configurable tmp path

Dockerfile

@@ -0,0 +1,26 @@
+FROM jupyter/base-notebook:python-3.9.13
+
+USER root
+
+# Install OS dependencies
+RUN apt-get update && apt-get install -y git-core ffmpeg espeak-ng && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+# Update Conda, create the voicecraft environment, and install dependencies
+RUN conda update -y -n base -c conda-forge conda && \
+    conda create -y -n voicecraft python=3.9.16 && \
+    conda run -n voicecraft conda install -y -c conda-forge montreal-forced-aligner=2.2.17 openfst=1.8.2 kaldi=5.5.1068 && \
+    conda run -n voicecraft pip install -e git+https://github.com/facebookresearch/audiocraft.git@c5157b5bf14bf83449c17ea1eeb66c19fb4bc7f0#egg=audiocraft && \
+    conda run -n voicecraft pip install xformers==0.0.22 && \
+    conda run -n voicecraft pip install torch==2.0.1 && \
+    conda run -n voicecraft pip install torchaudio==2.0.2 && \
+    conda run -n voicecraft pip install tensorboard==2.16.2 && \
+    conda run -n voicecraft pip install phonemizer==3.2.1 && \
+    conda run -n voicecraft pip install datasets==2.16.0 && \
+    conda run -n voicecraft pip install torchmetrics==0.11.1
+    
+
+# Install the Jupyter kernel
+RUN conda install -n voicecraft ipykernel --update-deps --force-reinstall -y && \
+    conda run -n voicecraft python -m ipykernel install --name=voicecraft

README.md

@@ -43,23 +43,26 @@ cd VoiceCraft
 # https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/1.13.5/install-guide.html
 # sudo apt-get install -y nvidia-container-toolkit-base || yay -Syu nvidia-container-toolkit || echo etc...
 
-# 3. Try to start an existing container otherwise create a new one passing in all GPUs
+# 3. First build the docker image
+docker build --tag "voicecraft" .
+
+# 4. Try to start an existing container otherwise create a new one passing in all GPUs
 ./start-jupyter.sh  # linux
 start-jupyter.bat   # windows
 
-# 4. now open a webpage on the host box to the URL shown at the bottom of:
+# 5. now open a webpage on the host box to the URL shown at the bottom of:
 docker logs jupyter
 
-# 5. optionally look inside from another terminal
+# 6. optionally look inside from another terminal
 docker exec -it jupyter /bin/bash
 export USER=(your_linux_username_used_above)
 export HOME=/home/$USER
 sudo apt-get update
 
-# 6. confirm video card(s) are visible inside container
+# 7. confirm video card(s) are visible inside container
 nvidia-smi
 
-# 7. Now in browser, open inference_tts.ipynb and work through one cell at a time
+# 8. Now in browser, open inference_tts.ipynb and work through one cell at a time
 echo GOOD LUCK
 ```
 
@@ -83,6 +86,10 @@ conda install -c conda-forge montreal-forced-aligner=2.2.17 openfst=1.8.2 kaldi=
 
 # to run ipynb
 conda install -n voicecraft ipykernel --no-deps --force-reinstall
+
+# below is only needed if you want to run gradio_app.py
+sudo apt-get install espeak # NOTE: only required if you want to use gradio_app, which is used by whisperx for forced alignment
+sudo apt-get install libespeak-dev # NOTE: only required if you want to use gradio_app, which is used by whisperx for forced alignment
 ```
 
 If you have encountered version issues when running things, checkout [environment.yml](./environment.yml) for exact matching.
@@ -93,12 +100,18 @@ Checkout [`inference_speech_editing.ipynb`](./inference_speech_editing.ipynb) an
 ## Gradio
 After environment setup install additional dependencies:
 ```bash
+apt-get install -y espeak espeak-data libespeak1 libespeak-dev
+apt-get install -y festival*
+apt-get install -y build-essential
+apt-get install -y flac libasound2-dev libsndfile1-dev vorbis-tools
+apt-get install -y libxml2-dev libxslt-dev zlib1g-dev
 pip install -r gradio_requirements.txt
 ```
 
