openadapt-ml 0.1.0__py3-none-any.whl → 0.2.0__py3-none-any.whl

{openadapt_ml-0.1.0.dist-info → openadapt_ml-0.2.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openadapt-ml
-Version: 0.1.0
+Version: 0.2.0
 Summary: Model-agnostic, domain-agnostic ML engine for GUI automation agents
 Project-URL: Homepage, https://github.com/OpenAdaptAI/openadapt-ml
 Project-URL: Repository, https://github.com/OpenAdaptAI/openadapt-ml
@@ -43,15 +43,29 @@ Requires-Dist: pytest>=9.0.0; extra == 'dev'
 Requires-Dist: ruff>=0.1.0; extra == 'dev'
 Provides-Extra: lambda-labs
 Requires-Dist: requests>=2.28.0; extra == 'lambda-labs'
+Provides-Extra: parquet
+Requires-Dist: pyarrow>=14.0.0; extra == 'parquet'
+Provides-Extra: training
+Requires-Dist: datasets>=2.18.0; extra == 'training'
+Requires-Dist: trl>=0.12.0; extra == 'training'
 Description-Content-Type: text/markdown
 
 # OpenAdapt-ML
 
+[![PyPI version](https://badge.fury.io/py/openadapt-ml.svg)](https://badge.fury.io/py/openadapt-ml)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![Python Version](https://img.shields.io/badge/python-3.12-blue)](https://www.python.org/)
 
 OpenAdapt-ML is a **model-agnostic, domain-agnostic ML engine** for GUI
-automation agents.
+automation agents. It sits above **TRL + Unsloth** (which we use directly for training performance) and provides the GUI-specific layer:
+
+- **Episode semantics**: Step/action/observation alignment, screenshot-action coupling, termination handling
+- **Demo-conditioned inference**: Retrieval-augmented prompting (validated: 33% → 100% first-action accuracy)
+- **Benchmark adapters**: WAA today, OSWorld/WebArena planned
+- **VLM adapters**: Updated with leading GUI-agent SOTA open-source models
+- **Training pipeline**: TRL + Unsloth integration for 2x faster training with 50% less VRAM
+
+OpenAdapt-ML is **not** a training framework, optimizer, hardware orchestrator, or experiment manager. We use TRL/Unsloth, Lambda Labs/Azure, and W&B/MLflow for those.
 
 It provides:
 
@@ -59,24 +73,41 @@ It provides:
 - **Synthetic semantic UI generation** for bootstrapping datasets.
 - **Dataset builders** that turn episodes into next-action SFT samples.
 - **VLM adapters** (Qwen3-VL, Qwen2.5-VL) using Hugging Face + PEFT.
-- A minimal **supervised training loop** for fine-tuning.
+- **SFT training via TRL** with Unsloth optimizations for efficient fine-tuning.
 - A simple **runtime policy** API that predicts the next GUI action.
 
 The design is described in detail in [`docs/design.md`](docs/design.md).
 
 ---
 
-## 1. Quickstart
+## 1. Installation
 
-### 1.1 Install dependencies
+### 1.1 From PyPI (recommended)
 
-From the repository root:
+```bash
+# Install the package
+uv add openadapt-ml
+
+# For training with TRL (recommended for fine-tuning)
+uv add openadapt-ml[training]
+
+# For API-backed VLMs (Claude, GPT)
+uv add openadapt-ml[api]
+```
+
+### 1.2 From source (development)
 
 ```bash
+git clone https://github.com/OpenAdaptAI/openadapt-ml.git
+cd openadapt-ml
 uv sync
 ```
 
-### 1.2 Run a small demo policy
+---
+
+## 2. Quickstart
+
+### 2.1 Run a small demo policy
 
 Run a fast, model-free smoke test:
 
