Added caching in docker and dmb type changed

Dockerfile

@@ -12,28 +12,33 @@ RUN pip install --trusted-host pypi.python.org -r requirements.txt && \
     pip uninstall transformers && \
     pip install transformers==4.29.2
 
+RUN --mount=type=cache,target=/home/user/.cache/
+
 # Make port 80 available to the world outside this container
 EXPOSE 80
 
 # Set the TORTOISE_MODELS_DIR environment variable
 ENV TORTOISE_MODELS_DIR tortoise/models/pretrained_models
 
+ENV DBM_TYPE=dbm.ndbm
+
+
 # Create the directory for pretrained models
 # RUN mkdir -p $TORTOISE_MODELS_DIR
 
 RUN echo "Downloading models through docker container..."
 
 # Download all the models
-# RUN wget -O $TORTOISE_MODELS_DIR/autoregressive.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/autoregressive.pth && \
-#     wget -O $TORTOISE_MODELS_DIR/classifier.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/classifier.pth && \
-#     wget -O $TORTOISE_MODELS_DIR/clvp2.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/clvp2.pth && \
-#     wget -O $TORTOISE_MODELS_DIR/cvvp.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/cvvp.pth && \
-#     wget -O $TORTOISE_MODELS_DIR/diffusion_decoder.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/diffusion_decoder.pth && \
-#     wget -O $TORTOISE_MODELS_DIR/vocoder.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/vocoder.pth && \
-#     wget -O $TORTOISE_MODELS_DIR/rlg_auto.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/rlg_auto.pth && \
-#     wget -O $TORTOISE_MODELS_DIR/rlg_diffuser.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/rlg_diffuser.pth && \
-#     wget -O $TORTOISE_MODELS_DIR/bigvgan_base_24khz_100band_g.pth https://drive.google.com/uc?id=1_cKskUDuvxQJUEBwdgjAxKuDTUW6kPdY && \
-#     wget -O $TORTOISE_MODELS_DIR/bigvgan_24khz_100band_g.pth https://drive.google.com/uc?id=1wmP_mAs7d00KHVfVEl8B5Gb72Kzpcavp
+RUN wget -O $TORTOISE_MODELS_DIR/autoregressive.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/autoregressive.pth && \
+    wget -O $TORTOISE_MODELS_DIR/classifier.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/classifier.pth && \
+    wget -O $TORTOISE_MODELS_DIR/clvp2.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/clvp2.pth && \
+    wget -O $TORTOISE_MODELS_DIR/cvvp.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/cvvp.pth && \
+    wget -O $TORTOISE_MODELS_DIR/diffusion_decoder.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/diffusion_decoder.pth && \
+    wget -O $TORTOISE_MODELS_DIR/vocoder.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/vocoder.pth && \
+    wget -O $TORTOISE_MODELS_DIR/rlg_auto.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/rlg_auto.pth && \
+    wget -O $TORTOISE_MODELS_DIR/rlg_diffuser.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/rlg_diffuser.pth && \
+    wget -O $TORTOISE_MODELS_DIR/bigvgan_base_24khz_100band_g.pth https://drive.google.com/uc?id=1_cKskUDuvxQJUEBwdgjAxKuDTUW6kPdY && \
+    wget -O $TORTOISE_MODELS_DIR/bigvgan_24khz_100band_g.pth https://drive.google.com/uc?id=1wmP_mAs7d00KHVfVEl8B5Gb72Kzpcavp
 
 RUN echo "Finished downloading models through docker container..."
 

app.py

@@ -33,7 +33,7 @@ LATENT_MODES = [
 
 def main():
     conf = TortoiseConfig()
-    
+    voice_samples, conditioning_latents = None, None
     with st.expander("Create New Voice", expanded=True):
         if "file_uploader_key" not in st.session_state:
             st.session_state["file_uploader_key"] = str(randint(1000, 100000000))
@@ -71,7 +71,11 @@ def main():
                     bytes_data = uploaded_file.read()
                     with open(f"{voices_dir}voice_sample{index}.wav", "wb") as wav_file:
                         wav_file.write(bytes_data)
-
+                
+                #create conditioning latents and save them
+                voice_samples, conditioning_latents = get_condi(
+                    [new_voice_name], []
+                )
                 st.session_state["text_input_key"] = str(randint(1000, 100000000))
                 st.session_state["file_uploader_key"] = str(randint(1000, 100000000))
                 st.experimental_rerun()