 Run gradio server from terminal or [`gradio_app.ipynb`](./gradio_app.ipynb):
 ```bash
 python gradio_app.py
+TMP_PATH=/tmp python gradio_app.py # if you want to change tmp folder path
 ```
 It is ready to use on [default url](http://127.0.0.1:7860).
 

gradio_app.ipynb

@@ -8,84 +8,28 @@
     "### Only do the below if you are using docker"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "270aa2cc",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# install OS deps\n",
-    "!sudo apt-get update && sudo apt-get install -y \\\n",
-    "    git-core \\\n",
-    "    ffmpeg \\\n",
-    "    espeak-ng"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "8ba5f452",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Update and setup Conda voicecraft environment\n",
-    "!conda update -y -n base -c conda-forge conda\n",
-    "!conda create -y -n voicecraft python=3.9.16 && \\\n",
-    "    conda init bash"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "4ef2935c",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# install conda and pip stuff in the activated conda above context\n",
-    "!echo -e \"Grab a cup a coffee and a slice of pizza...\\n\\n\"\n",
-    "\n",
-    "# make sure $HOME and $USER are setup so this will source the conda environment\n",
-    "!source ~/.bashrc && \\\n",
-    "    conda activate voicecraft && \\\n",
-    "    conda install -y -c conda-forge montreal-forced-aligner=2.2.17 openfst=1.8.2 kaldi=5.5.1068 && \\\n",
-    "    pip install torch==2.0.1 && \\\n",
-    "    pip install tensorboard==2.16.2 && \\\n",
-    "    pip install phonemizer==3.2.1 && \\\n",
-    "    pip install torchaudio==2.0.2 && \\\n",
-    "    pip install datasets==2.16.0 && \\\n",
-    "    pip install torchmetrics==0.11.1\n",
-    "\n",
-    "# do this one last otherwise you'll get an error about torch compiler missing due to xformer mismatch\n",
-    "!source ~/.bashrc && \\\n",
-    "    conda activate voicecraft && \\\n",
-    "    pip install -e git+https://github.com/facebookresearch/audiocraft.git@c5157b5bf14bf83449c17ea1eeb66c19fb4bc7f0#egg=audiocraft"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "2fca57eb",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# okay setup the conda environment such that jupyter notebook can find the kernel\n",
-    "!source ~/.bashrc && \\\n",
-    "    conda activate voicecraft && \\\n",
-    "    conda install -y -n voicecraft ipykernel --update-deps --force-reinstall\n",
-    "\n",
-    "# installs the Jupyter kernel into /home/myusername/.local/share/jupyter/kernels/voicecraft\n",
-    "!source ~/.bashrc && \\\n",
-    "    conda activate voicecraft && \\\n",
-    "    python3 -m ipykernel install --user --name=voicecraft"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,
    "id": "961faa43",
    "metadata": {},
    "outputs": [],
+   "source": [
+    "!source ~/.bashrc && \\\n",
+    "    apt-get update && \\\n",
+    "    apt-get install -y espeak espeak-data libespeak1 libespeak-dev && \\\n",
+    "    apt-get install -y festival* && \\\n",
+    "    apt-get install -y build-essential && \\\n",
+    "    apt-get install -y flac libasound2-dev libsndfile1-dev vorbis-tools && \\\n",
+    "    apt-get install -y libxml2-dev libxslt-dev zlib1g-dev"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "598d75cf",
+   "metadata": {},
+   "outputs": [],
    "source": [
     "!source ~/.bashrc && \\\n",
     "    conda activate voicecraft && \\\n",

gradio_app.py

@@ -1,3 +1,6 @@
+import os
+# os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
+# os.environ["CUDA_VISIBLE_DEVICES"] = "0" # for local use
 import gradio as gr
 import torch
 import torchaudio
@@ -6,14 +9,13 @@ from data.tokenizer import (
     TextTokenizer,
 )
 from models import voicecraft
-import os
 import io
 import numpy as np
 import random
 import uuid
 