@@ -84,7 +115,7 @@ Run a fast, model-free smoke test:
 uv run python -m openadapt_ml.scripts.demo_policy --backend dummy
 ```
 
-### 1.3 Run the synthetic login benchmark (end-to-end)
+### 2.2 Run the synthetic login benchmark (end-to-end)
 
 On a machine with a suitable GPU, you can reproduce the Qwen3-VL synthetic
 login benchmark (train → eval base/FT → plot) with a single command:
@@ -138,7 +169,7 @@ For complete documentation including training setup, evaluation metrics, SoM mod
 
 ---
 
-## 2. Repository Structure
+## 3. Repository Structure
 
 Key modules:
 
@@ -159,9 +190,9 @@ Key modules:
 - `openadapt_ml/models/dummy_adapter.py`
   - Tiny fake adapter used to validate training and runtime flows without
     loading a real VLM.
-- `openadapt_ml/training/trainer.py`
-  - Minimal supervised training loop (`train_supervised`) with gradient
-    accumulation and logging.
+- `openadapt_ml/training/trl_trainer.py`
+  - TRL-based SFT training (`train_with_trl`) with Unsloth optimizations
+    for 2x faster training and 50% less VRAM.
 - `openadapt_ml/runtime/policy.py`
   - `AgentPolicy` that formats inputs for a VLM and parses textual actions
     like `CLICK(x=..., y=...)` and `DONE()` into structured `Action`s.
@@ -184,12 +215,12 @@ Configs and docs:
 
 ---
 
-## 3. Environment Setup
+## 4. Environment Setup
 
 OpenAdapt-ML targets **Python 3.12** and uses [`uv`](https://github.com/astral-sh/uv)
 for dependency management.
 
-### 2.1 Install and sync
+### 4.1 Install and sync
 
 From the repository root:
 
@@ -202,7 +233,7 @@ uv sync
 This will create a virtual environment (e.g. `.venv/`) and install all
 packages declared in `pyproject.toml`.
 
-### 2.2 Working inside the environment
+### 4.2 Working inside the environment
 
 Use `uv run` to execute Python modules and scripts with the synced
 environment:
@@ -215,12 +246,12 @@ You can also run `pytest` or other tools via `uv run`.
 
 ---
 
-## 4. Synthetic Data & Datasets
+## 5. Synthetic Data & Datasets
 
 The v1 pipeline is validated on **synthetic, semantic UIs**, starting with a
 simple login flow.
 
-### 3.1 Synthetic scenarios
+### 5.1 Synthetic scenarios
 
 OpenAdapt-ML includes synthetic UI generators for structured GUI automation benchmarks.
 Currently two scenarios are supported:
@@ -255,7 +286,7 @@ A more complex registration form with first name, last name, email, password, co
 | Episode Success Rate | **100%** |
 | Episodes / Steps | 32 / 384 |
 
-### 3.2 Generating synthetic data
+### 5.2 Generating synthetic data
 
 Synthetic data is generated on the fly by `generate_synthetic_sessions` in
 `openadapt_ml/ingest/synthetic.py` and used internally by the training
@@ -286,7 +317,7 @@ Each session contains episodes with:
   - An observation (screenshot path).
   - An action (e.g. `CLICK`, `TYPE`, `DONE`).
 
-### 3.3 Next-action SFT samples
+### 5.3 Next-action SFT samples
 
 Episodes are converted into SFT-style samples by
 `build_next_action_sft_samples` in `openadapt_ml/datasets/next_action.py`.
@@ -312,21 +343,20 @@ and its invariants, see `docs/design.md` §7.4.
 
 ---
 
-## 5. Training
+## 6. Training
 
-Training is driven by `openadapt_ml/scripts/train.py` and YAML configs under
-`configs/`.
+Training uses **TRL (Transformer Reinforcement Learning)** with **Unsloth** optimizations
+for efficient VLM fine-tuning. This provides 2x faster training with 50% less VRAM compared
+to standard approaches.
 
-The training script:
+The training pipeline:
 
-1. Loads a config file (YAML).
-2. Generates synthetic sessions.
-3. Flattens to episodes and builds SFT samples.
-4. Wraps them in a `NextActionDataset`.
-5. Instantiates a VLM adapter (e.g. `QwenVLAdapter`).
-6. Runs `train_supervised` over the dataset.
+1. Loads episodes from synthetic data or real recordings.
+2. Converts to TRL-compatible SFT format with images and chat messages.
+3. Fine-tunes using SFTTrainer with LoRA adapters.
+4. Generates checkpoints and training logs for visualization.
 
-### 4.1 Qwen3-VL synthetic training
+### 6.1 Qwen3-VL synthetic training
 
 Config: `configs/qwen3vl_synthetic.yaml`
 
@@ -353,7 +383,7 @@ This will:
 - Run a single-epoch supervised fine-tuning loop.
 - Print loss values as training progresses.
 
-### 4.2 Qwen2.5-VL synthetic training
+### 6.2 Qwen2.5-VL synthetic training
 
 Config: `configs/qwen2_5vl_synthetic.yaml`
 
@@ -378,7 +408,7 @@ format expected by the Qwen2.5-VL processor.
 > Note: Both configs are sized for **small synthetic smoke runs**, not
 > large-scale production training.
 
-### 4.3 Qwen3-VL synthetic login benchmark (hero example)
+### 6.3 Qwen3-VL synthetic login benchmark (hero example)
 
 OpenAdapt-ML ships a **synthetic login** benchmark backed by Qwen3-VL,
 used to compare **base vs LoRA-fine-tuned** models on a hardened synthetic
@@ -413,7 +443,7 @@ It exposes step-level performance metrics, which let us visually answer the ques
 3. **Precision matters**: Fine-tuned models have excellent click precision (85-100% hit rate, <0.05 coord error) while API models struggle with the action format
 4. **Size vs specialization**: The fine-tuned 2B model outperforms the general-purpose Claude Sonnet 4.5, showing that domain-specific fine-tuning trumps raw model size
 
-### 4.4 Set-of-Marks (SoM) Mode: 100% Accuracy
+### 6.4 Set-of-Marks (SoM) Mode: 100% Accuracy
 
 With **Set-of-Marks** visual prompting, fine-tuned Qwen3-VL-2B achieves **100% accuracy** on both login (6-step) and registration (12-step) scenarios:
 
@@ -452,11 +482,11 @@ For the full SoM investigation report, see [`experiments/qwen_login/SOM_INVESTIG
 
 ---
 
-## 6. Grounding Module
+## 7. Grounding Module
 
 OpenAdapt-ML includes a **grounding module** for locating UI elements on screenshots using natural language descriptions. This enables policy/grounding separation where the policy decides *what* to do and the grounder finds *where* to do it.
 
-### 6.1 GeminiGrounder Demo
+### 7.1 GeminiGrounder Demo
 
 The `GeminiGrounder` uses Google's Gemini vision API to locate UI elements:
 
@@ -475,7 +505,7 @@ if candidates:
     print(f"Found at {best.centroid} with {best.confidence:.0%} confidence")
 ```
 
