Add /upload_sample route to allow upload new speakers using Web API

.gitignore

@@ -1,11 +1,10 @@
 dist
 venv
-speakers
 output
 test
 models
 xtts_api_server/models
 *.pyc
 xtts_api_server/RealtimeTTS/engines/coqui_engine_old.py
 xtts_models
-modules-xtts.txt
+modules-xtts.txt

XTTSv2_AIFF.ipynb

@@ -0,0 +1,116 @@
+{
+  "nbformat": 4,
+  "nbformat_minor": 0,
+  "metadata": {
+    "colab": {
+      "provenance": [],
+      "gpuType": "T4"
+    },
+    "kernelspec": {
+      "name": "python3",
+      "display_name": "Python 3"
+    },
+    "language_info": {
+      "name": "python"
+    },
+    "accelerator": "GPU"
+  },
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "source": [
+        "**Run next block, restart session when asked and run same block until everything installs.**\n",
+        "\n",
+        "Last message should be like this (yes, ignore ERROR)\n",
+        "\n",
+        "\n",
+        "```\n",
+        "ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\n",
+        "torchvision 0.19.0+cu121 requires torch==2.4.0, but you have torch 2.1.1+cu121 which is incompatible\n",
+        "Successfully installed torch-2.1.1+cu121 torchaudio-2.1.1+cu121 triton-2.1.0\n",
+        "\n",
+        "```\n",
+        "\n"
+      ],
+      "metadata": {
+        "id": "FmMMJppi2-Gh"
+      }
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "lT4ieCTnr_L7"
+      },
+      "outputs": [],
+      "source": [
+        "!apt-get update && apt-get install --no-install-recommends -y sox libsox-fmt-all curl wget gcc git git-lfs build-essential libaio-dev libsndfile1 ssh ffmpeg sshpass portaudio19-dev libportaudio2\n",
+        "!git clone https://github.com/abeiro/xtts-api-server\n",
+        "%cd xtts-api-server\n",
+        "!python3 -m pip install --upgrade pip wheel ninja virtualenv\n",
+        "!pip install setuptools==68.1.2\n",
+        "!pip install deepspeed\n",
+        "!pip install -r requirements.txt\n",
+        "!pip install torch==2.1.1+cu121 torchaudio==2.1.1+cu121 --index-url https://download.pytorch.org/whl/cu121"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "**Run this once to get model downloaded and deepspeed compiled. Once uvicorn starts, stop and proceed to next block**\n",
+        "\n",
+        "\n",
+        "\n"
+      ],
+      "metadata": {
+        "id": "u1Ir_s5U3ZcT"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "!python3 -m xtts_api_server --listen --deepspeed"
+      ],
+      "metadata": {
+        "id": "Xs5dXNL43VFB"
+      },
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "**Run and note cloudflare address. Wait for message:**\n",
+        "\n",
+        "```\n",
+        "INFO:     Uvicorn running on http://0.0.0.0:8020 (Press CTRL+C to quit)\n",
+        "\n",
+        "```\n",
+        "\n",
+        "Use http://generated-domain-name:80 in AI-FF conf.\n",
+        "\n"
+      ],
+      "metadata": {
+        "id": "xOowr61z3nT2"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "#Run\n",
+        "!rm nohup.out\n",
+        "!wget -c https://github.com/cloudflare/cloudflared/releases/latest/download/cloudflared-linux-amd64 -O cloudflared-linux-amd64\n",
+        "!chmod +x cloudflared-linux-amd64\n",
+        "!nohup ./cloudflared-linux-amd64 tunnel --url http://localhost:8020 &\n",
+        "!sleep 8\n",
+        "!cat nohup.out\n",
+        "!python3 -m xtts_api_server --listen --deepspeed"
+      ],
+      "metadata": {
+        "id": "xuLmVnEW2PYM"
+      },
+      "execution_count": null,
+      "outputs": []
+    }
+  ]
+}