 
-TMP_PATH = "./demo/temp"
+TMP_PATH = os.getenv("TMP_PATH", "./demo/temp")
 device = "cuda" if torch.cuda.is_available() else "cpu"
 whisper_model, align_model, voicecraft_model = None, None, None
 
@@ -64,7 +66,7 @@ class WhisperModel:
 class WhisperxModel:
     def __init__(self, model_name, align_model: WhisperxAlignModel):
         from whisperx import load_model
-        self.model = load_model(model_name, device, asr_options={"suppress_numerals": True})
+        self.model = load_model(model_name, device, asr_options={"suppress_numerals": True, "max_new_tokens": None, "clip_timestamps": None, "hallucination_silence_threshold": None})
         self.align_model = align_model
 
     def transcribe(self, audio_path):
@@ -75,9 +77,6 @@ class WhisperxModel:
 def load_models(whisper_backend_name, whisper_model_name, alignment_model_name, voicecraft_model_name):
     global transcribe_model, align_model, voicecraft_model
 
-    os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
-    os.environ["CUDA_VISIBLE_DEVICES"] = "0"
-
     if alignment_model_name is not None:
         align_model = WhisperxAlignModel()
 
@@ -178,7 +177,6 @@ def align(seed, transcript, audio_path):
     } for fragment in fragments["fragments"]]
     segments = align_model.align(segments, audio_path)
     state = get_transcribe_state(segments)
-    print(state)
 
     return [
         state["transcript_with_start_time"], state["transcript_with_end_time"],
@@ -237,10 +235,10 @@ def run(seed, left_margin, right_margin, codec_audio_sr, codec_sr, top_k, top_p,
                 target_transcript = ""
                 for word in transcribe_state["words_info"]:
                     if word["end"] < prompt_end_time:
-                        target_transcript += word["word"]
+                        target_transcript += word["word"] + (" " if word["word"][-1] != " " else "")
                     elif (word["start"] + word["end"]) / 2 < prompt_end_time:
                         # include part of the word it it's big, but adjust prompt_end_time
-                        target_transcript += word["word"]
+                        target_transcript += word["word"] + (" " if word["word"][-1] != " " else "")
                         prompt_end_time = word["end"]
                         break
                     else:
@@ -265,13 +263,13 @@ def run(seed, left_margin, right_margin, codec_audio_sr, codec_sr, top_k, top_p,
                 target_transcript = ""
                 for word in transcribe_state["words_info"]:
                     if word["start"] < edit_start_time:
-                        target_transcript += word["word"]
+                        target_transcript += word["word"] + (" " if word["word"][-1] != " " else "")
                     else:
                         break
                 target_transcript += f" {sentence}"
                 for word in transcribe_state["words_info"]:
                     if word["end"] > edit_end_time:
-                        target_transcript += word["word"]
+                        target_transcript += word["word"] + (" " if word["word"][-1] != " " else "")
             else:
                 target_transcript = sentence
 
@@ -443,7 +441,7 @@ with gr.Blocks() as app:
 
     with gr.Row():
         with gr.Column(scale=2):
-            input_audio = gr.Audio(value="./demo/84_121550_000074_000000.wav", label="Input Audio", type="filepath")
+            input_audio = gr.Audio(value="./demo/84_121550_000074_000000.wav", label="Input Audio", type="filepath", interactive=True)
             with gr.Group():
                 original_transcript = gr.Textbox(label="Original transcript", lines=5, value=demo_original_transcript,
                                                  info="Use whisper model to get the transcript. Fix and align it if necessary.")
@@ -496,22 +494,22 @@ with gr.Blocks() as app:
                 rerun_btn = gr.Button(value="Rerun")
 
