ci: lint autofix on pull requests#48

Open

miguelenes wants to merge 85 commits into

VideoVerses:mainfrom

miguelenes:ci/lint-autofix-test

Commits on Jun 22, 2026

chore: update dependencies and Python version to 3.11; upgrade deepspeed, accelerate, and other libraries; add pytest configuration and smoke tests
miguelenes
committed
chore: add pytest configuration to suppress third-party import warnings; refactor imports in multiple files to use updated timm library structure
miguelenes
committed
feat: add benchmark script for attention backends and update pyproject.toml; enhance README with performance tuning details
miguelenes
committed
chore: update pyproject.toml formatting, enhance README with low-VRAM presets, and add new inference scripts for Hunyuan and WanVideo models
miguelenes
committed
chore: update pre-commit configuration to enable linting and unit tests, adjust deepspeed version in pyproject.toml, and optimize DataLoader settings in multiple config files
miguelenes
committed
chore: update dependencies in poetry.lock, enhance pyproject.toml with new inference scripts for various models, and improve README with upgrade notes and CI instructions
miguelenes
committed
chore: add handoff documentation for Cursor integration, update README with installation instructions for AMD ROCm and CPU support, and enhance poetry.lock and pyproject.toml for improved dependenc…
miguelenes
committed
chore: update .gitignore to exclude Jolli Memory logs and add CUDA-related markers in poetry.lock for improved dependency management
miguelenes
committed
chore: enhance dependency management in poetry.lock and pyproject.toml, add new CUDA presets and verification script, and update README with multi-GPU support and inference configurations
miguelenes
committed
chore: add .env.example for environment variable configuration, introduce .gitmodules for SimpleTuner submodule, and create AGENTS.md for project guidelines and agent workflow
miguelenes
committed
chore: refactor imports across multiple files to remove unused dependencies, enhance code clarity, and improve overall project structure
miguelenes
committed
chore: add new configuration files for domain-specific adult content training and inference, update .gitignore to include new data directories, and enhance documentation for fine-tuning processes
miguelenes
committed
chore: add production inference documentation for Wan 2.2 Diffusers 720p, implement benchmark matrix function for improved backend performance testing, and refactor optimization functions for clari…
miguelenes
committed
chore: add CPU smoke presets for various models, enhance capability matrix documentation, and implement new tests for flow tier compatibility and inference limits
miguelenes
committed
chore: implement cloud provisioning scripts for Vast.ai, add environment configuration examples, and enhance documentation for GPU training workflows
miguelenes
committed
refactor: rename project from VideoTuna to PrivTune, update environment variables and documentation, and streamline agent instructions for LoRA training workflows
miguelenes
committed
refactor: update benchmark script for clarity, enhance test cases for attention backends, and remove obsolete test files
miguelenes
committed
refactor: move utility functions to sched_utils.py for better organization and clarity in scheduler files
miguelenes
committed
refactor: remove legacy commands and dependencies, update documentation for PrivTune, and streamline training scripts for improved clarity and usability
miguelenes
committed
refactor: update PrivTune documentation for clarity, streamline training commands, and enhance testing procedures for domain LoRA workflows
miguelenes
committed
docs: add Cursor Cloud specific instructions to AGENTS.md

cursoragent
and
miguelenes
committed
feat: add validation scripts and configurations for Wan 2.2 domain LoRA, enhance testing procedures, and update documentation for improved clarity and usability
miguelenes
committed
feat: introduce Wan 2.1 I2V LoRA training and validation configurations, enhance dataset handling for image-video pairs, and update documentation for improved clarity on I2V workflows
miguelenes
committed
feat: introduce centralized settings management for environment variables, streamline access to configuration in scripts, and enhance CPU inference handling
miguelenes
committed
feat: integrate cyclopts for CLI management, add new inference entrypoints, and streamline inference options handling
miguelenes
committed
refactor: remove black and isort from dev dependencies, update formatting commands to use ruff, and streamline code for improved clarity
miguelenes
committed
feat: add support for fast HF downloads via VIDEOTUNA_FAST_HF_DOWNLOAD, enabling high-performance model pulls during bootstrap and updating documentation for cloud users
miguelenes
committed
refactor: replace decord with PyAV for video processing, update video I/O utilities, and enhance video frame handling across the codebase
miguelenes
committed
Upgrade pytest to 9.1.1 and consolidate test config.
miguelenes
committed
Unify training metrics on TensorBoard and remove unused wandb.
miguelenes
committed
Bump diffusers to 0.38 and harden Wan 2.2 LoRA bridge pin audit.
miguelenes
committed
Add optional torchao transformer quantization for Wan 2.2 low-VRAM inference.
miguelenes
committed
Add tenacity retries for Vast bootstrap network steps.
miguelenes
committed
Remove unused beartype and scope easydict to Wan vendor boundary.
miguelenes
committed
Document DeepSpeed/Lightning pins and fix ZeRO-3 PEFT dtype mismatch.
miguelenes
committed
Restore generation_base.py after accidental mixin refactor in prior commit.
miguelenes
committed
Split generation_base into focused mixins for clearer separation of concerns.
miguelenes
committed
Unify training and flow logging on loguru with structured context.
miguelenes
committed
Add Pydantic v2 validation for Wan domain LoRA YAML configs.
miguelenes
committed
Remove sys.path hacks from Wan entry scripts.
miguelenes
committed
Extend ruff lint to scripts/ and tools/.
miguelenes
committed
Consolidate configs under domain/ and inference/presets/.
miguelenes
committed
Scope type-check to the gradual mypy allowlist.
miguelenes
committed

Commits on Jun 23, 2026