Upload folder using huggingface_hub
Browse files
run.ipynb
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"cells": [{"cell_type": "markdown", "id": "302934307671667531413257853548643485645", "metadata": {}, "source": ["# Gradio Demo: blocks_neural_instrument_coding"]}, {"cell_type": "code", "execution_count": null, "id": "272996653310673477252411125948039410165", "metadata": {}, "outputs": [], "source": ["!pip install -q gradio "]}, {"cell_type": "code", "execution_count": null, "id": "288918539441861185822528903084949547379", "metadata": {}, "outputs": [], "source": ["# Downloading files from the demo repo\n", "import os\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/flute.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/new-sax-1.mp3\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/new-sax-1.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/new-sax.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/sax.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/sax2.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/trombone.wav"]}, {"cell_type": "code", "execution_count": null, "id": "44380577570523278879349135829904343037", "metadata": {}, "outputs": [], "source": ["# A Blocks implementation of https://erlj.notion.site/Neural-Instrument-Cloning-from-very-few-samples-2cf41d8b630842ee8c7eb55036a1bfd6\n", "\n", "import datetime\n", "import os\n", "import random\n", "\n", "import gradio as gr\n", "from gradio.components import Markdown as m\n", "\n", "
|
|
|
|
| 1 |
+
{"cells": [{"cell_type": "markdown", "id": "302934307671667531413257853548643485645", "metadata": {}, "source": ["# Gradio Demo: blocks_neural_instrument_coding"]}, {"cell_type": "code", "execution_count": null, "id": "272996653310673477252411125948039410165", "metadata": {}, "outputs": [], "source": ["!pip install -q gradio "]}, {"cell_type": "code", "execution_count": null, "id": "288918539441861185822528903084949547379", "metadata": {}, "outputs": [], "source": ["# Downloading files from the demo repo\n", "import os\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/flute.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/new-sax-1.mp3\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/new-sax-1.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/new-sax.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/sax.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/sax2.wav\n", "!wget -q https://github.com/gradio-app/gradio/raw/main/demo/blocks_neural_instrument_coding/trombone.wav"]}, {"cell_type": "code", "execution_count": null, "id": "44380577570523278879349135829904343037", "metadata": {}, "outputs": [], "source": ["# A Blocks implementation of https://erlj.notion.site/Neural-Instrument-Cloning-from-very-few-samples-2cf41d8b630842ee8c7eb55036a1bfd6\n", "\n", "import datetime\n", "import os\n", "import random\n", "\n", "import gradio as gr\n", "from gradio.components import Markdown as m\n", "\n", "def get_time():\n", " now = datetime.datetime.now()\n", " return now.strftime(\"%m/%d/%Y, %H:%M:%S\")\n", "\n", "def generate_recording():\n", " return random.choice([\"new-sax-1.mp3\", \"new-sax-1.wav\"])\n", "\n", "def reconstruct(audio):\n", " return random.choice([\"new-sax-1.mp3\", \"new-sax-1.wav\"])\n", "\n", "io1 = gr.Interface(\n", " lambda x, y, z: os.path.join(os.path.abspath(''),\"sax.wav\"),\n", " [\n", " gr.Slider(label=\"pitch\"),\n", " gr.Slider(label=\"loudness\"),\n", " gr.Audio(label=\"base audio file (optional)\"),\n", " ],\n", " gr.Audio(),\n", ")\n", "\n", "io2 = gr.Interface(\n", " lambda x, y, z: os.path.join(os.path.abspath(''),\"flute.wav\"),\n", " [\n", " gr.Slider(label=\"pitch\"),\n", " gr.Slider(label=\"loudness\"),\n", " gr.Audio(label=\"base audio file (optional)\"),\n", " ],\n", " gr.Audio(),\n", ")\n", "\n", "io3 = gr.Interface(\n", " lambda x, y, z: os.path.join(os.path.abspath(''),\"trombone.wav\"),\n", " [\n", " gr.Slider(label=\"pitch\"),\n", " gr.Slider(label=\"loudness\"),\n", " gr.Audio(label=\"base audio file (optional)\"),\n", " ],\n", " gr.Audio(),\n", ")\n", "\n", "io4 = gr.Interface(\n", " lambda x, y, z: os.path.join(os.path.abspath(''),\"sax2.wav\"),\n", " [\n", " gr.Slider(label=\"pitch\"),\n", " gr.Slider(label=\"loudness\"),\n", " gr.Audio(label=\"base audio file (optional)\"),\n", " ],\n", " gr.Audio(),\n", ")\n", "\n", "demo = gr.Blocks(title=\"Neural Instrument Cloning\")\n", "\n", "with demo.clear():\n", " m(\n", " \"\"\"\n", " ## Neural Instrument Cloning from Very Few Samples\n", " <center><img src=\"https://media.istockphoto.com/photos/brass-trombone-picture-id490455809?k=20&m=490455809&s=612x612&w=0&h=l9KJvH_25z0QTLggHrcH_MsR4gPLH7uXwDPUAZ_C5zk=\" width=\"400px\"></center>\"\"\"\n", " )\n", " m(\n", " \"\"\"\n", " This Blocks implementation is an adaptation [a report written](https://erlj.notion.site/Neural-Instrument-Cloning-from-very-few-samples-2cf41d8b630842ee8c7eb55036a1bfd6) by Nicolas Jonason and Bob L.T. Sturm.