optimum-rbln 0.2.1a0__tar.gz

optimum_rbln-0.2.1a0/.github/ISSUE_TEMPLATE/bug_report.md

@@ -0,0 +1,53 @@
+---
+name: 🐛 Bug Report
+about: Create a report to help us improve optimum-rbln
+title: '[BUG] '
+labels: 'bug'
+assignees: ''
+---
+
+### Environment Information
+- optimum-rbln version:
+- rebel-compiler version:
+- Python version:
+- OS :
+- RBLN hardware (e.g. ATOM-CA25):
+
+### Description
+A clear and concise description of what the bug is.
+
+### Steps To Reproduce
+1. Model/code being used
+```python
+# Your code here
+```
+
+2. Steps to reproduce the behavior:
+   1. 
+   2. 
+   3. 
+
+### Expected Behavior
+A clear and concise description of what you expected to happen.
+
+### Actual Behavior
+A clear and concise description of what actually happened.
+
+### Error Message
+If applicable, paste the full error message here:
+```
+Error message and stack trace
+```
+
+### Model Information
+- Model name/type:
+- Model source (HuggingFace/Custom):
+- Task type:
+
+### Additional Context
+- Have you modified the original model?
+- Are you using quantization?
+- Any other relevant information about your setup or use case
+
+### Logs
+Please include relevant logs, screenshots, or other diagnostic information if available.

optimum_rbln-0.2.1a0/.github/ISSUE_TEMPLATE/config.yml

@@ -0,0 +1,11 @@
+blank_issues_enabled: false
+contact_links:
+  - name: 📚 Documentation
+    url: https://docs.rbln.ai/software/optimum/optimum_rbln.html
+    about: Check our documentation before creating an issue
+  - name: 🛠️ Developers Community
+    url: https://discuss.rebellions.ai/
+    about: Join our developers community
+  - name: 📧 Email Support
+    url: mailto:support@rebellions.ai
+    about: Contact our support team

optimum_rbln-0.2.1a0/.github/ISSUE_TEMPLATE/feature_request.md

@@ -0,0 +1,32 @@
+---
+name: 🚀 Feature Request
+about: Suggest an idea for optimum-rbln
+title: '[FEATURE] '
+labels: 'enhancement'
+assignees: ''
+---
+
+### Is your feature request related to a problem?
+A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]
+
+### Describe the solution you'd like
+A clear and concise description of what you want to happen.
+
+### Describe alternatives you've considered
+A clear and concise description of any alternative solutions or features you've considered.
+
+### Use Case
+Please describe the use case for this feature in detail:
+- Who would use this feature?
+- What is the expected workflow?
+- What are the benefits?
+
+### Technical Details
+If applicable, provide technical details about:
+- Required model architectures
+- Hardware requirements
+- Performance expectations
+- API design suggestions
+
+### Additional context
+Add any other context, code examples, or references about the feature request here.

optimum_rbln-0.2.1a0/.github/ISSUE_TEMPLATE/model_request.md

@@ -0,0 +1,47 @@
+---
+name: 🎯 New Model Support Request
+about: Request support for a new model in optimum-rbln
+title: '[MODEL] '
+labels: 'model-request'
+assignees: ''
+---
+
+### Model Information
+- Model name: 
+- Model link (Hugging Face/Paper/GitHub): 
+- Task type (e.g., Text Generation, Image Classification): 
+
+### Model Details
+- Input modality (Text/Image/Audio/Other): 
+- Output modality (Text/Image/Audio/Other): 
+- Model size (parameters): 
+- Required memory (if known): 
+
+### Input/Output Tensor Information
+- Can input/output shapes be inferred from model config? Yes/No
+- If yes, please explain the relationship:
+  ```
+  Example:
+  - Input shape: (batch_size, sequence_length, hidden_size)
+    - sequence_length from config.max_position_embeddings
+    - hidden_size from config.hidden_size
+  - Output shape: (batch_size, sequence_length, vocab_size)
+    - vocab_size from config.vocab_size
+  ```
+- If no, please specify:
+  - Input tensor shapes:
+  - Output tensor shapes:
+  - Dynamic dimensions (if any):
+
+### Usage Example
+Please provide a minimal example of how the model is typically used:
+```python
+# Example code showing basic model usage
+```
+
+### Popularity & Impact
+Please help us understand the importance and potential impact of this model. You can include metrics such as:
+- Number of downloads/usage statistics (if available)
+- Industry adoption examples
+- Community interest indicators
+- Any other relevant information demonstrating the model's significance

