Add files

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.whl filter=lfs diff=lfs merge=lfs -text

.pre-commit-config.yaml

@@ -0,0 +1,33 @@
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v5.0.0
+    hooks:
+      - id: check-executables-have-shebangs
+      - id: check-json
+      - id: check-merge-conflict
+      - id: check-shebang-scripts-are-executable
+      - id: check-toml
+      - id: check-yaml
+      - id: end-of-file-fixer
+      - id: mixed-line-ending
+        args: ["--fix=lf"]
+      - id: requirements-txt-fixer
+      - id: trailing-whitespace
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.9.9
+    hooks:
+      - id: ruff
+        args: ["--fix"]
+      - id: ruff-format
+  - repo: https://github.com/pre-commit/mirrors-mypy
+    rev: v1.15.0
+    hooks:
+      - id: mypy
+        args: ["--ignore-missing-imports"]
+        additional_dependencies:
+          [
+            "types-python-slugify",
+            "types-pytz",
+            "types-PyYAML",
+            "types-requests",
+          ]

.python-version

@@ -0,0 +1 @@
+3.10

.vscode/extensions.json

@@ -0,0 +1,8 @@
+{
+    "recommendations": [
+        "ms-python.python",
+        "charliermarsh.ruff",
+        "streetsidesoftware.code-spell-checker",
+        "tamasfe.even-better-toml"
+    ]
+}

.vscode/settings.json

@@ -0,0 +1,17 @@
+{
+    "editor.formatOnSave": true,
+    "files.insertFinalNewline": false,
+    "[python]": {
+        "editor.defaultFormatter": "charliermarsh.ruff",
+        "editor.formatOnType": true,
+        "editor.codeActionsOnSave": {
+            "source.fixAll.ruff": "explicit",
+            "source.organizeImports": "explicit"
+        }
+    },
+    "[jupyter]": {
+        "files.insertFinalNewline": false
+    },
+    "notebook.output.scrolling": true,
+    "notebook.formatOnSave.enabled": true
+}