-### 6.2 Set-of-Marks (SoM) Support
+### 7.2 Set-of-Marks (SoM) Support
 
 The grounding module includes functions for extracting all UI elements and overlaying numbered labels (Set-of-Marks):
 
@@ -497,7 +527,7 @@ This enables element-based actions using indices instead of coordinates:
 
 See `docs/gemini_grounding.md` for full documentation and `examples/test_gemini_grounding.py` for a complete example.
 
-### 6.3 Available Grounders
+### 7.3 Available Grounders
 
 | Grounder | Description | Latency | Use Case |
 |----------|-------------|---------|----------|
@@ -505,7 +535,7 @@ See `docs/gemini_grounding.md` for full documentation and `examples/test_gemini_
 | `OracleGrounder` | Ground-truth bboxes | ~0ms | Evaluation |
 | `DetectorGrounder` | Generic wrapper with backend selection | varies | Flexible |
 
-### 6.4 Grounding Evaluation
+### 7.4 Grounding Evaluation
 
 The `openadapt_ml.evals.grounding` module provides metrics for evaluating grounding accuracy:
 
@@ -523,7 +553,7 @@ print(metrics)
 
 ---
 
-## 7. VLM Adapters
+## 8. VLM Adapters
 
 All VLM backends implement the shared `BaseVLMAdapter` interface in
 `openadapt_ml/models/base_adapter.py` (prepare inputs, compute loss, generate
@@ -542,7 +572,7 @@ Current adapters include:
 For full adapter internals and training-time vs runtime behavior, see
 `docs/design.md` §8.
 
-### 7.1 API-backed adapters
+### 8.1 API-backed adapters
 
 To use the API-backed adapter from Python, you can configure API keys via `.env`
 file, environment variables, or pass them explicitly:
@@ -565,12 +595,12 @@ The existing CLI scripts `scripts/demo_policy.py` and
 
 ---
 
-## 8. Runtime Policy & Demos
+## 9. Runtime Policy & Demos
 
 The runtime policy is implemented in `openadapt_ml/runtime/policy.py` as
 `AgentPolicy`.
 
