have a working model

README.md

@@ -1,6 +1,6 @@
 ---
 title: Emoji Predictor
-emoji: 📊
+emoji: 😎
 colorFrom: pink
 colorTo: indigo
 sdk: gradio

app.py

@@ -1,10 +1,16 @@
 import gradio as gr
 import torch
+import os
+
 from PIL import Image
+from pathlib import Path
+from more_itertools import chunked
+
 from transformers import CLIPProcessor, CLIPModel
 
 checkpoint = "vincentclaes/emoji-predictor"
-no_of_emojis = range(20)
+x_, _, files = next(os.walk("./emojis"))
+no_of_emojis = range(len(files))
 emojis_as_images = [Image.open(f"emojis/{i}.png") for i in no_of_emojis]
 K = 4
 
@@ -12,6 +18,29 @@ processor = CLIPProcessor.from_pretrained(checkpoint)
 model = CLIPModel.from_pretrained(checkpoint)
 
 
+def concat_images(*images):
+    """Generate composite of all supplied images.
+    https://stackoverflow.com/a/71315656/1771155
+    """
+    # Get the widest width.
+    width = max(image.width for image in images)
+    # Add up all the heights.
+    height = max(image.height for image in images)
+    # set the correct size of width and heigtht of composite.
+    composite = Image.new('RGB', (2*width, 2*height))
+    assert K == 4, "We expect 4 suggestions, other numbers won't work."
+    for i, image in enumerate(images):
+        if i == 0:
+            composite.paste(image, (0, 0))
+        elif i == 1:
+            composite.paste(image, (width, 0))
+        elif i == 2:
+            composite.paste(image, (0, height))
+        elif i == 3:
+            composite.paste(image, (width, height))
+    return composite
+
+
 def get_emoji(text, model=model, processor=processor, emojis=emojis_as_images, K=4):
     inputs = processor(text=text, images=emojis, return_tensors="pt", padding=True, truncation=True)
     outputs = model(**inputs)
@@ -23,11 +52,35 @@ def get_emoji(text, model=model, processor=processor, emojis=emojis_as_images, K
     predictions_suggestions_for_chunk = [torch.topk(prob, K).indices.tolist() for prob in probs][0]
     predictions_suggestions_for_chunk
 
-    return [f"emojis/{i}.png" for i in predictions_suggestions_for_chunk]
+    images = [Image.open(f"emojis/{i}.png") for i in predictions_suggestions_for_chunk]
+    images_concat = concat_images(*images)
+    return images_concat
 
 
-text = gr.inputs.Textbox()
+text = gr.inputs.Textbox(placeholder="Enter a text and we will try to predict an emoji...")
 title = "Predicting an Emoji"
-description = "Enter a text and we will try to predict an emoji.\nThe model is a few shot fine tuned CLIP model trained on images of emoji's."
-examples = ["I'm so glad I finally arrived in my holiday resort!"]
-gr.Interface(fn=get_emoji, inputs=text, outputs=gr.Gallery(), examples=examples, title=title).launch()
+description = """You provide a sentence and our few-shot fine tuned CLIP model will predict from the following emoji's:
+\n❤️ 😍 😂 💕 🔥 😊 😎 ✨ 💙 😘 📷 🇺🇸 ☀ 💜 😉 💯 😁 🎄 📸 😜 ☹️ 😭 😔 😡 💢 😤 😳 🙃 😩 😠 🙈 🙄\n
+"""
+article = """
+\n
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+\n
+We fine tuned Open Ai's CLIP model on both text (tweets) and images of emoji's!\n
+The current model is fine-tuned on 15 samples per emoji.
+
+- model: https://huggingface.co/vincentclaes/emoji-predictor \n
+- dataset: https://huggingface.co/datasets/vincentclaes/emoji-predictor \n
+- code: https://github.com/vincentclaes/emoji-predictor \n
+- profile: https://huggingface.co/vincentclaes \n
+"""
+examples = [
+    "I'm so happy for you!",
+    "I'm not feeling great today.",
+    "This makes me angry!",
+    "Can I follow you?",
+    "I'm so bored right now ...",
+]
+gr.Interface(fn=get_emoji, inputs=text, outputs=gr.Image(shape=(72,72)), 
+             examples=examples, title=title, description=description,
+             article=article).launch()

poetry.lock

@@ -155,17 +155,6 @@ category = "main"
 optional = false
 python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
 