app.py

@@ -0,0 +1,198 @@
+#!/usr/bin/env python
+
+import os
+from collections.abc import Iterator
+from threading import Thread
+
+# TODO: remove this once the code is released  # noqa: FIX002, TD002, TD003
+if os.getenv("SPACE_ID"):
+    import shlex
+    import subprocess
+
+    subprocess.run(shlex.split("pip install assets/transformers-4.50.0.dev0-py3-none-any.whl"), check=True)  # noqa: S603
+
+
+import gradio as gr
+import spaces
+import torch
+from transformers import AutoProcessor, Gemma3ForConditionalGeneration, TextIteratorStreamer
+
+model_id = "gg-hf-g/gemma-3-12b-it"
+processor = AutoProcessor.from_pretrained(model_id)
+model = Gemma3ForConditionalGeneration.from_pretrained(
+    model_id, device_map="auto", torch_dtype=torch.bfloat16, attn_implementation="eager"
+)
+
+
+def process_new_user_message(message: dict) -> list[dict]:
+    return [{"type": "text", "text": message["text"]}, *[{"type": "image", "url": path} for path in message["files"]]]
+
+
+def process_history(history: list[dict]) -> list[dict]:
+    messages = []
+    current_user_content: list[dict] = []
+    for item in history:
+        if item["role"] == "assistant":
+            if current_user_content:
+                messages.append({"role": "user", "content": current_user_content})
+                current_user_content = []
+            messages.append({"role": "assistant", "content": [{"type": "text", "text": item["content"]}]})
+        else:
+            content = item["content"]
+            if isinstance(content, str):
+                current_user_content.append({"type": "text", "text": content})
+            else:
+                current_user_content.append({"type": "image", "url": content[0]})
+    return messages
+
+
+@spaces.GPU(duration=120)
+def run(message: dict, history: list[dict], system_prompt: str = "", max_new_tokens: int = 512) -> Iterator[str]:
+    messages = []
+    if system_prompt:
+        messages.append({"role": "system", "content": [{"type": "text", "text": system_prompt}]})
+    messages.extend(process_history(history))
+    messages.append({"role": "user", "content": process_new_user_message(message)})
+
+    inputs = processor.apply_chat_template(
+        messages,
+        add_generation_prompt=True,
+        tokenize=True,
+        return_dict=True,
+        return_tensors="pt",
+    ).to(device=model.device, dtype=torch.bfloat16)
+
+    streamer = TextIteratorStreamer(processor, timeout=60.0, skip_prompt=True, skip_special_tokens=True)
+    generate_kwargs = dict(
+        inputs,
+        streamer=streamer,
+        max_new_tokens=max_new_tokens,
+    )
+    t = Thread(target=model.generate, kwargs=generate_kwargs)
+    t.start()
+
+    output = ""
+    for delta in streamer:
+        output += delta
+        yield output
+
+
+examples = [
+    [
+        {
+            "text": "caption this image",
+            "files": ["assets/sample-images/01.png"],
+        }
+    ],
+    [
+        {
+            "text": "What's the sign says?",
+            "files": ["assets/sample-images/02.png"],
+        }
+    ],
+    [
+        {
+            "text": "Compare and contrast the two images.",
+            "files": ["assets/sample-images/03.png"],
+        }
+    ],
+    [
+        {
+            "text": "List all the objects in the image and their colors.",
+            "files": ["assets/sample-images/04.png"],
+        }
+    ],
+    [
+        {
+            "text": "Describe the atmosphere of the scene.",
+            "files": ["assets/sample-images/05.png"],
+        }
+    ],
+    [
+        {
+            "text": "Write a poem inspired by the visual elements of the images.",
+            "files": ["assets/sample-images/06-1.png", "assets/sample-images/06-2.png"],
+        }
+    ],
+    [
+        {
+            "text": "Compose a short musical piece inspired by the visual elements of the images.",
+            "files": [
+                "assets/sample-images/07-1.png",
+                "assets/sample-images/07-2.png",
+                "assets/sample-images/07-3.png",
+                "assets/sample-images/07-4.png",
+            ],
+        }
+    ],
+    [
+        {
+            "text": "Write a short story about what might have happened in this house.",
+            "files": ["assets/sample-images/08.png"],
+        }
+    ],
+    [
+        {
+            "text": "Create a short story based on the sequence of images.",
+            "files": [
+                "assets/sample-images/09-1.png",
+                "assets/sample-images/09-2.png",
+                "assets/sample-images/09-3.png",
+                "assets/sample-images/09-4.png",
+                "assets/sample-images/09-5.png",
+            ],
+        }
+    ],
+    [
+        {
+            "text": "Describe the creatures that would live in this world.",
+            "files": ["assets/sample-images/10.png"],
+        }
+    ],
+    [
+        {
+            "text": "Read text in the image.",
+            "files": ["assets/additional-examples/1.png"],
+        }
+    ],
+    [
+        {
+            "text": "When is this ticket dated and how much did it cost?",
+            "files": ["assets/additional-examples/2.png"],
+        }
+    ],
+    [
+        {
+            "text": "Read the text in the image into markdown.",
+            "files": ["assets/additional-examples/3.png"],
+        }
+    ],
+    [
+        {
+            "text": "Evaluate this integral.",
+            "files": ["assets/additional-examples/4.png"],
+        }
+    ],
+]
+
+demo = gr.ChatInterface(
+    fn=run,
+    type="messages",
+    textbox=gr.MultimodalTextbox(file_types=["image"], file_count="multiple"),
+    multimodal=True,
+    additional_inputs=[
+        gr.Textbox(label="System Prompt", value="You are a helpful assistant."),
+        gr.Slider(label="Max New Tokens", minimum=100, maximum=2000, step=10, value=500),
+    ],
+    stop_btn=False,
+    title="Gemma 3 12B it",
+    description="<img src='https://huggingface.co/spaces/huggingface-projects/gemma-3-12b-it/resolve/main/assets/logo.png' id='logo' />",
+    examples=examples,
+    run_examples_on_click=False,
+    cache_examples=False,
+    css_paths="style.css",
+    delete_cache=(1800, 1800),
+)
+
+if __name__ == "__main__":
+    demo.launch()

assets/transformers-4.50.0.dev0-py3-none-any.whl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78972eef871bae197d8fd577623f017eb243f730ffecf02a3b70cacdc3d127fd
+size 10933656