conf.sh

@@ -0,0 +1,67 @@
+#!/bin/bash
+clear
+cat << EOF
+CHIM XTTS
+
+This will install CHIM XTTS. This is a high quality TTS service that works with Skyrim voices. 
+You can also generate your own voices.
+However it will require around 4GB of VRAM! 
+
+Options:
+* deepseed = Uses more VRAM but it is faster. Needs cuda13 for 50xx series GPUs.
+* lowvram = Uses less VRAM but it is slower. 
+* regular = Middle ground of both options above. RECOMMENDED!
+
+If you are not sure use lowvram.
+
+EOF
+
+if [ ! -d /home/dwemer/python-tts ]; then
+        exit "XTTSv2 not installed"
+fi
+
+mapfile -t files < <(find /home/dwemer/xtts-api-server/ -name "start-*.sh")
+# Check if any files were found
+
+if [ ${#files[@]} -eq 0 ]; then
+    echo "No files found matching the pattern."
+    exit 1
+fi
+
+# Display the files in a numbered list
+echo -e "Select a an option from the list:\n\n"
+for i in "${!files[@]}"; do
+    echo "$((i+1)). ${files[$i]}"
+done
+
+echo "0. Disable service";
+echo
+
+# Prompt the user to make a selection
+read -p "Select an option by picking the matching number: " selection
+
+# Validate the input
+
+if [ "$selection" -eq "0" ]; then
+    echo "Disabling service. Run this again to enable"
+    rm /home/dwemer/xtts-api-server/start.sh &>/dev/null
+    exit 0
+fi
+
+if ! [[ "$selection" =~ ^[0-9]+$ ]] || [ "$selection" -lt 1 ] || [ "$selection" -gt ${#files[@]} ]; then
+    echo "Invalid selection."
+    exit 1
+fi
+
+# Get the selected file
+selected_file="${files[$((selection-1))]}"
+
+echo "You selected: $selected_file"
+
+ln -sf $selected_file /home/dwemer/xtts-api-server/start.sh
+
+
+# Ensure all start scripts are executable
+chmod +x /home/dwemer/xtts-api-server/start-*.sh 2>/dev/null || true
+
+

