,
|
| 2793 |
+
"S0": 1.8978450298309326,
|
| 2794 |
+
"SD": 1.2470426559448242,
|
| 2795 |
+
"ratio": 1.5218765818192972,
|
| 2796 |
+
"erank": 3.9423370361328125,
|
| 2797 |
+
"row_cv": 1.019960082717305,
|
| 2798 |
+
"cv_in_band": true,
|
| 2799 |
+
"s_delta": 0.023922037333250046,
|
| 2800 |
+
"alpha_mean": 0.02476486563682556,
|
| 2801 |
+
"alpha_std": 0.001562486868351698,
|
| 2802 |
+
"epoch_max_grad": 0.00038052692342238544
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 47,
|
| 2806 |
+
"global_batch": 45500,
|
| 2807 |
+
"train_recon": 3.789441535484954e-07,
|
| 2808 |
+
"test_mse": 3.658151115359942e-07,
|
| 2809 |
+
"S0": 1.8980814218521118,
|
| 2810 |
+
"SD": 1.24757981300354,
|
| 2811 |
+
"ratio": 1.5214108042261325,
|
| 2812 |
+
"erank": 3.942470073699951,
|
| 2813 |
+
"row_cv": 0.9316666809082372,
|
| 2814 |
+
"cv_in_band": true,
|
| 2815 |
+
"s_delta": 0.023923061788082123,
|
| 2816 |
+
"alpha_mean": 0.024763910099864006,
|
| 2817 |
+
"alpha_std": 0.0015622881473973393,
|
| 2818 |
+
"epoch_max_grad": 0.0004929647633273873
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 47,
|
| 2822 |
+
"global_batch": 45872,
|
| 2823 |
+
"epoch_test_mse": 3.8963481420069e-07,
|
| 2824 |
+
"train_recon": 3.791474125405404e-07,
|
| 2825 |
+
"S0": 1.8973217010498047,
|
| 2826 |
+
"SD": 1.2469170093536377,
|
| 2827 |
+
"ratio": 1.5216102600097656,
|
| 2828 |
+
"erank": 3.942500591278076,
|
| 2829 |
+
"row_cv": 0.9719612083214753,
|
| 2830 |
+
"cv_in_band": true,
|
| 2831 |
+
"s_delta": 0.02390497922897339,
|
| 2832 |
+
"alpha_mean": 0.02476312220096588,
|
| 2833 |
+
"alpha_std": 0.0015621393686160445,
|
| 2834 |
+
"epoch_max_grad": 0.0006675699265927935,
|
| 2835 |
+
"epoch_time_s": 264.8346450328827,
|
| 2836 |
+
"tree_metrics": null,
|
| 2837 |
+
"sp_metrics": null,
|
| 2838 |
+
"bt_metrics": {
|
| 2839 |
+
"per_byte_acc": 0.998138427734375,
|
| 2840 |
+
"per_byte_l1": 0.0020039875525981188,
|
| 2841 |
+
"trigram_exact_rate": 0.99554443359375,
|
| 2842 |
+
"per_channel_acc": [
|
| 2843 |
+
0.998077392578125,
|
| 2844 |
+
0.998199462890625,
|
| 2845 |
+
0.998138427734375
|
| 2846 |
+
]
|
| 2847 |
+
}
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 48,
|
| 2851 |
+
"global_batch": 46000,
|
| 2852 |
+
"train_recon": 3.6859197871663696e-07,
|
| 2853 |
+
"test_mse": 3.5664351116793114e-07,
|
| 2854 |
+
"S0": 1.8983609676361084,
|
| 2855 |
+
"SD": 1.2479228973388672,
|
| 2856 |
+
"ratio": 1.5212165402783315,
|
| 2857 |
+
"erank": 3.9425625801086426,
|
| 2858 |
+
"row_cv": 0.9520906418936101,
|
| 2859 |
+
"cv_in_band": true,
|
| 2860 |
+
"s_delta": 0.023922448977828026,
|
| 2861 |
+
"alpha_mean": 0.024762988090515137,
|
| 2862 |
+
"alpha_std": 0.0015620560152456164,
|
| 2863 |
+
"epoch_max_grad": 0.0005894988654341063
|
| 2864 |
+
},
|
| 2865 |
+
{
|
| 2866 |
+
"epoch": 48,
|
| 2867 |
+
"global_batch": 46500,
|
| 2868 |
+
"train_recon": 3.6940807773232053e-07,
|
| 2869 |
+
"test_mse": 3.568206921045203e-07,
|
| 2870 |
+
"S0": 1.8983187675476074,
|
| 2871 |
+
"SD": 1.2481660842895508,
|
| 2872 |
+