-### 8.1 AgentPolicy
+### 9.1 AgentPolicy
 
 `AgentPolicy` is initialized with a VLM adapter (dummy or real). Given an
 SFT-style sample, it:
@@ -581,7 +611,7 @@ SFT-style sample, it:
    - `DONE()`
 3. Returns a structured `Action` plus an optional free-form `thought`.
 
-### 8.2 Demo script
+### 9.2 Demo script
 
 `openadapt_ml/scripts/demo_policy.py` demonstrates how to use
 `AgentPolicy` with different backends.
@@ -613,7 +643,7 @@ Each invocation will:
 
 ---
 
-## 9. Testing
+## 10. Testing
 
 Basic tests are provided under `tests/`.
 
@@ -623,26 +653,26 @@ Run the test suite with:
 uv run pytest
 ```
 
-In particular:
+Key test files:
 
-- `tests/test_training_dummy.py` runs a smoke test over the training loop
-  using `DummyAdapter`.
+- `tests/test_training_dummy.py` - Tests TRL training configuration and sample conversion
+- `tests/test_local_cli.py` - Tests local training CLI commands (status, check, viewer)
 
 ---
 
-## 10. Training on Real Data
+## 11. Training on Real Data
 
 OpenAdapt-ML supports training on real GUI recordings from two sources:
 1. **openadapt-capture** - New lightweight recording format
 2. **OpenAdapt database** - Original OpenAdapt recordings (legacy)
 
-### 10.1 Training on openadapt-capture recordings
+### 11.1 Training on openadapt-capture recordings
 
 [openadapt-capture](https://github.com/OpenAdaptAI/openadapt-capture) is a lightweight GUI recording tool.
 
 ```bash
 # Install openadapt-capture
-uv pip install openadapt-capture
+uv add openadapt-capture
 
 # Record a workflow (e.g., turning off Night Shift)
 openadapt-capture record --output ~/captures/turn-off-nightshift
@@ -656,7 +686,7 @@ uv run python -m openadapt_ml.scripts.train \
 
 The goal is automatically derived from the directory name (e.g., `"Turn off nightshift"`).
 
-### 10.2 Compare human vs AI predictions
+### 11.2 Compare human vs AI predictions
 
 ```bash
 uv run python -m openadapt_ml.scripts.compare \
@@ -673,11 +703,11 @@ The comparison viewer shows:
 
 ---
 
-## 11. Local Training (CUDA / Apple Silicon)
+## 12. Local Training (CUDA / Apple Silicon)
 
 Train locally on your own GPU. Auto-detects CUDA or Apple Silicon (MPS).
 
-### 11.1 Quick start
+### 12.1 Quick start
 
 ```bash
 # Train on a capture (auto-detects device and config)
@@ -686,7 +716,7 @@ uv run python -m openadapt_ml.cloud.local train \
   --open  # Opens dashboard in browser
 ```
 
-### 11.2 Training workflow
+### 12.2 Training workflow
 
 ```bash
 # Check device and training status
@@ -713,11 +743,11 @@ uv run python -m openadapt_ml.cloud.local compare \
 
 ---
 
-## 12. Cloud GPU Training (Lambda Labs)
+## 13. Cloud GPU Training (Lambda Labs)
 
 For faster training on powerful GPUs, use Lambda Labs. Full documentation: [`docs/cloud_gpu_training.md`](docs/cloud_gpu_training.md).
 
-### 12.1 Quick start
+### 13.1 Quick start
 
 ```bash
 # Set API key
@@ -729,7 +759,7 @@ uv run python -m openadapt_ml.cloud.lambda_labs train \
   --goal "Turn off Night Shift in System Settings"
 ```
 
-### 12.2 Manual workflow
+### 13.2 Manual workflow
 
 ```bash
 # List available instances and pricing
@@ -751,7 +781,7 @@ uv run python -m openadapt_ml.cloud.lambda_labs download <instance_id>
 uv run python -m openadapt_ml.cloud.lambda_labs terminate <instance_id>
 ```
 
-### 12.3 Training visualization
+### 13.3 Training visualization
 
 The training process generates:
 - **`training_output/dashboard.html`** - Real-time training dashboard with loss curves
@@ -790,9 +820,15 @@ uv run python -m openadapt_ml.cloud.local serve --port 8080 --open
 - `Home` / `End` - First/last frame
 - `O` - Toggle click overlay
 
