Pinned issues
Issues
- 1
Error during run time in 1st epoch using conformer_ctc2 OTC
#1288 opened by kerolos - 5
nbest C++ api use too much vram
#1281 opened by binhtranmcs - 0
aux labels id range limit?
#1286 opened by huangruizhe - 13
fail k2.compose(H, LG, inner_labels="tokens")
#1285 opened by liveroomand - 2
Torchscript export error for torch.finfo function
#1284 opened by vinitunni - 4
argmax_internal::PairInputIterator and argmax_internal::PairOutputIterator don't implement the random access iterator concept
#1277 opened by galv - 15
Nbest online decoding bug
#1276 opened by binhtranmcs - 7
- 8
Possible to install k2 in MacOS Apple Silicon
#1240 opened by nabil6391 - 5
ModuleNotFoundError: No module named '_k2'
#1278 opened by AKG-0501 - 2
ModuleNotFoundError: No module named '_k2'
#1275 opened by MisakaMikoto-o - 7
FST is going down abruptly.
#1264 opened by kbramhendra - 2
A warning is always printed when there is an extreme reduction in the encoder output frame rate.
#1271 opened by TeaPoly - 4
K2 installation is not working (not up to date for the new cuda and pytorch versions)
#1267 opened by maxlvov - 2
Installation on Windows with GPU
#1263 opened by ron-wein - 1
- 15
K2 installation on Windowd
#1256 opened by AlexandderGorodetski - 4
Question about k2.union() function.
#1262 opened by lawlict - 7
Question Clarification on streaming decoding for HLG
#1242 opened by kbramhendra - 2
Question about the implementation of MWER
#1260 opened by lawlict - 6
- 26
zipformer error
#1255 opened by AlexandderGorodetski - 5
k2 source build PyTorch 2.1 + Cuda 12.1 fail.
#1252 opened by videodanchik - 2
- 1
Use system kaldifeat
#1250 opened by meerfrau - 2
- 6
Install k2 problem
#1238 opened by HyacinthJingjing - 15
K2 decoder slow 🐌
#1239 opened by david-macleod - 6
Compile LG: When arpa is large, an error occurs
#1233 opened by jingyu1220 - 13
Question about frame number attribute on lattice
#1229 opened by desh2608 - 5
LG decoding,FSA is not safe for different utt
#1235 opened by hmylk - 4
Recommended ways to install k2
#1232 opened by csukuangfj - 2
Provide manylinux wheels
#1227 opened by csukuangfj - 2
- 3
When to use k2.ctc_topo with modified=True?
#1224 opened by wangtiance - 4
Provide a script to install k2 automatically
#1223 opened by csukuangfj - 1
Training with custom CTC topology (with no blanks)
#1222 opened by desh2608 - 0
- 15
Missing complete Segment output
#1216 opened by kbramhendra - 1
RuntimeError: CUDA error: invalid configuration argument
#1220 opened by hy-li751 - 1
k2 compiled without CUDA support
#1219 opened by hy-li751 - 11
y = k2.RaggedTensor(y).to(device) RuntimeError: Some bad things happened. Please read the above error messages and stack
#1214 opened by TszSimLaw - 5
- 2
adding inference_mode() causes "Inference tensors do not track version counter"
#1210 opened by Slyne - 3
Issue with swoosh_r
#1209 opened by AmirHussein96 - 7
Online determinization
#1208 opened by francisr - 2
Decode a Kaldi model with k2
#1207 opened by francisr - 22
- 4
Installation
#1206 opened by Ankita-Phd - 5
Questions for compiling k2 on the ROCm environment
#1202 opened by Mashiro009