"ratio": 1.5208863437586966,
|
| 2873 |
+
"erank": 3.942652702331543,
|
| 2874 |
+
"row_cv": 1.035353292248424,
|
| 2875 |
+
"cv_in_band": true,
|
| 2876 |
+
"s_delta": 0.023920953273773193,
|
| 2877 |
+
"alpha_mean": 0.02476237528026104,
|
| 2878 |
+
"alpha_std": 0.001561684301123023,
|
| 2879 |
+
"epoch_max_grad": 0.0005894988654341063
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 48,
|
| 2883 |
+
"global_batch": 46848,
|
| 2884 |
+
"epoch_test_mse": 3.605949851589685e-07,
|
| 2885 |
+
"train_recon": 3.688228652614927e-07,
|
| 2886 |
+
"S0": 1.8977155685424805,
|
| 2887 |
+
"SD": 1.2475641965866089,
|
| 2888 |
+
"ratio": 1.5211366415023804,
|
| 2889 |
+
"erank": 3.9426560401916504,
|
| 2890 |
+
"row_cv": 0.8861206421167027,
|
| 2891 |
+
"cv_in_band": true,
|
| 2892 |
+
"s_delta": 0.02390536665916443,
|
| 2893 |
+
"alpha_mean": 0.02476186864078045,
|
| 2894 |
+
"alpha_std": 0.001561404555104673,
|
| 2895 |
+
"epoch_max_grad": 0.0005894988654341063,
|
| 2896 |
+
"epoch_time_s": 265.08042001724243,
|
| 2897 |
+
"tree_metrics": null,
|
| 2898 |
+
"sp_metrics": null,
|
| 2899 |
+
"bt_metrics": {
|
| 2900 |
+
"per_byte_acc": 0.9981486201286316,
|
| 2901 |
+
"per_byte_l1": 0.001983642578125,
|
| 2902 |
+
"trigram_exact_rate": 0.99560546875,
|
| 2903 |
+
"per_channel_acc": [
|
| 2904 |
+
0.998138427734375,
|
| 2905 |
+
0.9981689453125,
|
| 2906 |
+
0.998138427734375
|
| 2907 |
+
]
|
| 2908 |
+
}
|
| 2909 |
+
},
|
| 2910 |
+
{
|
| 2911 |
+
"epoch": 49,
|
| 2912 |
+
"global_batch": 47000,
|
| 2913 |
+
"train_recon": 3.616739441270056e-07,
|
| 2914 |
+
"test_mse": 3.4919793279186706e-07,
|
| 2915 |
+
"S0": 1.8985830545425415,
|
| 2916 |
+
"SD": 1.248485803604126,
|
| 2917 |
+
"ratio": 1.5207085525959774,
|
| 2918 |
+
"erank": 3.942721366882324,
|
| 2919 |
+
"row_cv": 1.060079404488267,
|
| 2920 |
+
"cv_in_band": true,
|
| 2921 |
+
"s_delta": 0.0239193607121706,
|
| 2922 |
+
"alpha_mean": 0.024761782959103584,
|
| 2923 |
+
"alpha_std": 0.0015613470459356904,
|
| 2924 |
+
"epoch_max_grad": 0.0002041612852542687
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 49,
|
| 2928 |
+
"global_batch": 47500,
|
| 2929 |
+
"train_recon": 3.636457585224766e-07,
|
| 2930 |
+
"test_mse": 3.583414525110129e-07,
|
| 2931 |
+
"S0": 1.8987205028533936,
|
| 2932 |
+
"SD": 1.2488330602645874,
|
| 2933 |
+
"ratio": 1.520395758298678,
|
| 2934 |
+
"erank": 3.9427995681762695,
|
| 2935 |
+
"row_cv": 1.0466524041234704,
|
| 2936 |
+
"cv_in_band": true,
|
| 2937 |
+
"s_delta": 0.0239215400069952,
|
| 2938 |
+
"alpha_mean": 0.024761423468589783,
|
| 2939 |
+
"alpha_std": 0.0015611095586791635,
|
| 2940 |
+
"epoch_max_grad": 0.00028654783812587416
|
| 2941 |
+
},
|
| 2942 |
+
{
|
| 2943 |
+
"epoch": 49,
|
| 2944 |
+
"global_batch": 47824,
|
| 2945 |
+
"epoch_test_mse": 3.5706435755855637e-07,
|
| 2946 |
+
"train_recon": 3.629518004614142e-07,
|
| 2947 |
+
"S0": 1.8978402614593506,
|
| 2948 |
+
"SD": 1.248028039932251,
|
| 2949 |
+
"ratio": 1.5206711292266846,
|
| 2950 |
+
"erank": 3.9427871704101562,
|
| 2951 |
+
"row_cv": 0.9463377217856472,
|