ddistro_install.sh

@@ -0,0 +1,128 @@
+#!/bin/bash
+
+echo "=== CHIM XTTS Installation ==="
+echo ""
+echo "NOTE: CHIM XTTS and Chatterbox use the same port (8020)."
+echo "      Only one can be enabled at a time."
+echo ""
+
+cd /home/dwemer/xtts-api-server
+python3 -m venv /home/dwemer/python-tts
+source /home/dwemer/python-tts/bin/activate
+
+
+PIDS=$(lsof -ti:8020)
+if [ -n "$PIDS" ]; then
+    echo "Killing processes listening on port 8020: $PIDS"
+    kill -9 $PIDS
+fi
+
+
+echo "This will take a while so please wait."
+
+read -p "Do you want to perform a clean install? (yes/no): " clean_install
+if [[ "$clean_install" =~ ^[Yy][Ee][Ss]$ || "$clean_install" =~ ^[Yy]$ ]]; then
+    rm -fr /home/dwemer/python-tts/*
+	python3 -m venv /home/dwemer/python-tts
+    source /home/dwemer/python-tts/bin/activate
+fi
+
+# Clean previous deepspeed stuff
+rm /home/dwemer/.cache/torch_extensions/py311_cu130/transformer_inference/*
+# Ask user about GPU
+read -p "Are you using a GT10XX series GPU? (yes/no): " gpu_answer
+if [[ "$gpu_answer" =~ ^[Yy][Ee][Ss]$ || "$gpu_answer" =~ ^[Yy]$ ]]; then
+    cu_tag="cu118"
+    torch_url="https://download.pytorch.org/whl/${cu_tag}"
+    torch_ver="2.2.2"
+    torchaudio_ver="2.2.2"
+    python3 -m pip install --upgrade pip wheel ninja virtualenv
+    pip install setuptools==68.1.2
+    # Install app requirements without auto-pulling torch/torchaudio from deps
+    pip install --no-deps -r requirements.txt
+    # Pin to stable, CUDA-tagged PyTorch/Torchaudio that do not require TorchCodec
+    pip cache purge || true
+    pip uninstall -y torch torchaudio torchcodec torchvision || true
+    pip install --no-deps --no-cache-dir --index-url "$torch_url" "torch==${torch_ver}+${cu_tag}" "torchaudio==${torchaudio_ver}+${cu_tag}"
+    pip check || true
+    # Ensure fallback audio loader is available
+    pip install --no-cache-dir soundfile
+    sed -i 's/checkpoint = load_fsspec(model_path, map_location=torch.device("cpu"))\["model"\]/checkpoint = load_fsspec(model_path, map_location=torch.device("cpu"), weights_only=False)["model"]/' /home/dwemer/python-tts/lib/python3.11/site-packages/TTS/tts/models/xtts.py
+
+else
+    read -p "New: Use CUDA13. Needs windows updated drivers. RTX 50XX should use this. Usey cuda13? (yes/no): " gpu5_answer
+    if [[ "$gpu5_answer" =~ ^[Yy][Ee][Ss]$ || "$gpu5_answer" =~ ^[Yy]$ ]]; then
+	    cu_tag="cu130"
+	    torch_url="https://download.pytorch.org/whl/${cu_tag}"
+	    echo "Using torch: $torch_url"
+	    python3 -m pip install --upgrade pip wheel ninja virtualenv
+	    pip install setuptools==68.1.2
+	    # Install app requirements without auto-pulling torch/torchaudio from deps
+	    #pip install --no-deps -r requirements5.txt  --index-url=$torch_url
+	    pip install -r requirements5.txt  --extra-index-url=https://download.pytorch.org/whl/cu130
+	    # Pin to stable, CUDA-tagged PyTorch/Torchaudio that do not require TorchCodec
+	    pip check || true
+	    # Ensure fallback audio loader is available
+	    pip install --no-cache-dir soundfile
+        # Fix symlinks
+		LIBDIR=$(python3 -c 'import site; print(site.getsitepackages()[0])')/nvidia/cu13/lib
+	    for f in "$LIBDIR"/lib*.so.*; do
+    	    base=$(basename "$f")
+        	link="${f%%.so.*}.so"
+            if [ ! -e "$link" ]; then
+				echo "Creating symlink: $(basename "$link") -> $base"
+				ln -s "$base" "$link"
+			fi
+		done		
+	    #pip install xtts-api-server #Fails
+
+    else
+	    cu_tag="cu128"
+	    torch_url="https://download.pytorch.org/whl/${cu_tag}"
+	    echo "Using torch: $torch_url"
+	    python3 -m pip install --upgrade pip wheel ninja virtualenv
+	    pip install setuptools==68.1.2
+	    # Install app requirements without auto-pulling torch/torchaudio from deps
+	    pip install --no-deps -r requirements.txt
+	    # Pin to stable, CUDA-tagged PyTorch/Torchaudio that do not require TorchCodec
+	    pip cache purge || true
+	    pip uninstall -y torch torchaudio torchcodec torchvision || true
+	    #pip install --index-url "$torch_url" torch torchaudio torchcodec torchvision 
+	    pip install torch==2.7.0+cu128 torchaudio==2.7.0+cu128 torchvision==0.22.0+cu128 --index-url=https://download.pytorch.org/whl/
+	    pip check || true
+	    # Ensure fallback audio loader is available
+	    pip install --no-cache-dir soundfile
+	    #pip install xtts-api-server #Fails
+	    sed -i 's/checkpoint = load_fsspec(model_path, map_location=torch.device("cpu"))\["model"\]/checkpoint = load_fsspec(model_path, map_location=torch.device("cpu"), weights_only=False)["model"]/' /home/dwemer/python-tts/lib/python3.11/site-packages/TTS/tts/models/xtts.py
+
+    fi
+fi
+
+
+cp /home/dwemer/TheNarrator.wav speakers/TheNarrator.wav
+
+source /home/dwemer/python-tts/bin/activate
+
+./conf.sh
+
+echo 
+echo "This will start CHIM XTTS to download the selected model"
+echo "Wait for the message 'Uvicorn running on http://0.0.0.0:8020 (Press CTRL+C to quit)'"
+echo "Then close this window. Press ENTER to continue"
+read
+
+echo "please wait...."
+
+# Add CUDA to PATH if the directory exists
+if [ -d "/home/dwemer/python-tts/lib/python3.11/site-packages/nvidia/cu13/lib/" ];
+then
+  export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/dwemer/python-tts/lib/python3.11/site-packages/nvidia/cu13/lib/
+  export PATH=/home/dwemer/python-tts/lib/python3.11/site-packages/nvidia/cu13/bin:$PATH
+  export CUDA_HOME=/home/dwemer/python-tts/lib/python3.11/site-packages/nvidia/cu13
+fi
+
+readlink start.sh | grep -q '/home/dwemer/xtts-api-server/start-deepspeed.sh' && export DEEPSPEED="--deepspeed" ||  export DEEPSPEED=""
+
+python -m xtts_api_server --listen $DEEPSPEED
+
+echo "Press Enter"