issues
search
hirofumi0810
/
neural_sp
End-to-end ASR/LM implementation with PyTorch
Apache License 2.0
589
stars
140
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
loss ctc fluctuates
#349
qkrguswn2401
opened
2 years ago
1
librispeech dict issue vocab
#348
qkrguswn2401
closed
2 years ago
3
data preprocessing
#347
artificertxj1
opened
2 years ago
0
Recommended versions of Python, Pytorch, CUDA?
#346
SuryodayBasak
opened
2 years ago
0
StableEmit
#345
hirofumi0810
closed
2 years ago
1
Streamable architecture
#344
Durgesh92
opened
3 years ago
0
How to decode empty target files which only contain background noise
#343
CSLujunyu
opened
3 years ago
1
Refactor model loading
#342
hirofumi0810
closed
3 years ago
0
Add AISHELL-2 recipe
#341
hirofumi0810
closed
3 years ago
1
Fix bug in loading already averaged model
#340
hirofumi0810
closed
3 years ago
0
Internal LM estimation
#339
hirofumi0810
closed
3 years ago
1
Conv subsampling
#338
hirofumi0810
closed
3 years ago
1
Minor update for results in README.md
#337
hirofumi0810
closed
3 years ago
1
Refactor LM loading during evaluation
#336
hirofumi0810
closed
3 years ago
1
Fix dtype in relative MHA
#335
hirofumi0810
closed
3 years ago
1
Add LaboroTVSpeech recipe
#334
hirofumi0810
closed
3 years ago
1
Fix ensemble in LAS
#333
hirofumi0810
closed
3 years ago
1
Remove GRU support
#332
hirofumi0810
closed
3 years ago
1
Refactor frontend CNN
#331
hirofumi0810
closed
3 years ago
0
Fix yaml config
#330
hirofumi0810
closed
3 years ago
1
Fix bug in resuming multi-GPU training
#329
hirofumi0810
closed
3 years ago
1
Add zero_infinity option for CTC in torch>=1.7
#328
hirofumi0810
closed
3 years ago
0
Add set_to_none option for torch>=1.7
#327
hirofumi0810
closed
3 years ago
1
Add CI supports for various pytorch versions
#326
hirofumi0810
closed
3 years ago
1
Update CSJ recipe
#325
hirofumi0810
closed
3 years ago
0
Fix CI error
#324
hirofumi0810
closed
3 years ago
1
Update CSJ result
#323
hirofumi0810
closed
3 years ago
0
quetion about the loss and grad of "mbr"
#322
Cescfangs
opened
3 years ago
3
Update Librispeech recipe
#321
hirofumi0810
closed
3 years ago
1
Fix DDP launch
#320
hirofumi0810
closed
3 years ago
1
Update TEDLIUM2 recipe
#319
hirofumi0810
closed
3 years ago
1
Update AISHELL-1 results
#318
hirofumi0810
closed
3 years ago
0
Fix sort_stop_epoch
#317
hirofumi0810
closed
3 years ago
1
Remove redundant lines for DDP
#316
hirofumi0810
closed
3 years ago
1
Efficient CSV loading
#315
hirofumi0810
closed
3 years ago
0
Fix seed in dataloader for resuming training
#314
hirofumi0810
closed
3 years ago
0
Refine speed perturbation script
#313
hirofumi0810
closed
3 years ago
0
Add batch size type for efficient training
#312
hirofumi0810
closed
3 years ago
1
Update AISHELL-1 recipe
#311
hirofumi0810
closed
3 years ago
1
Fix gradient accumulation for resuming multi-GPU training
#310
hirofumi0810
closed
3 years ago
0
Fix LM inference
#309
hirofumi0810
closed
3 years ago
0
Fix LM dataloader
#308
hirofumi0810
closed
3 years ago
1
Support DDP
#307
hirofumi0810
closed
3 years ago
0
Refactor args_common.py
#306
hirofumi0810
closed
3 years ago
0
Update recipe for wandb
#305
hirofumi0810
closed
3 years ago
0
Fix treatment of block tail for streaming inference
#304
hirofumi0810
closed
3 years ago
1
Add safeguard for state reset during streaming inference
#303
hirofumi0810
closed
3 years ago
0
Fix dataloader for RNN models
#302
hirofumi0810
closed
3 years ago
0
Add speaker check for LM state carryover during streaming infernece
#301
hirofumi0810
closed
3 years ago
0
Add missing torch.no_grad() to streaming inference
#300
hirofumi0810
closed
3 years ago
1
Next