+**Benchmark Viewer:**
+
+![Benchmark Viewer](docs/images/benchmark_viewer.png)
+
+*View benchmark evaluation results with task-level filtering, success/failure status, and run comparison. Shows Claude achieving 30% on mock evaluation tasks (simulated environment for testing the pipeline - real WAA evaluation requires Windows VMs).*
+
 ---
 
-## 13. Limitations & Notes
+## 14. Limitations & Notes
 
 - **Apple Silicon / bitsandbytes**:
   - Example configs are sized for CPU / Apple Silicon development runs; see
@@ -805,12 +841,18 @@ uv run python -m openadapt_ml.cloud.local serve --port 8080 --open
 - **Evaluation**:
   - v1 focuses on smoke tests and qualitative behavior on synthetic data.
     More formal evaluation scripts and metrics are planned.
+- **Windows Agent Arena (WAA) on Azure**:
+  - WAA requires nested virtualization (Windows VM inside Docker via QEMU)
+  - Azure ML managed compute does not support nested virtualization
+  - For real WAA evaluation, use dedicated VMs with Dv3/Ev3 series or run locally
+  - Mock evaluation (`test-mock`) validates the pipeline without Windows VMs
+  - See `CLAUDE.md` for detailed workarounds and infrastructure setup
 
 For deeper architectural details, see [`docs/design.md`](docs/design.md).
 
 ---
 
-## 14. Roadmap
+## 15. Roadmap
 
 For the up-to-date, prioritized roadmap (including concrete implementation
 targets and agent-executable acceptance criteria), see

