Introduce modular files for speech models (#35902)

* WAV_2_VEC_2 to WAV2VEC2

* added modular files for hubert, wavlm, wav2vec2_bert, data2vec_audio

* remove unnessary definitions in modulars

* added modular files for UniSpeech, UniSpeechSat, Wav2Vec2Conformer

* docstring fix for UniSpeechForCTC

* removed unneccessary re-definition of modular classes

* reverted lazy imports change on modular_model_converter, type-alias for Wav2Vec2BaseModelOutput

* top-level import of deepspeed in seamless_m4t, speecht5

* avoid tracking imports inside classes, relocate lazy deepspeed, peft imports in their original locations

* convert modular

* tiny modular typing fixes

* some more modular fixes

* make style

---------

Co-authored-by: eustlb <94853470+eustlb@users.noreply.github.com>
Co-authored-by: Eustache Le Bihan <eulebihan@gmail.com>

setup.py

@@ -307,7 +307,12 @@ extras["hub-kernels"] = deps_list("kernels")
 extras["integrations"] = extras["hub-kernels"] + extras["optuna"] + extras["ray"] + extras["sigopt"]
 
 extras["serving"] = deps_list("pydantic", "uvicorn", "fastapi", "starlette")
-extras["audio"] = deps_list("librosa", "pyctcdecode", "phonemizer", "kenlm@git+https://github.com/ydshieh/kenlm@78f664fb3dafe1468d868d71faf19534530698d5")
+extras["audio"] = deps_list(
+    "librosa",
+    "pyctcdecode",
+    "phonemizer",
+    "kenlm@git+https://github.com/ydshieh/kenlm@78f664fb3dafe1468d868d71faf19534530698d5",
+)
 # `pip install ".[speech]"` is deprecated and `pip install ".[torch-speech]"` should be used instead
 extras["speech"] = deps_list("torchaudio") + extras["audio"]
 extras["torch-speech"] = deps_list("torchaudio") + extras["audio"]