optimum_rbln-0.2.1a0/.github/pull_request_template.md

@@ -0,0 +1,59 @@
+# Pull Request Description
+
+## Type of Change
+<!-- Mark the appropriate option with an [x] -->
+- [ ] New Model Support
+- [ ] Bug fix (non-breaking change which fixes an issue)
+- [ ] New feature (non-breaking change which adds functionality)
+- [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
+- [ ] Documentation update
+- [ ] Performance improvement
+- [ ] Code refactoring
+- [ ] Other (please describe):
+
+
+## Changes Overview
+<!-- Provide a brief summary of the changes in this PR -->
+
+## Motivation and Context
+<!-- Explain why this change is necessary and what problem it solves -->
+
+## Checklist
+<!-- Mark completed items with an [x] -->
+- [ ] I have performed a self-review of my own code
+- [ ] I have added tests that prove my fix is effective or that my feature works (If needed)
+
+## Additional Information
+<!-- Any additional information, configuration, or data that might be necessary to reproduce the issue or use the new feature -->
+
+## Related Issues
+<!-- Link any related issues here using the syntax: Closes #123, Fixes #456 -->
+
+
+
+----
+# Conventional commit
+```
+type(optional scope): description
+```
+----
+# Type candidate
+  - Model Updates
+    - `model`: Adding New models or Bugfix for existing models
+      - ex) Add LlavaNext 
+      - ex) Bugfix Whisper
+  - Enhancements
+    - `performance`: Optimizing some models or this library itself
+      - ex) Loading RBLNModel faster
+      - ex) Optimizing Memory Usage of DecoderOnlyModel
+  - Code Refactor
+    - `refactor`: Re-arrange class architecture, or more.
+      - ex) Refactor Seq2Seq
+  - Documentation
+    - `doc`: Update docstring only     
+  - Library Dependencies
+    - `dependency`: Update requirements, something like that.
+  - Other
+    - `other`: None of above.
+      - ex) ci update
+      - ex) pdm update

