diff --git a/egs/rm/s5/RESULTS b/egs/rm/s5/RESULTS
index ecafb588cfea373c683d23dafa2608d9ff7c5ace..a8156e10e14e3f1da7fd58dd1aebd42dddd2927b 100644 (file)
--- a/egs/rm/s5/RESULTS
+++ b/egs/rm/s5/RESULTS
@@ -234,6 +234,9 @@ for x in exp/nnet2_online_wsj/nnet_ms_a_smbr_0.00005/1/decode_*; do grep WER $x/
%WER 2.86 [ 358 / 12533, 46 ins, 61 del, 251 sub ] exp/chain/tdnn_5g/decode/wer_5_0.0
%WER 2.71 [ 340 / 12533, 58 ins, 59 del, 223 sub ] exp/chain/tdnn_5n/decode/wer_4_0.0
+### WSJ->RM Transfer learning using chain model ###
+%WER 1.68 [ 210 / 12533, 25 ins, 33 del, 152 sub ] exp/chain/tdnn_wsj_rm_1a/decode/wer_2_0.0
+
### nnet1 results ###
# dnn4b, MFCC,LDA,fMLLR feaures, (Karel - 30.7.2015)