fixed voicefixing not working as intended, load TTS before Gradio in the webui due to how long it takes to initialize tortoise (instead of just having a block to preload it)

README.md

@@ -132,6 +132,7 @@ I'll try and make a list of "common" (or what I feel may be common that I experi
 * `torch.cuda.OutOfMemoryError: CUDA out of memory.`: You most likely have a GPU with low VRAM (~4GiB), and the small optimizations with keeping data on the GPU is enough to OOM. Please open the `start.bat` file and add `--low-vram` to the command (for example: `py app.py --low-vram`) to disable those small optimizations.
 * `WavFileWarning: Chunk (non-data) not understood, skipping it.`: something about your WAVs are funny, and its best to remux your audio files with FFMPEG (included batch file in `.\convert\`).
 	- Honestly, I don't know if this does impact output quality, as I feel it's placebo when I do try and correct this.
+* `Unable to find a valid cuDNN algorithm to run convolution`: a rather weird error message that occurs in the colab notebook. The vanilla auto-batch size calculation is a bit flawed, so try and reduce it to a fixed number in `Settings`, like eight or so.
 
 #### Non-"""Issues"""
 

tortoise_tts.ipynb

@@ -42,8 +42,7 @@
             "%cd tortoise-tts\n",
             "!python -m pip install --upgrade pip\n",
             "!pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu116\n",
-            "!python -m pip install -r ./requirements.txt\n",
-            "!python setup.py install"
+            "!python -m pip install -r ./requirements.txt\n"
          ]
       },
       {
@@ -78,9 +77,9 @@
             "sys.argv = [\"\"]\n",
             "\n",
             "mrq.args = mrq.setup_args()\n",
+            "mrq.tts = mrq.setup_tortoise()\n",
             "mrq.webui = mrq.setup_gradio()\n",
             "mrq.webui.launch(share=True, prevent_thread_lock=True, height=1000)\n",
-            "mrq.tts = mrq.setup_tortoise()\n",
             "mrq.webui.block_thread()"
          ],
          "metadata":{

webui.py

@@ -190,7 +190,7 @@ def generate(
                 'time': run_time
             }
             # save here in case some error happens mid-batch
-            torchaudio.save(f'{outdir}/{voice}_{name}.wav', audio, args.output_sample_rate)
+            torchaudio.save(f'{outdir}/{voice}_{name}.wav', audio, tts.output_sample_rate)
 
     for k in audio_cache:
         audio = audio_cache[k]['audio']
@@ -265,10 +265,10 @@ def generate(
     if args.voice_fixer and voicefixer:
         # we could do this on the pieces before they get stiched up anyways to save some compute
         # but the stitching would need to read back from disk, defeating the point of caching the waveform
-        for path in progress.tqdm(audio_cache, desc="Running voicefix..."):
+        for path in progress.tqdm(output_voices, desc="Running voicefix..."):
             voicefixer.restore(
-                input=f'{outdir}/{voice}_{k}.wav',
-                output=f'{outdir}/{voice}_{k}.wav',
+                input=path,
+                output=path,
                 #cuda=False,
                 #mode=mode,
             )