pyproject.toml

@@ -0,0 +1,64 @@
+[project]
+name = "gemma-3-12b-it"
+version = "0.1.0"
+description = ""
+readme = "README.md"
+requires-python = ">=3.10"
+dependencies = [
+    "accelerate>=1.4.0",
+    "gradio>=5.20.1",
+    "hf-transfer>=0.1.9",
+    "protobuf>=6.30.0",
+    "sentencepiece>=0.2.0",
+    "spaces>=0.32.0",
+    "torch==2.4.0",
+]
+
+[tool.ruff]
+line-length = 119
+
+[tool.ruff.lint]
+select = ["ALL"]
+ignore = [
+    "COM812", # missing-trailing-comma
+    "D203",   # one-blank-line-before-class
+    "D213",   # multi-line-summary-second-line
+    "E501",   # line-too-long
+    "SIM117", # multiple-with-statements
+    #
+    "D100",    # undocumented-public-module
+    "D101",    # undocumented-public-class
+    "D102",    # undocumented-public-method
+    "D103",    # undocumented-public-function
+    "D104",    # undocumented-public-package
+    "D105",    # undocumented-magic-method
+    "D107",    # undocumented-public-init
+    "EM101",   # raw-string-in-exception
+    "FBT001",  # boolean-type-hint-positional-argument
+    "FBT002",  # boolean-default-value-positional-argument
+    "PD901",   # pandas-df-variable-name
+    "PGH003",  # blanket-type-ignore
+    "PLR0913", # too-many-arguments
+    "PLR0915", # too-many-statements
+    "TRY003",  # raise-vanilla-args
+]
+unfixable = [
+    "F401", # unused-import
+]
+
+[tool.ruff.lint.pydocstyle]
+convention = "google"
+
+[tool.ruff.lint.per-file-ignores]
+"*.ipynb" = ["T201", "T203"]
+
+[tool.ruff.format]
+docstring-code-format = true
+
+[tool.uv.sources]
+transformers = { path = "assets/transformers-4.50.0.dev0-py3-none-any.whl" }
+
+[dependency-groups]
+dev = [
+    "transformers",
+]