openadapt_ml-0.2.0.dist-info/RECORD

@@ -0,0 +1,86 @@
+openadapt_ml/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+openadapt_ml/config.py,sha256=eH5WTKRPkkidjkNb25Wn_dUAizUQTsVPMYLDq_ekPJQ,1865
+openadapt_ml/benchmarks/__init__.py,sha256=RRAoXm_al-DgStCKCycnC60_iZo52mLmUu5nN5IPfxY,3855
+openadapt_ml/benchmarks/agent.py,sha256=kuq-dWqWJH8ogjklAFdnbG_6zVzHtd6Ab2rPxRjvHtU,29724
+openadapt_ml/benchmarks/azure.py,sha256=AI5ZdsNxmQ197ymOj68lTLAtl-eByuguhACqv7UZhIg,28145
+openadapt_ml/benchmarks/base.py,sha256=d7T_zMlMPlN0beDWkpzOvOHYQO6QnsePLQ45iKbi66Y,11667
+openadapt_ml/benchmarks/cli.py,sha256=ElBZDcmDOpA0uAoHYKRx3fLAJ5v8dH4R9SxfhmwTZWw,204898
+openadapt_ml/benchmarks/data_collection.py,sha256=EYOsYnFQifF3MXD0TZxznd-HbODiovnGDtxGjSMpO-Y,14652
+openadapt_ml/benchmarks/live_tracker.py,sha256=1SukwgRYbAzSMzHOhPZOSgZ58L44CYtX-KOfAzyAJZw,5130
+openadapt_ml/benchmarks/runner.py,sha256=a52GasPKEnNgevxilQAI1z8FnzwWddLiDY60NYWxpZk,13616
+openadapt_ml/benchmarks/viewer.py,sha256=Mjbt67gfnSw29rNhU4JYL0cdAc-UZ3QpxmG6GhKnD68,41111
+openadapt_ml/benchmarks/vm_monitor.py,sha256=NKhLF66hcBqxIDHk-iPKn31AfzpXyE-9IkBTYX2k70I,20330
+openadapt_ml/benchmarks/waa.py,sha256=V_RtDfOLW2jD8MN-M9YJYvt67NQCXrUkVxOtcNYVLHo,27008
+openadapt_ml/benchmarks/waa_live.py,sha256=U92KaBWCpJmwqSURJeG5biEolTyBCvyCikWRz05IV-A,22982
+openadapt_ml/benchmarks/waa_deploy/Dockerfile,sha256=m9Nsei1v_m0jF1HWc7yCMz-_wD3rHHv9H9HBC2lFVSE,10969
+openadapt_ml/benchmarks/waa_deploy/__init__.py,sha256=KV71HrrgETytfY0i4vFSi-yM0KjoQP2hd9Bl03cZ9yc,320
+openadapt_ml/benchmarks/waa_deploy/api_agent.py,sha256=dbazhRxc554901LFiVuj6sMmOgoHtTKl8XIAiIJrFWU,20024
+openadapt_ml/benchmarks/waa_deploy/start_waa_server.bat,sha256=YxgrSWh76zLijlpxEpulWf9To2JtJ-yR42lr2WyTXiY,1496
+openadapt_ml/cloud/__init__.py,sha256=XYrvxivJeZ8qYnuGod5kodMlm3iT2OK2GAApO3CNB0c,133
+openadapt_ml/cloud/azure_inference.py,sha256=lqkG86Dn6x2Rp7Y-X9tK_8mn0QfRvaDQNRHg-kJkQyc,15736
+openadapt_ml/cloud/lambda_labs.py,sha256=jZ700gNR8zxO2qGa7yqByKXgb4-18zN30YQdDZ3Fbuk,102206
+openadapt_ml/cloud/local.py,sha256=AX7gxgkwGwOJ1ogp-rQefKDyK3MDNMWA9W2qkwaV66g,104782
+openadapt_ml/cloud/ssh_tunnel.py,sha256=q4VEKT4cEU0-hAP1t9fL2jRur0FMRxg8ZV6j1IQV45k,20824
+openadapt_ml/datasets/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+openadapt_ml/datasets/next_action.py,sha256=eNNwwdRE-pJa1DCgEIr9eTipZ7pDAEDVxkc9UPObJLs,21219
+openadapt_ml/evals/__init__.py,sha256=Kx7bSvPHwmoGVI3q1wS_lC17W2S32YHj0459JRqu6Ow,573
+openadapt_ml/evals/grounding.py,sha256=un4bWrhqwrJ4O9QqF40NfnhKeOMntGdCJcwXz6ZzNJ4,8447
+openadapt_ml/evals/plot_eval_metrics.py,sha256=gsikQ3MSUY7Pw61D8lB286q0MPpBL9E05UFHPrawViw,5237
+openadapt_ml/evals/trajectory_matching.py,sha256=p40wDDlD0AyiY6vsgkcp6FBDISyKKQKycsrEz7uGF8Y,20616
+openadapt_ml/experiments/demo_prompt/__init__.py,sha256=dwS0bI53jXMzHE-DPhb_mhmPdoqSZRIcNbV79wt8KPM,454
+openadapt_ml/experiments/demo_prompt/format_demo.py,sha256=rIXcjiYhI2YuLDUjY4iTxPp4ZmtvKibQQYTCfEn4lZs,6553
+openadapt_ml/experiments/demo_prompt/run_experiment.py,sha256=uchhadnqxauBXxlalTh7wXPLOOXk7NBY1mwXiN7rpHI,16309
+openadapt_ml/experiments/demo_prompt/results/experiment_20251231_002125.json,sha256=08oryOF126toTQDN9xciodavvfsaWNnXuBs0aULwpfI,5326
+openadapt_ml/experiments/demo_prompt/results/experiment_n30_20251231_165958.json,sha256=u03VgYTQia_HzilzNjxdGLpUSdbo4SzmHqI-GXlvurg,26915
+openadapt_ml/experiments/demo_prompt/results/multistep_20251231_025051.json,sha256=FA1JgXXLor6on3lHlfJdNSuKzBca30ggH8IWSJEmmfA,11517
+openadapt_ml/experiments/waa_demo/__init__.py,sha256=9M8iLxO9GWAw-FIB-0tzsqaweLcO5EVP1Sc5BoK16iU,363
+openadapt_ml/experiments/waa_demo/demos.py,sha256=UwO0EYy8wUEggaBaI_cXuYe_jwSB1hx3ZtPf-z9bhjc,13796
+openadapt_ml/experiments/waa_demo/runner.py,sha256=OxgQhZIqhYeGDYmAcQLEsFh3B053rnuPL0ZEIoXz0bI,24327
+openadapt_ml/experiments/waa_demo/tasks.py,sha256=jw1QwbOt8xmWBW2lmBWcJzKBXssjv_e0j49MlC2rVJY,5425
+openadapt_ml/export/__init__.py,sha256=mKehKHOio4jGcK-3r0-pb446GdKMPs0O9hAu4S0_R7s,266
+openadapt_ml/export/__main__.py,sha256=0ObtWcdzf6p7gPwhNlCKpNm2FIhmusdYNkuk8tyt77U,149