     with gr.Row():
-        with gr.Accordion("VoiceCraft config", open=False):
+        with gr.Accordion("Generation Parameters - change these if you are unhappy with the generation", open=False):
-            seed = gr.Number(label="seed", value=-1, precision=0)
+            stop_repetition = gr.Radio(label="stop_repetition", choices=[-1, 1, 2, 3, 4], value=3,
-            left_margin = gr.Number(label="left_margin", value=0.08)
+                                       info="if there are long silence in the generated audio, reduce the stop_repetition to 2 or 1. -1 = disabled")
-            right_margin = gr.Number(label="right_margin", value=0.08)
+            sample_batch_size = gr.Number(label="speech rate", value=4, precision=0,
-            codec_audio_sr = gr.Number(label="codec_audio_sr", value=16000)
+                                          info="The higher the number, the faster the output will be. Under the hood, the model will generate this many samples and choose the shortest one")
-            codec_sr = gr.Number(label="codec_sr", value=50)
+            seed = gr.Number(label="seed", value=-1, precision=0, info="random seeds always works :)")
-            top_k = gr.Number(label="top_k", value=0)
-            top_p = gr.Number(label="top_p", value=0.8)
-            temperature = gr.Number(label="temperature", value=1)
-            stop_repetition = gr.Radio(label="stop_repetition", choices=[-1, 1, 2, 3], value=3,
-                                       info="if there are long silence in the generated audio, reduce the stop_repetition to 3, 2 or even 1, -1 = disabled")
-            sample_batch_size = gr.Number(label="sample_batch_size", value=4, precision=0,
-                                          info="generate this many samples and choose the shortest one")
             kvcache = gr.Radio(label="kvcache", choices=[0, 1], value=1,
                                 info="set to 0 to use less VRAM, but with slower inference")
-            silence_tokens = gr.Textbox(label="silence tokens", value="[1388,1898,131]")
+            left_margin = gr.Number(label="left_margin", value=0.08, info="margin to the left of the editing segment")
+            right_margin = gr.Number(label="right_margin", value=0.08, info="margin to the right of the editing segment")
+            top_p = gr.Number(label="top_p", value=0.8, info="0.8 is a good value, 0.9 is also good")
+            temperature = gr.Number(label="temperature", value=1, info="haven't try other values, do not recommend to change")
+            top_k = gr.Number(label="top_k", value=0, info="0 means we don't use topk sampling, because we use topp sampling")
+            codec_audio_sr = gr.Number(label="codec_audio_sr", value=16000, info='encodec specific, Do not change')
+            codec_sr = gr.Number(label="codec_sr", value=50, info='encodec specific, Do not change')
+            silence_tokens = gr.Textbox(label="silence tokens", value="[1388,1898,131]", info="encodec specific, do not change")
 
     
     audio_tensors = gr.State()