optimum_rbln-0.2.1a0/.github/scripts/auto_code_review.py

@@ -0,0 +1,207 @@
+# Copyright 2024 Rebellions Inc.
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at:
+
+#     http://www.apache.org/licenses/LICENSE-2.0
+
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Portions of this software are licensed under the Apache License,
+# Version 2.0. See the NOTICE file distributed with this work for
+# additional information regarding copyright ownership.
+
+# All other portions of this software, including proprietary code,
+# are the intellectual property of Rebellions Inc. and may not be
+# copied, modified, or distributed without prior written permission
+# from Rebellions Inc.
+
+import os
+import sys
+
+import google.generativeai as genai
+import requests
+from github import Github
+
+
+model_name = os.environ["GOOGLE_MODEL_ID"]
+genai.configure(api_key=os.environ["GOOGLE_API_KEY"])
+max_context_token = 100000
+force_review = False
+
+
+def get_pr_diff():
+    api_url = f"https://api.github.com/repos/{os.getenv('GITHUB_REPOSITORY')}/pulls/{os.getenv('PR_NUMBER')}"
+    headers = {
+        "Authorization": f"token {os.getenv('GITHUB_TOKEN')}",
+        "Accept": "application/vnd.github.v3.diff",
+    }
+    response = requests.get(api_url, headers=headers)
+    return response.text if response.status_code == 200 else ""
+
+
+def get_prompt(diff, pr):
+    system_prompt = """You are an experienced software engineer specializing in code reviews for deep learning libraries. Your task is to review code changes and related pull request (PR) information for `optimum-rbln`, a Python library that optimizes Hugging Face models for execution on RBLN NPUs.
+
+Focus on providing actionable and constructive feedback. Don't make generalized suggestions."""
+
+    prompt = f"""
+Review the following code changes(GIT DIFF) along with the pull request (PR) details and provide feedback:
+
+<PR_DESCRIPTION>
+  title : {pr.title}
+  body :
+{pr.body[: pr.body.find("## Related Issues")] if pr.body is not None else ""}
+</PR_DESCRIPTION>
+
+
+<GIT_DIFF>
+{diff}
+</GIT_DIFF>
+"""
+    return system_prompt, prompt
+
+
+def translate_review(review):
+    model = genai.GenerativeModel(
+        model_name,
+        system_instruction="You are a professional translator specializing in technical and software-related content. Keep the technical words in English, but understand the whole sentence and rephrase it in Korean.",
+    )
+    prompt = f"""Translate the following English text into Korean, maintaining technical accuracy and clarity. Include ONLY the translation, NO OTHER EXPLANATIONS or RESPONSES as a chatbot. :
+
+{review}"""
+    response = model.generate_content(prompt)
+
+    return response.text
+
+
+def review_code(system_prompt, prompt):
+    model = genai.GenerativeModel(model_name, system_instruction=system_prompt)
+    response = model.generate_content(prompt)
+    print(prompt)
+    return response.text
+
+
+def remove_file_from_diff(diff_content, file_to_remove):
+    lines = diff_content.splitlines()
+    result = []
+    skip = False
+    file_header = f"diff --git a/{file_to_remove} b/{file_to_remove}"
+
+    for line in lines:
+        if line.startswith("diff --git"):
+            if line == file_header:
+                skip = True
+            else:
+                skip = False
+
+        if not skip:
+            result.append(line)
+
+    return "\n".join(result)
+
+
+def skip_bot(pr):
+    global force_review
+    """
+    Review if
+    1. last commit messages starts with "[autoreview]"
+    2. last comment contains "/autoreview"
+    """
+
+    # Check commit message
+    commits = list(pr.get_commits())
+    if len(commits) == 0:
+        return True
+
+    last_commit = commits[-1]
+    try:
+        commit_message = last_commit.raw_data["commit"]["message"]
+    except KeyError:
+        commit_message = ""
+
+    if commit_message.startswith("[autoreview]"):
+        return False
+
+    # Check the last comment
+    comments = list(pr.get_issue_comments())
+    if len(comments) == 0:
+        return True
+
+    last = comments[-1]
+    if last.user.login.find("github-actions") != -1:
+        return True
+
+    if last.body.find("/autoreview") == -1:
+        return True
+
+    if last.reactions["heart"] > 0:
+        return True
+
+    if last.body.find("force") != -1:
+        force_review = True
+
+    last.create_reaction("heart")
+    return False
+
+
+def main():
+    github_token = os.getenv("GITHUB_TOKEN")
+    pr_number = os.getenv("PR_NUMBER")
+    if not pr_number:
+        pr_number = os.getenv("INPUT_PR_NUMBER")
+
+    if not all([github_token, pr_number]):
+        print("Missing required environment variables")
+        sys.exit(1)
+
+    g = Github(github_token)
+    repo = g.get_repo(os.getenv("GITHUB_REPOSITORY"))
+    pr = repo.get_pull(int(pr_number))
+
+    if skip_bot(pr):
+        print(
+            "To invoke review, Write '/autoreview' and re-run github actions,"
+            " or start the commit message with '[autoreview]'. "
+        )
+        sys.exit(0)
+
+    # Get PR diff
+    diff = get_pr_diff()
+    diff = remove_file_from_diff(diff, "uv.lock")
+
+    # Check diff is available
+    if len(diff) == 0:
+        print("Failed to get the contents of PR Diff. Skipping review.")
+        pr.create_issue_comment("Auto Code Review skipped: Failed to get the diff.")
+        sys.exit(0)
+
+    # check token count
+    system_prompt, prompt = get_prompt(diff, pr)
+    model = genai.GenerativeModel(model_name=model_name, system_instruction=system_prompt)
+    num_tokens = model.count_tokens(prompt).total_tokens
+    if num_tokens > max_context_token and not force_review:
+        msg = f"Diff ({len(diff)}) exceeds maximum allowed tokens ({max_context_token}) > ({num_tokens}). Skipping review."
+        print(msg)
+        pr.create_issue_comment(msg)
+        sys.exit(0)
+
+    # Get Auto review
+    review = review_code(system_prompt, prompt)
+    translation = translate_review(review)
+
+    # Post comment on PR
+    pr.create_issue_comment(f"""# Auto Code Review
+
+- [참고] Auto Code Review를 invoke하려면, commit message의 시작을 [autoreview]로 시작하거나, "/autoreview" 를 comment로 작성한 후,
+해당 commit의 github action에서 code review를 re-run 하시면 됩니다.
+\n\n{review}\n\n{translation}""")
+
+
+if __name__ == "__main__":
+    main()