+openadapt_ml/export/cli.py,sha256=goTKNq9cOO9wsdNluLMH_-f9kdWShH3FPP8sCZ6KaPI,2331
+openadapt_ml/export/parquet.py,sha256=1BGHJKJc302trn64mUbuhxYMBPNvH80QW5f8GM63njk,9664
+openadapt_ml/grounding/__init__.py,sha256=uMvcALFRXmKD6PHhqLZ24Y6zhRUs46_PnWYqiqJP5cM,1412
+openadapt_ml/grounding/base.py,sha256=mnjT25nxltZCD0VBzgIgj2kuCcB4sgXBN97MBaW5P6c,7688
+openadapt_ml/grounding/detector.py,sha256=z-6Y_jnUNnhviUjKv6okjJ0r13DmBiNZMzNJo0rTlBY,19786
+openadapt_ml/ingest/__init__.py,sha256=7YASU-pOXtjalcRJ3WCbGuEWne0aVZtypsfcEmp-yFE,1437
+openadapt_ml/ingest/capture.py,sha256=SR22U6M7hU_QAsCYAG4JaxGFHpLZg96k_SJ-tkXAgbs,10218
+openadapt_ml/ingest/loader.py,sha256=PPBFMA9d7oc2bjXgrdPCZQBnv3MvoDFQNJvXeI90_j8,9865
+openadapt_ml/ingest/synthetic.py,sha256=DSUyqbLxHtJjELitTP2C_3bv3-E0UW7P-RFAO9W8XFs,39302
+openadapt_ml/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+openadapt_ml/models/api_adapter.py,sha256=9EfQrXuFYIT-ea-wmGzJlM4thgVVDIRZnqqS_gL-PbU,6541
+openadapt_ml/models/base_adapter.py,sha256=AG40BhdDORwUNYYg3DM1wsErX57aEJrkK0tyn0tEKhI,2050
+openadapt_ml/models/dummy_adapter.py,sha256=h4Zu-rjWgtG1r8jRtcsrX-FZm8iImrhrTQ7TsLfjE8A,1581
+openadapt_ml/models/qwen_vl.py,sha256=sIJUtDRXAcz9zh3uRWOMYVOxdWIXlcM3vazdNOAsY_U,17239
+openadapt_ml/retrieval/README.md,sha256=j4gXhTo6yH-5cuw4ER4174V-U6TQakOVT6Hj4kj7B0I,5696
+openadapt_ml/retrieval/USAGE.md,sha256=XDIrX-94Z5nC-wvnBY5yF5gTqUYixxCC3wwUFvQx5YM,9278
+openadapt_ml/retrieval/__init__.py,sha256=xocb84riKLUCezUioKssFRhAQsnvexh4W932o368_qg,2726
+openadapt_ml/retrieval/demo_retriever.py,sha256=fYPLKzlG7yGHfV-F-TECBbxkMyP1fqdqPzTJ7G0oXYU,29184
+openadapt_ml/retrieval/embeddings.py,sha256=W4Bqo48Ds4BI9zZg1awkSi9p5kplRRALEjEgVh-jbsY,19239
+openadapt_ml/retrieval/index.py,sha256=UBFnSxp5T5eKt2txFcd0FytKCw1qxONZfxnFJVrduRQ,5710
+openadapt_ml/retrieval/retriever.py,sha256=idJcz4pUHgPHuZvX3VIUmO8Vs-iw4_1w6UUypynRfVY,4579
+openadapt_ml/runtime/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+openadapt_ml/runtime/policy.py,sha256=M-OyhaE5gWh7e3KJ0Ip_YavlfRaEO7S4UKNjvqH_qsg,6724
+openadapt_ml/schema/__init__.py,sha256=W1Rx58WjFpUE2D1hdujD6tkxr5m8U3nE4JBMwdw7kLc,2787
+openadapt_ml/schema/converters.py,sha256=3qK1z8EATFaDi1M0w9T7PLiRtJu6OtQM7JG7qbE4EKU,18392
+openadapt_ml/schema/episode.py,sha256=_QQ34V39DYLaOx5GnH4mKHoXteekqRSq1C2aJe_Y_5Y,15399
+openadapt_ml/scripts/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+openadapt_ml/scripts/compare.py,sha256=IUi3lQfOV8qm1JMvPDAnk3UqT7vPYNjLGeN8qda1wXc,56921
+openadapt_ml/scripts/demo_policy.py,sha256=NqB1akCKT5dlLER5ToBpF15QYaWnbXtRgWBvrOb6HJc,2105
+openadapt_ml/scripts/eval_policy.py,sha256=Tk1QMNlkfcyHToAqdTgdLkJyEGyVLXMcE5EVmJb89Ng,10381
+openadapt_ml/scripts/make_gif.py,sha256=H9fevBZFH31_7vs-OROfg9A2U6eboWXUNjyU2-XLMqw,4439
+openadapt_ml/scripts/prepare_synthetic.py,sha256=RbnY3QiH_RFk2du3awXFn90jxJQNUkhWlei2NSQ1aUs,1109
+openadapt_ml/scripts/run_qwen_login_benchmark.py,sha256=NWIhCAFSX5pYKFRCec7RkrYtzvz2LNMqhDfXcKxlagM,5655
+openadapt_ml/scripts/train.py,sha256=BrDJxiZx1S8igpNi6hC287kh51qU5RtPcdVq1yxEJkQ,6685
+openadapt_ml/training/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+openadapt_ml/training/benchmark_viewer.py,sha256=iD56K467n0qS-nrcxxGZtABQs6qswB2je5Uj4xEacNI,174238
+openadapt_ml/training/shared_ui.py,sha256=7ZdBDo-__pHUMiuvSot-waYvpinAullRXQTsiPoruBw,4823
+openadapt_ml/training/stub_provider.py,sha256=fw2b0EzsCDDFhTZQIKR49BEFmSEN1d4g9rhFihzErWY,10508
+openadapt_ml/training/trainer.py,sha256=K4LMDe-GxnGZB9gt0IGMU67DcqtcggaaQIbJp4GYcYc,90713
+openadapt_ml/training/trl_trainer.py,sha256=9NQnNt2MDLUUUA_V8f2PwsDLeHnb2fOnmxol24-PyjU,11355
+openadapt_ml/training/viewer.py,sha256=7uA2SHW1Uh7v65s8sOSaYuAVUlyp6GsCOGy6YA0hnBQ,121132
+openadapt_ml-0.2.0.dist-info/METADATA,sha256=-pGXAMvRsWMHCdjsj3GzALIcAr0fPaMvEaUv8jkGYfA,28839
+openadapt_ml-0.2.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+openadapt_ml-0.2.0.dist-info/licenses/LICENSE,sha256=2E5UY67RVLedJuNnwGudkAMtfM3LZNUcHgmaL89TAfw,1068
+openadapt_ml-0.2.0.dist-info/RECORD,,