inference_tts.ipynb

@@ -5,106 +5,14 @@
    "metadata": {},
    "source": [
     "VoiceCraft Inference Text To Speech Demo\n",
-    "===\n",
+    "==="
-    "This will install a ton of dependencies all over so consider using the provided docker container start-jupyter script to keep the cruft off your dev box.\n",
-    "\n",
-    "Run the next cells one at a time up until the *STOP* and follow those instructions before continuing. You only have to do this the first time to setup the container."
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Only do the below if you are using docker"
+    "### Select 'voicecraft' as the kernel"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# install OS deps\n",
-    "!sudo apt-get update && sudo apt-get install -y \\\n",
-    "    git-core \\\n",
-    "    ffmpeg \\\n",
-    "    espeak-ng"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Update and setup Conda voicecraft environment\n",
-    "!conda update -y -n base -c conda-forge conda\n",
-    "!conda create -y -n voicecraft python=3.9.16 && \\\n",
-    "    conda init bash"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# install conda and pip stuff in the activated conda above context\n",
-    "!echo -e \"Grab a cup a coffee and a slice of pizza...\\n\\n\"\n",
-    "\n",
-    "# make sure $HOME and $USER are setup so this will source the conda environment\n",
-    "!source ~/.bashrc && \\\n",
-    "    conda activate voicecraft && \\\n",
-    "    conda install -y -c conda-forge montreal-forced-aligner=2.2.17 openfst=1.8.2 kaldi=5.5.1068 && \\\n",
-    "    pip install torch==2.0.1 && \\\n",
-    "    pip install tensorboard==2.16.2 && \\\n",
-    "    pip install phonemizer==3.2.1 && \\\n",
-    "    pip install torchaudio==2.0.2 && \\\n",
-    "    pip install datasets==2.16.0 && \\\n",
-    "    pip install torchmetrics==0.11.1\n",
-    "\n",
-    "# do this one last otherwise you'll get an error about torch compiler missing due to xformer mismatch\n",
-    "!source ~/.bashrc && \\\n",
-    "    conda activate voicecraft && \\\n",
-    "    pip install -e git+https://github.com/facebookresearch/audiocraft.git@c5157b5bf14bf83449c17ea1eeb66c19fb4bc7f0#egg=audiocraft"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# okay setup the conda environment such that jupyter notebook can find the kernel\n",
-    "!source ~/.bashrc && \\\n",
-    "    conda activate voicecraft && \\\n",
-    "    conda install -y -n voicecraft ipykernel --update-deps --force-reinstall\n",
-    "\n",
-    "# installs the Jupyter kernel into /home/myusername/.local/share/jupyter/kernels/voicecraft\n",
-    "!source ~/.bashrc && \\\n",
-    "    conda activate voicecraft && \\\n",
-    "    python3 -m ipykernel install --user --name=voicecraft"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "# STOP\n",
-    "You have to do this part manually using the mouse/keyboard and the tabs at the top.\n",
-    "\n",
-    "* Refresh your browser to make sure it picks up the new kernel.\n",
-    "* Kernel -> Change Kernel -> Select Kernel -> voicecraft\n",
-    "* Kernel -> Restart Kernel -> Yes\n",
-    "\n",
-    "Now you can run the rest of the notebook and get an audio sample output. It will automatically download more models and such. The next time you use this container, you can just start below here as the dependencies will remain available until you delete the docker container."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Only do the above if you are using docker"
    ]
   },
   {
@@ -280,10 +188,6 @@
     "# torchaudio.save(seg_save_fn_gen, gen_audio, codec_audio_sr)\n",
     "# torchaudio.save(seg_save_fn_concat, concated_audio, codec_audio_sr)\n",
     "\n",
-    "# if you get error importing T5 in transformers\n",
-    "# try \n",
-    "# pip uninstall Pillow\n",
-    "# pip install Pillow\n",
     "# you are might get warnings like WARNING:phonemizer:words count mismatch on 300.0% of the lines (3/1), this can be safely ignored"
    ]
   }

start-jupyter.bat

@@ -14,7 +14,7 @@ docker run -it -d ^
     -e JUPYTER_TOKEN=mytoken ^
     -w "/home/%username%" ^
     -v "%cd%":"/home/%username%/work" ^
-    jupyter/base-notebook
+    voicecraft
 
 if %errorlevel% == 0 (
     echo Jupyter container created and running.

start-jupyter.sh

@@ -16,7 +16,7 @@ docker run -it \
     -e GRANT_SUDO=yes \
     -w "/home/${NB_USER}" \
     -v "$PWD":"/home/$USER/work" \
-    jupyter/base-notebook
+    voicecraft
 
 ## `docker logs jupyter` to get the URL link and token e.g.
 ## http://127.0.0.1:8888/lab?token=blahblahblahblabhlaabhalbhalbhal