requirements.txt

@@ -0,0 +1,229 @@
+# This file was autogenerated by uv via the following command:
+#    uv pip compile pyproject.toml -o requirements.txt
+accelerate==1.4.0
+    # via gemma-3-12b-it (pyproject.toml)
+aiofiles==23.2.1
+    # via gradio
+annotated-types==0.7.0
+    # via pydantic
+anyio==4.8.0
+    # via
+    #   gradio
+    #   httpx
+    #   starlette
+certifi==2025.1.31
+    # via
+    #   httpcore
+    #   httpx
+    #   requests
+charset-normalizer==3.4.1
+    # via requests
+click==8.1.8
+    # via
+    #   typer
+    #   uvicorn
+exceptiongroup==1.2.2
+    # via anyio
+fastapi==0.115.11
+    # via gradio
+ffmpy==0.5.0
+    # via gradio
+filelock==3.17.0
+    # via
+    #   huggingface-hub
+    #   torch
+    #   triton
+fsspec==2025.3.0
+    # via
+    #   gradio-client
+    #   huggingface-hub
+    #   torch
+gradio==5.20.1
+    # via
+    #   gemma-3-12b-it (pyproject.toml)
+    #   spaces
+gradio-client==1.7.2
+    # via gradio
+groovy==0.1.2
+    # via gradio
+h11==0.14.0
+    # via
+    #   httpcore
+    #   uvicorn
+hf-transfer==0.1.9
+    # via gemma-3-12b-it (pyproject.toml)
+httpcore==1.0.7
+    # via httpx
+httpx==0.28.1
+    # via
+    #   gradio
+    #   gradio-client
+    #   safehttpx
+    #   spaces
+huggingface-hub==0.29.2
+    # via
+    #   accelerate
+    #   gradio
+    #   gradio-client
+idna==3.10
+    # via
+    #   anyio
+    #   httpx
+    #   requests
+jinja2==3.1.6
+    # via
+    #   gradio
+    #   torch
+markdown-it-py==3.0.0
+    # via rich
+markupsafe==2.1.5
+    # via
+    #   gradio
+    #   jinja2
+mdurl==0.1.2
+    # via markdown-it-py
+mpmath==1.3.0
+    # via sympy
+networkx==3.4.2
+    # via torch
+numpy==2.2.3
+    # via
+    #   accelerate
+    #   gradio
+    #   pandas
+nvidia-cublas-cu12==12.1.3.1
+    # via
+    #   nvidia-cudnn-cu12
+    #   nvidia-cusolver-cu12
+    #   torch
+nvidia-cuda-cupti-cu12==12.1.105
+    # via torch
+nvidia-cuda-nvrtc-cu12==12.1.105
+    # via torch
+nvidia-cuda-runtime-cu12==12.1.105
+    # via torch
+nvidia-cudnn-cu12==9.1.0.70
+    # via torch
+nvidia-cufft-cu12==11.0.2.54
+    # via torch
+nvidia-curand-cu12==10.3.2.106
+    # via torch
+nvidia-cusolver-cu12==11.4.5.107
+    # via torch
+nvidia-cusparse-cu12==12.1.0.106
+    # via
+    #   nvidia-cusolver-cu12
+    #   torch
+nvidia-nccl-cu12==2.20.5
+    # via torch
+nvidia-nvjitlink-cu12==12.8.93
+    # via
+    #   nvidia-cusolver-cu12
+    #   nvidia-cusparse-cu12
+nvidia-nvtx-cu12==12.1.105
+    # via torch
+orjson==3.10.15
+    # via gradio
+packaging==24.2
+    # via
+    #   accelerate
+    #   gradio
+    #   gradio-client
+    #   huggingface-hub
+    #   spaces
+pandas==2.2.3
+    # via gradio
+pillow==11.1.0
+    # via gradio
+protobuf==6.30.0
+    # via gemma-3-12b-it (pyproject.toml)
+psutil==5.9.8
+    # via
+    #   accelerate
+    #   spaces
+pydantic==2.10.6
+    # via
+    #   fastapi
+    #   gradio
+    #   spaces
+pydantic-core==2.27.2
+    # via pydantic
+pydub==0.25.1
+    # via gradio
+pygments==2.19.1
+    # via rich
+python-dateutil==2.9.0.post0
+    # via pandas
+python-multipart==0.0.20
+    # via gradio
+pytz==2025.1
+    # via pandas
+pyyaml==6.0.2
+    # via
+    #   accelerate
+    #   gradio
+    #   huggingface-hub
+requests==2.32.3
+    # via
+    #   huggingface-hub
+    #   spaces
+rich==13.9.4
+    # via typer
+ruff==0.9.10
+    # via gradio
+safehttpx==0.1.6
+    # via gradio
+safetensors==0.5.3
+    # via accelerate
+semantic-version==2.10.0
+    # via gradio
+sentencepiece==0.2.0
+    # via gemma-3-12b-it (pyproject.toml)
+shellingham==1.5.4
+    # via typer
+six==1.17.0
+    # via python-dateutil
+sniffio==1.3.1
+    # via anyio
+spaces==0.32.0
+    # via gemma-3-12b-it (pyproject.toml)
+starlette==0.46.1
+    # via
+    #   fastapi
+    #   gradio
+sympy==1.13.3
+    # via torch
+tomlkit==0.13.2
+    # via gradio
+torch==2.4.0
+    # via
+    #   gemma-3-12b-it (pyproject.toml)
+    #   accelerate
+tqdm==4.67.1
+    # via huggingface-hub
+triton==3.0.0
+    # via torch
+typer==0.15.2
+    # via gradio
+typing-extensions==4.12.2
+    # via
+    #   anyio
+    #   fastapi
+    #   gradio
+    #   gradio-client
+    #   huggingface-hub
+    #   pydantic
+    #   pydantic-core
+    #   rich
+    #   spaces
+    #   torch
+    #   typer
+    #   uvicorn
+tzdata==2025.1
+    # via pandas
+urllib3==2.3.0
+    # via requests
+uvicorn==0.34.0
+    # via gradio
+websockets==15.0.1
+    # via gradio-client

style.css

@@ -0,0 +1,11 @@
+h1 {
+  text-align: center;
+  display: block;
+}
+
+#logo {
+  display: block;
+  margin: 0 auto;
+  width: 40%;
+  object-fit: contain;
+}