openadapt_ml/schemas/__init__.py

@@ -1,53 +0,0 @@
-"""Schema definitions and validation for openadapt-ml.
-
-Core data structures:
-    - Action: A single GUI action (click, type, scroll, etc.)
-    - Observation: GUI state observation (screenshot, accessibility tree, etc.)
-    - Step: One timestep containing observation + action
-    - Episode: A single task attempt / workflow instance
-    - Session: Container for multiple episodes
-
-Validation:
-    - validate_episode(): Validate an Episode object
-    - validate_session(): Validate a Session object
-    - validate_episodes(): Validate a list of Episodes
-    - ValidationError: Raised on schema violations
-"""
-
-from openadapt_ml.schemas.sessions import (
-    Action,
-    ActionType,
-    Episode,
-    Observation,
-    Session,
-    Step,
-)
-from openadapt_ml.schemas.validation import (
-    ValidationError,
-    summarize_episodes,
-    validate_action,
-    validate_episode,
-    validate_episodes,
-    validate_observation,
-    validate_session,
-    validate_step,
-)
-
-__all__ = [
-    # Core types
-    "Action",
-    "ActionType",
-    "Episode",
-    "Observation",
-    "Session",
-    "Step",
-    # Validation
-    "ValidationError",
-    "validate_action",
-    "validate_episode",
-    "validate_episodes",
-    "validate_observation",
-    "validate_session",
-    "validate_step",
-    "summarize_episodes",
-]