dusty-nv / jetson-containers

Machine Learning Containers for NVIDIA Jetson and JetPack-L4T
MIT License
2.27k stars 461 forks source link

[BUG] Cannot build dusty/xtts:r36.4, old XTTS repo is deprecated #696

Open wyhanz opened 6 days ago

wyhanz commented 6 days ago

Thank you for your work on this repository, it has been very helpful to me!

I encountered the following error when attempting to build the dusty/xtts:r36.4 image using jetson-containers:

      Error compiling Cython file:
      ------------------------------------------------------------
      ...
          cdef Pool mem
          cpdef readonly StringStore strings
          cpdef public Morphology morphology
          cpdef public object vectors
          cpdef public object _lookups
          cpdef public object writing_system
                ^
      ------------------------------------------------------------

      spacy\vocab.pxd:32:10: Variables cannot be declared with 'cpdef'. Use 'cdef' instead.

      Error compiling Cython file:
      ------------------------------------------------------------
      ...
          cpdef readonly StringStore strings
          cpdef public Morphology morphology
          cpdef public object vectors
          cpdef public object _lookups
          cpdef public object writing_system
          cpdef public object get_noun_chunks
                ^
      ------------------------------------------------------------

      spacy\vocab.pxd:33:10: Variables cannot be declared with 'cpdef'. Use 'cdef' instead.

      Error compiling Cython file:
      ------------------------------------------------------------
      ...
          cdef float prior_prob

      cdef class KnowledgeBase:
          cdef Pool mem
          cpdef readonly Vocab vocab

Upon investigation, I found a similar issue in the original TTS repository issue #4029, which identifies the cause as the old version of the XTTS repository being deprecated and migrated to a new repository idiap/coqui-ai-TTS. It appears that your build scripts still reference a fork of the deprecated version, which could lead to this problem persisting in the future, given that the old repository is no longer supported.

wyhanz commented 6 days ago

directly clone and then install works for me. (JP61, L4T36.4.0, orin 64GB)

RUN git clone https://github.com/idiap/coqui-ai-TTS ${TTS_PATH} && \
    cd ${TTS_PATH} && \
    pip3 install -e .

Test scripts failed, just skip them.