optimum_rbln-0.2.1a0/.github/scripts/validate_pr_checklist.py

@@ -0,0 +1,84 @@
+# Copyright 2024 Rebellions Inc.
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at:
+
+#     http://www.apache.org/licenses/LICENSE-2.0
+
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Portions of this software are licensed under the Apache License,
+# Version 2.0. See the NOTICE file distributed with this work for
+# additional information regarding copyright ownership.
+
+# All other portions of this software, including proprietary code,
+# are the intellectual property of Rebellions Inc. and may not be
+# copied, modified, or distributed without prior written permission
+# from Rebellions Inc.
+
+import os
+import re
+import sys
+
+from github import Github
+
+
+def read_checklist_from_template():
+    template_path = ".github/pull_request_template.md"
+    checklist_items = []
+
+    with open(template_path, "r") as file:
+        content = file.read()
+        # Find the checklist section
+        checklist_section = re.search(r"## Checklist\n(.*?)\n\n", content, re.DOTALL)
+        if checklist_section:
+            checklist = checklist_section.group(1)
+            # Extract individual checklist items
+            checklist_items = re.findall(r"- \[ \] (.*)", checklist)
+
+    return checklist_items
+
+
+def validate_checklist(body, expected_items):
+    for item in expected_items:
+        if f"- [x] {item}" not in body:
+            print(f"item : {item}")
+            return False
+    return True
+
+
+def main():
+    github_token = os.getenv("GITHUB_TOKEN")
+    pr_number = os.getenv("PR_NUMBER")
+    repo_name = os.getenv("GITHUB_REPOSITORY")
+
+    if not all([github_token, pr_number, repo_name]):
+        print("Missing required environment variables")
+        sys.exit(1)
+
+    g = Github(github_token)
+    repo = g.get_repo(repo_name)
+    pr = repo.get_pull(int(pr_number))
+
+    expected_items = read_checklist_from_template()
+
+    if not expected_items:
+        print("No checklist items found in the PR template.")
+        sys.exit(1)
+
+    if validate_checklist(pr.body, expected_items):
+        print("All checklist items are marked. PR is valid.")
+        sys.exit(0)
+    else:
+        print(f"expected items : {expected_items}")
+        print("Not all checklist items are marked. PR is invalid.")
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()

optimum_rbln-0.2.1a0/.github/workflows/auto_code_review.yml

@@ -0,0 +1,33 @@
+name: Auto Code Review
+
+on:
+  pull_request:
+
+env:
+  GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
+  GOOGLE_MODEL_ID: ${{ vars.GOOGLE_MODEL_ID }}
+
+jobs:
+  code-review:
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout repository
+      uses: actions/checkout@v3
+      with:
+        fetch-depth: 0
+
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.x'
+
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        pip install google-generativeai PyGithub
+
+    - name: Run Auto Code Review
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        PR_NUMBER: ${{ github.event.pull_request.number }}
+      run: python .github/scripts/auto_code_review.py

optimum_rbln-0.2.1a0/.github/workflows/check_code_quality.yml

@@ -0,0 +1,49 @@
+name: check_code_quality
+
+on:
+  push:
+    branches: [main]
+    paths:
+      - "src/**/*.py"
+      - "tests/**/*.py"
+      - "examples/**/*.py"
+
+  pull_request:
+    branches: [main]
+    paths:
+      - "src/**/*.py"
+      - "tests/**/*.py"
+      - "examples/**/*.py"
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
+  cancel-in-progress: true
+
+jobs:
+  build:
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: [3.9]
+        os: [ubuntu-latest]
+
+    runs-on: ${{ matrix.os }}
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Setup uv
+        uses: astral-sh/setup-uv@v3
+        with:
+          enable-cache: true
+          cache-dependency-glob: "uv.lock"
+
+      - name: Setup Python ${{ matrix.python-version }}
+        run: uv python install ${{ matrix.python-version }}
+
+      - name: Install dependencies
+        run: uv sync --frozen --only-group quality
+
+      - name: Check style with ruff
+        run: |
+          uv run --no-sync ruff format --check .
+          uv run --no-sync ruff check .

optimum_rbln-0.2.1a0/.github/workflows/deploy-on-tag.yaml

@@ -0,0 +1,14 @@
+name: Deploy package on tag released
+
+on:
+  push:
+    tags:
+      - 'v*'
+  workflow_dispatch:
+
+jobs:
+  deploy-on-tag:
+    uses: ./.github/workflows/deploy.yaml
+    with:
+      version: ${{ github.ref_name }}
+    secrets: inherit

optimum_rbln-0.2.1a0/.github/workflows/deploy.yaml

@@ -0,0 +1,48 @@
+name: Deploy package to official PyPI
+
+on:
+  workflow_call:
+    inputs:
+      version:
+        required: true
+        type: string
+
+env:
+  PYTHON_VERSION: "3.10"
+
+jobs:
+  deploy:
+    runs-on: rebel-k8s-runner
+    steps:
+      - name: Checkout the repository
+        uses: actions/checkout@v4
+        with:
+          repository: rebellions-sw/optimum-rbln
+          fetch-depth: 0
+          ref: ${{ inputs.version }}
+          token: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Setup uv
+        uses: astral-sh/setup-uv@v3
+        with:
+          enable-cache: true
+          cache-dependency-glob: "uv.lock"
+
+      - name: Setup Python${{ env.PYTHON_VERSION }}
+        run: uv python install ${{ env.PYTHON_VERSION }}
+
+      - name: Install dependencies
+        run: uv sync --frozen --only-group deploy
+
+      - name: Build package
+        run: uv build
+
+      # "uv publish" seems having problem, so use twine until it's fixed
+      - name: Publish package
+        run: |
+          uv run --no-sync \
+            twine upload \
+            --verbose \
+            -u __token__ \
+            -p ${{ secrets.PYPI_ORG_PASSWORD }} \
+            dist/*