\n", " \n", " I've implemented it in Blocks to show off some cool features, such as embedding live ML demos. More on that ahead...\n", " \n", " ### What does this machine learning model do?\n", " It combines techniques from neural voice cloning with musical instrument synthesis. This makes it possible to produce neural instrument synthesisers from just seconds of target instrument audio.\n", " \n", " ### Audio Examples\n", " Here are some **real** 16 second saxophone recordings:\n", " \"\"\"\n", " )\n", " gr.Audio(os.path.join(os.path.abspath(''),\"sax.wav\"), label=\"Here is a real 16 second saxophone recording:\")\n", " gr.Audio(os.path.join(os.path.abspath(''),\"sax.wav\"))\n", "\n", " m(\n", " \"\"\"\\n\n", " Here is a **generated** saxophone recordings:\"\"\"\n", " )\n", " a = gr.Audio(os.path.join(os.path.abspath(''),\"new-sax.wav\"))\n", "\n", " gr.Button(\"Generate a new saxophone recording\")\n", "\n", " m(\n", " \"\"\"\n", " ### Inputs to the model\n", " The inputs to the model are:\n", " * pitch\n", " * loudness\n", " * base audio file\n", " \"\"\"\n", " )\n", "\n", " m(\n", " \"\"\"\n", " Try the model live!\n", " \"\"\"\n", " )\n", "\n", " gr.TabbedInterface(\n", " [io1, io2, io3, io4], [\"Saxophone\", \"Flute\", \"Trombone\", \"Another Saxophone\"]\n", " )\n", "\n", " m(\n", " \"\"\"\n", " ### Using the model for cloning\n", " You can also use this model a different way, to simply clone the audio file and reconstruct it \n", " using machine learning. Here, we'll show a demo of that below:\n", " \"\"\"\n", " )\n", "\n", " a2 = gr.Audio()\n", " a2.change(reconstruct, a2, a2)\n", "\n", " m(\n", " \"\"\"\n", " Thanks for reading this! As you may have realized, all of the \"models\" in this demo are fake. They are just designed to show you what is possible using Blocks \ud83e\udd17.\n", " \n", " For details of the model, read the [original report here](https://erlj.notion.site/Neural-Instrument-Cloning-from-very-few-samples-2cf41d8b630842ee8c7eb55036a1bfd6).\n", " \n", " *Details for nerds*: this report was \"launched\" on:\n", " \"\"\"\n", " )\n", "\n", " t = gr.Textbox(label=\"timestamp\")\n", "\n", " demo.load(get_time, [], t)\n", "\n", "if __name__ == \"__main__\":\n", " demo.launch()\n"]}], "metadata": {}, "nbformat": 4, "nbformat_minor": 5}
|
run.py
CHANGED
|
@@ -7,20 +7,16 @@ import random
|
|
| 7 |
import gradio as gr
|
| 8 |
from gradio.components import Markdown as m
|
| 9 |
|
| 10 |
-
|
| 11 |
def get_time():
|
| 12 |
now = datetime.datetime.now()
|
| 13 |
return now.strftime("%m/%d/%Y, %H:%M:%S")
|
| 14 |
|
| 15 |
-
|
| 16 |
def generate_recording():
|
| 17 |
return random.choice(["new-sax-1.mp3", "new-sax-1.wav"])
|
| 18 |
|
| 19 |
-
|
| 20 |
def reconstruct(audio):
|
| 21 |
return random.choice(["new-sax-1.mp3", "new-sax-1.wav"])
|
| 22 |
|
| 23 |
-
|
| 24 |
io1 = gr.Interface(
|
| 25 |
lambda x, y, z: os.path.join(os.path.dirname(__file__),"sax.wav"),
|
| 26 |
[
|
|
@@ -138,6 +134,5 @@ with demo.clear():
|
|
| 138 |
|
| 139 |
demo.load(get_time, [], t)
|
| 140 |
|
| 141 |
-
|
| 142 |
if __name__ == "__main__":
|
| 143 |
demo.launch()
|
|
|
|
| 7 |
import gradio as gr
|
| 8 |
from gradio.components import Markdown as m
|
| 9 |
|
|
|
|
| 10 |
def get_time():
|
| 11 |
now = datetime.datetime.now()
|
| 12 |
return now.strftime("%m/%d/%Y, %H:%M:%S")
|
| 13 |
|
|
|
|
| 14 |
def generate_recording():
|
| 15 |
return random.choice(["new-sax-1.mp3", "new-sax-1.wav"])
|
| 16 |
|
|
|
|
| 17 |
def reconstruct(audio):
|
| 18 |
return random.choice(["new-sax-1.mp3", "new-sax-1.wav"])
|
| 19 |
|
|
|
|
| 20 |
io1 = gr.Interface(
|
| 21 |
lambda x, y, z: os.path.join(os.path.dirname(__file__),"sax.wav"),
|
| 22 |
[
|
|
|
|
| 134 |
|
| 135 |
demo.load(get_time, [], t)
|
| 136 |
|
|
|
|
| 137 |
if __name__ == "__main__":
|
| 138 |
demo.launch()
|