diff --git a/.github/actions/deps/action.yaml b/.github/actions/deps/action.yaml index d38d141..04b8c5b 100644 --- a/.github/actions/deps/action.yaml +++ b/.github/actions/deps/action.yaml @@ -13,9 +13,9 @@ runs: python-version: ${{inputs.python-version}} - name: Setup pip shell: sh - run: | + run: | python3 -m ensurepip python3 -m pip install --upgrade pip - - name: Install project + - name: Install project shell: sh - run: pip install ".[dev,train]" \ No newline at end of file + run: pip install ".[dev,train]" diff --git a/.github/pull_request_template.md b/.github/pull_request_template.md index 76e043c..2745f42 100644 --- a/.github/pull_request_template.md +++ b/.github/pull_request_template.md @@ -1,17 +1,17 @@ - ## Change Description - -This project was automatically generated using the LINCC-Frameworks -[python-project-template](https://github.com/lincc-frameworks/python-project-template). For more information about the project template see the +This project was automatically generated using the LINCC-Frameworks +[python-project-template](https://github.com/lincc-frameworks/python-project-template). For more information about the project template see the [documentation](https://lincc-ppt.readthedocs.io/en/latest/). ## Dev Guide - Getting Started @@ -41,7 +41,7 @@ Notes: 3) Look at `pyproject.toml` for other optional dependencies, e.g. you can do `pip install -e ."[dev,train,cuda]"` if you want to use CUDA. 2) `pre-commit install` will initialize pre-commit for this local repository, so that a set of tests will be run prior to completing a local commit. For more - information, see the Python Project Template documentation on + information, see the Python Project Template documentation on [pre-commit](https://lincc-ppt.readthedocs.io/en/latest/practices/precommit.html) 3) Install `pandoc` allows you to verify that automatic rendering of Jupyter notebooks into documentation for ReadTheDocs works as expected. For more information, see @@ -49,7 +49,7 @@ Notes: [Sphinx and Python Notebooks](https://lincc-ppt.readthedocs.io/en/latest/practices/sphinx.html#python-notebooks) -## Running AzureML pipelines +## Running AzureML pipelines This repo contains the evaluation and training pipelines for AutoDoc. @@ -90,4 +90,4 @@ Notes: Example: ```sh az storage blob upload --account-name --container > --file data/data.jsonl -n data/sweetpea/data.jsonl - ``` \ No newline at end of file + ``` diff --git a/azureml/conda.yml b/azureml/conda.yml index f772397..fab3656 100644 --- a/azureml/conda.yml +++ b/azureml/conda.yml @@ -15,4 +15,4 @@ dependencies: - xformers - scipy # This works, while installing from pytorch and cuda from conda does not - - torch==2.0.1 \ No newline at end of file + - torch==2.0.1 diff --git a/azureml/eval.yml b/azureml/eval.yml index 551bd9c..1f31ea4 100644 --- a/azureml/eval.yml +++ b/azureml/eval.yml @@ -1,6 +1,6 @@ $schema: https://azuremlschemas.azureedge.net/latest/commandJob.schema.json command: > - python -m autora.doc.pipelines.main eval + python -m autora.doc.pipelines.main eval ${{inputs.data_dir}}/data.jsonl --model-path ${{inputs.model_path}} --sys-id ${{inputs.sys_id}} @@ -11,12 +11,12 @@ command: > code: ../src inputs: data_dir: - type: uri_folder + type: uri_folder path: azureml://datastores/workspaceblobstore/paths/data/sweetpea/ # Currently models are loading faster directly from HuggingFace vs Azure Blob Storage # model_dir: - # type: uri_folder - # path: azureml://datastores/workspaceblobstore/paths/base_models + # type: uri_folder + # path: azureml://datastores/workspaceblobstore/paths/base_models model_path: meta-llama/Llama-2-7b-chat-hf temperature: 0.7 top_p: 0.95 @@ -37,4 +37,4 @@ environment: # azureml://registries/azureml/environments/acpt-pytorch-2.0-cuda11 display_name: autodoc_prediction compute: azureml:t4cluster experiment_name: evaluation -description: | \ No newline at end of file +description: | diff --git a/azureml/generate.yml b/azureml/generate.yml index fedf7f5..7e3f902 100644 --- a/azureml/generate.yml +++ b/azureml/generate.yml @@ -1,6 +1,6 @@ $schema: https://azuremlschemas.azureedge.net/latest/commandJob.schema.json command: > - python -m autora.doc.pipelines.main generate + python -m autora.doc.pipelines.main generate --model-path ${{inputs.model_path}} --output ./outputs/output.txt --sys-id ${{inputs.sys_id}} @@ -8,23 +8,23 @@ command: > --param temperature=${{inputs.temperature}} --param top_k=${{inputs.top_k}} --param top_p=${{inputs.top_p}} - autora/doc/pipelines/main.py + autora/doc/pipelines/main.py code: ../src inputs: # Currently models are loading faster directly from HuggingFace vs Azure Blob Storage # model_dir: - # type: uri_folder - # path: azureml://datastores/workspaceblobstore/paths/base_models + # type: uri_folder + # path: azureml://datastores/workspaceblobstore/paths/base_models model_path: meta-llama/Llama-2-7b-chat-hf temperature: 0.7 top_p: 0.95 top_k: 40 sys_id: SYS_1 instruc_id: INSTR_SWEETP_1 -environment: +environment: image: mcr.microsoft.com/azureml/curated/acpt-pytorch-2.0-cuda11.7:21 conda_file: conda.yml display_name: autodoc_prediction compute: azureml:t4cluster experiment_name: prediction -description: | \ No newline at end of file +description: |