-[[package]]
-name = "commonmark"
-version = "0.9.1"
-description = "Python parser for the CommonMark Markdown spec"
-category = "main"
-optional = false
-python-versions = "*"
-
-[package.extras]
-test = ["flake8 (==3.7.8)", "hypothesis (==3.55.3)"]
-
 [[package]]
 name = "contourpy"
 version = "1.0.5"
@@ -211,29 +200,6 @@ category = "main"
 optional = false
 python-versions = ">=3.6"
 
-[[package]]
-name = "docarray"
-version = "0.16.5"
-description = "The data structure for unstructured data"
-category = "main"
-optional = false
-python-versions = "*"
-
-[package.dependencies]
-numpy = "*"
-rich = ">=12.0.0"
-
-[package.extras]
-annlite = ["annlite (>=0.3.10)"]
-benchmark = ["pandas", "seaborn"]
-common = ["protobuf (>=3.13.0)", "lz4", "requests", "matplotlib", "pillow", "fastapi", "uvicorn", "jina-hubble-sdk (>=0.11.0)"]
-elasticsearch = ["elasticsearch (>=8.2.0)"]
-full = ["protobuf (>=3.13.0)", "lz4", "requests", "matplotlib", "pillow", "trimesh", "scipy", "jina-hubble-sdk (>=0.10.0)", "av", "fastapi", "uvicorn", "strawberry-graphql"]
-qdrant = ["qdrant-client (>=0.7.3,<0.8.0)"]
-redis = ["redis (>=4.3.0)"]
-test = ["pytest", "pytest-timeout", "pytest-mock", "pytest-cov", "pytest-repeat", "pytest-reraise", "mock", "pytest-custom-exit-code", "black (==22.3.0)", "tensorflow (==2.7.0)", "paddlepaddle (==2.2.0)", "torch (==1.9.0)", "torchvision (==0.10.0)", "datasets", "onnx", "onnxruntime", "jupyterlab", "transformers (>=4.16.2)", "weaviate-client (>=3.3.0,<3.4.0)", "annlite (>=0.3.10)", "elasticsearch (>=8.2.0)", "redis (>=4.3.0)", "jina"]
-weaviate = ["weaviate-client (>=3.3.0,<3.4.0)"]
-
 [[package]]
 name = "fastapi"
 version = "0.85.0"
@@ -567,6 +533,14 @@ category = "main"
 optional = false
 python-versions = "*"
 
+[[package]]
+name = "more-itertools"
+version = "8.14.0"
+description = "More routines for operating on iterables, beyond itertools"
+category = "main"
+optional = false
+python-versions = ">=3.5"
+
 [[package]]
 name = "multidict"
 version = "6.0.2"
@@ -694,17 +668,6 @@ category = "main"
 optional = false
 python-versions = "*"
 
-[[package]]
-name = "pygments"
-version = "2.13.0"
-description = "Pygments is a syntax highlighting package written in Python."
-category = "main"
-optional = false
-python-versions = ">=3.6"
-
-[package.extras]
-plugins = ["importlib-metadata"]
-
 [[package]]
 name = "pynacl"
 version = "1.5.0"
@@ -809,21 +772,6 @@ idna = {version = "*", optional = true, markers = "extra == \"idna2008\""}
 [package.extras]
 idna2008 = ["idna"]
 