| 2952 |
+
"cv_in_band": true,
|
| 2953 |
+
"s_delta": 0.02390417456626892,
|
| 2954 |
+
"alpha_mean": 0.024761129170656204,
|
| 2955 |
+
"alpha_std": 0.0015609136316925287,
|
| 2956 |
+
"epoch_max_grad": 0.00028654783812587416,
|
| 2957 |
+
"epoch_time_s": 264.817599773407,
|
| 2958 |
+
"tree_metrics": null,
|
| 2959 |
+
"sp_metrics": null,
|
| 2960 |
+
"bt_metrics": {
|
| 2961 |
+
"per_byte_acc": 0.9981689453125,
|
| 2962 |
+
"per_byte_l1": 0.0019632976036518812,
|
| 2963 |
+
"trigram_exact_rate": 0.99566650390625,
|
| 2964 |
+
"per_channel_acc": [
|
| 2965 |
+
0.99822998046875,
|
| 2966 |
+
0.998138427734375,
|
| 2967 |
+
0.998138427734375
|
| 2968 |
+
]
|
| 2969 |
+
}
|
| 2970 |
+
},
|
| 2971 |
+
{
|
| 2972 |
+
"epoch": 50,
|
| 2973 |
+
"global_batch": 48000,
|
| 2974 |
+
"train_recon": 3.5861390527843057e-07,
|
| 2975 |
+
"test_mse": 3.4681306715356186e-07,
|
| 2976 |
+
"S0": 1.8987514972686768,
|
| 2977 |
+
"SD": 1.2489345073699951,
|
| 2978 |
+
"ratio": 1.5202970779181166,
|
| 2979 |
+
"erank": 3.9428343772888184,
|
| 2980 |
+
"row_cv": 0.9276686770646706,
|
| 2981 |
+
"cv_in_band": true,
|
| 2982 |
+
"s_delta": 0.02392035536468029,
|
| 2983 |
+
"alpha_mean": 0.024761123582720757,
|
| 2984 |
+
"alpha_std": 0.001560910721309483,
|
| 2985 |
+
"epoch_max_grad": 0.00014251736660701449
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 50,
|
| 2989 |
+
"global_batch": 48500,
|
| 2990 |
+
"train_recon": 3.606511140539016e-07,
|
| 2991 |
+
"test_mse": 3.525062197695661e-07,
|
| 2992 |
+
"S0": 1.8988665342330933,
|
| 2993 |
+
"SD": 1.2492074966430664,
|
| 2994 |
+
"ratio": 1.5200569354052502,
|
| 2995 |
+
"erank": 3.942873477935791,
|
| 2996 |
+
"row_cv": 1.1088441703976846,
|
| 2997 |
+
"cv_in_band": true,
|
| 2998 |
+
"s_delta": 0.023924175649881363,
|
| 2999 |
+
"alpha_mean": 0.02476108819246292,
|
| 3000 |
+
"alpha_std": 0.0015608883695676923,
|
| 3001 |
+
"epoch_max_grad": 0.00014251736660701449
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 50,
|
| 3005 |
+
"global_batch": 48800,
|
| 3006 |
+
"epoch_test_mse": 3.5937614575232145e-07,
|
| 3007 |
+
"train_recon": 3.6029499806220354e-07,
|
| 3008 |
+
"S0": 1.897929310798645,
|
| 3009 |
+
"SD": 1.2482919692993164,
|
| 3010 |
+
"ratio": 1.520421028137207,
|
| 3011 |
+
"erank": 3.9428369998931885,
|
| 3012 |
+
"row_cv": 0.9681119813360423,
|
| 3013 |
+
"cv_in_band": true,
|
| 3014 |
+
"s_delta": 0.02390742301940918,
|
| 3015 |
+
"alpha_mean": 0.02476106397807598,
|
| 3016 |
+
"alpha_std": 0.001560872420668602,
|
| 3017 |
+
"epoch_max_grad": 0.00014251736660701449,
|
| 3018 |
+
"epoch_time_s": 264.8415734767914,
|
| 3019 |
+
"tree_metrics": null,
|
| 3020 |
+
"sp_metrics": null,
|
| 3021 |
+
"bt_metrics": {
|
| 3022 |
+
"per_byte_acc": 0.9981689453125,
|
| 3023 |
+
"per_byte_l1": 0.0019632976036518812,
|
| 3024 |
+
"trigram_exact_rate": 0.995574951171875,
|
| 3025 |
+
"per_channel_acc": [
|
| 3026 |
+
0.998077392578125,
|
| 3027 |
+
0.9981689453125,
|
| 3028 |
+
0.998260498046875
|
| 3029 |
+
]
|
| 3030 |
+
}
|
| 3031 |
+
}
|
| 3032 |
+
]
|
| 3033 |
+
}
|