|
|
@@ -36,7 +36,11 @@ import pdb
|
|
|
import random
|
|
|
import math
|
|
|
class MFCCA(AbsESPnetModel):
|
|
|
- """CTC-attention hybrid Encoder-Decoder model"""
|
|
|
+ """
|
|
|
+ Author: Audio, Speech and Language Processing Group (ASLP@NPU), Northwestern Polytechnical University
|
|
|
+ MFCCA:Multi-Frame Cross-Channel attention for multi-speaker ASR in Multi-party meeting scenario
|
|
|
+ https://arxiv.org/abs/2210.05265
|
|
|
+ """
|
|
|
|
|
|
def __init__(
|
|
|
self,
|