-[[package]]
-name = "rich"
-version = "12.5.1"
-description = "Render rich text, tables, progress bars, syntax highlighting, markdown and more to the terminal"
-category = "main"
-optional = false
-python-versions = ">=3.6.3,<4.0.0"
-
-[package.dependencies]
-commonmark = ">=0.9.0,<0.10.0"
-pygments = ">=2.6.0,<3.0.0"
-
-[package.extras]
-jupyter = ["ipywidgets (>=7.5.1,<8.0.0)"]
-
 [[package]]
 name = "setuptools-scm"
 version = "7.0.5"
@@ -1051,7 +999,7 @@ multidict = ">=4.0"
 [metadata]
 lock-version = "1.1"
 python-versions = "^3.9"
-content-hash = "d1503a7bf493757c63052449403b2d5ed7275e673eaf4ebfcd1c0930e2fada42"
+content-hash = "5bc12d64b69b9c1f0f68ae6858e97ba26663256bae5a9172c0f5bb69402f6c62"
 
 [metadata.files]
 aiohttp = [
@@ -1152,17 +1100,12 @@ click = [
     {file = "click-8.1.3.tar.gz", hash = "sha256:7682dc8afb30297001674575ea00d1814d808d6a36af415a82bd481d37ba7b8e"},
 ]
 colorama = []
-commonmark = [
-    {file = "commonmark-0.9.1-py2.py3-none-any.whl", hash = "sha256:da2f38c92590f83de410ba1a3cbceafbc74fee9def35f9251ba9a971d6d66fd9"},
-    {file = "commonmark-0.9.1.tar.gz", hash = "sha256:452f9dc859be7f06631ddcb328b6919c67984aca654e5fefb3914d54691aed60"},
-]
 contourpy = []
 cryptography = []
 cycler = [
     {file = "cycler-0.11.0-py3-none-any.whl", hash = "sha256:3a27e95f763a428a739d2add979fa7494c912a32c17c4c38c4d5f082cad165a3"},
     {file = "cycler-0.11.0.tar.gz", hash = "sha256:9c87405839a19696e837b3b818fed3f5f69f16f1eec1a1ad77e043dcea9c772f"},
 ]
-docarray = []
 fastapi = []
 ffmpy = []
 filelock = []
@@ -1225,6 +1168,7 @@ matplotlib = []
 mdit-py-plugins = []
 mdurl = []
 monotonic = []
+more-itertools = []
 multidict = [
     {file = "multidict-6.0.2-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:0b9e95a740109c6047602f4db4da9949e6c5945cefbad34a1299775ddc9a62e2"},
     {file = "multidict-6.0.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:ac0e27844758d7177989ce406acc6a83c16ed4524ebc363c1f748cba184d89d3"},
@@ -1302,7 +1246,6 @@ pycparser = [
 pycryptodome = []
 pydantic = []
 pydub = []
-pygments = []
 pynacl = []
 pyparsing = [
     {file = "pyparsing-3.0.9-py3-none-any.whl", hash = "sha256:5026bae9a10eeaefb61dab2f09052b9f4307d44aee4eda64b309723d8d206bbc"},
@@ -1354,7 +1297,6 @@ pyyaml = [
 regex = []
 requests = []
 rfc3986 = []
-rich = []
 setuptools-scm = []
 six = [
     {file = "six-1.16.0-py2.py3-none-any.whl", hash = "sha256:8abb2f1d86890a2dfb989f9a77cfcfd3e47c2a354b01111771326f8aa26e0254"},

pyproject.toml

@@ -9,6 +9,7 @@ python = "^3.9"
 torch = "^1.12.1"
 gradio = "^3.3.1"
 transformers = "^4.22.1"
+more-itertools = "^8.14.0"
 
 [tool.poetry.dev-dependencies]
 

requirements.txt

@@ -34,6 +34,7 @@ matplotlib==3.6.0
 mdit-py-plugins==0.3.0
 mdurl==0.1.2
 monotonic==1.6
+more-itertools==8.14.0
 multidict==6.0.2
 numpy==1.23.3
 orjson==3.8.0