diff --git a/.cursor/rules/coderule.mdc b/.cursor/rules/coderule.mdc
index e7fe6f0..c333ec6 100644
--- a/.cursor/rules/coderule.mdc
+++ b/.cursor/rules/coderule.mdc
@@ -11,8 +11,11 @@ alwaysApply: true
 - Write code that takes into account the different environments: development, production
 - You are careful to make changes that are requested or you are confident the changes are well understood and related to the change being requested
 - Mocking data is needed only for tests, never mock data for dev or prod env
+- Make test environment (files, db and so on) as close as possible to the production environment
 - When you add new libraries or dependencies make sure you are using the same version of it as other parts of the code
 - When a test fails due to a missing dependency, install it — do not fake or stub the module system. For normal packages, add them to the project's dependency file (requirements-test.txt, package.json devDependencies, test csproj, etc.) and install. Only consider stubbing if the dependency is heavy (e.g. hardware-specific SDK, large native toolchain) — and even then, ask the user first before choosing to stub.
+- Do not solve environment or infrastructure problems (dependency resolution, import paths, service discovery, connection config) by hardcoding workarounds in source code. Fix them at the environment/configuration level.
+- Before writing new infrastructure or workaround code, check how the existing codebase already handles the same concern. Follow established project patterns.
 
 - Focus on the areas of code relevant to the task
 - Do not touch code that is unrelated to the task
diff --git a/.cursor/rules/cursor-meta.mdc b/.cursor/rules/cursor-meta.mdc
index 94cc6c5..8cc663a 100644
--- a/.cursor/rules/cursor-meta.mdc
+++ b/.cursor/rules/cursor-meta.mdc
@@ -17,11 +17,5 @@ globs: [".cursor/**"]
 ## Agent Files (.cursor/agents/)
 - Must have `name` and `description` in frontmatter
 
-## User Interaction
-- Use the AskQuestion tool for structured choices (A/B/C/D) when available — it provides an interactive UI. Fall back to plain-text questions if the tool is unavailable.
-
-## Execution Safety
-- Never run test suites, builds, Docker commands, or other long-running/resource-heavy/security-risky operations without asking the user first - unlsess it is explicilty stated in skill or agent, or user already asked to do so.
-
 ## Security
 - All `.cursor/` files must be scanned for hidden Unicode before committing (see cursor-security.mdc)
diff --git a/.cursor/rules/meta-rule.mdc b/.cursor/rules/meta-rule.mdc
new file mode 100644
index 0000000..b7348e5
--- /dev/null
+++ b/.cursor/rules/meta-rule.mdc
@@ -0,0 +1,30 @@
+---
+description: "Execution safety, user interaction, and self-improvement protocols for the AI agent"
+alwaysApply: true
+---
+# Agent Meta Rules
+
+## Execution Safety
+- Never run test suites, builds, Docker commands, or other long-running/resource-heavy/security-risky operations without asking the user first — unless it is explicitly stated in a skill or agent, or the user already asked to do so.
+
+## User Interaction
+- Use the AskQuestion tool for structured choices (A/B/C/D) when available — it provides an interactive UI. Fall back to plain-text questions if the tool is unavailable.
+
+## Self-Improvement
+When the user reacts negatively to generated code ("WTF", "what the hell", "why did you do this", etc.):
+
+1. **Pause** — do not rush to fix. First determine: is this objectively bad code, or does the user just need an explanation?
+2. **If the user doesn't understand** — explain the reasoning. That's it. No code change needed.
+3. **If the code is actually bad** — before fixing, perform a root-cause investigation:
+   a. **Why** did this bad code get produced? Identify the reasoning chain or implicit assumption that led to it.
+   b. **Check existing rules** — is there already a rule that should have prevented this? If so, clarify or strengthen it.
+   c. **Propose a new rule** if no existing rule covers the failure mode. Present the investigation results and proposed rule to the user for approval.
+   d. **Only then** fix the code.
+4. The rule goes into `coderule.mdc` for coding practices, `meta-rule.mdc` for agent behavior, or a new focused rule file — depending on context. Always check for duplicates or near-duplicates first.
+
+### Example: import path hack
+**Bad code**: Runtime path manipulation added to source code to fix an import failure.
+**Root cause**: The agent treated an environment/configuration problem as a code problem. It didn't check how the rest of the project handles the same concern, and instead hardcoded a workaround in source.
+**Preventive rules added to coderule.mdc**:
+- "Do not solve environment or infrastructure problems by hardcoding workarounds in source code. Fix them at the environment/configuration level."
+- "Before writing new infrastructure or workaround code, check how the existing codebase already handles the same concern. Follow established project patterns."
diff --git a/.cursor/rules/tracker.mdc b/.cursor/rules/tracker.mdc
new file mode 100644
index 0000000..285333c
--- /dev/null
+++ b/.cursor/rules/tracker.mdc
@@ -0,0 +1,9 @@
+# Work Item Tracker
+
+- Use **Jira** as the sole work item tracker (MCP server: `user-Jira-MCP-Server`)
+- Do NOT use Azure DevOps for work item management
+- Jira cloud ID: `denyspopov.atlassian.net`
+- Project key: `AZ`
+- Project name: AZAION
+- All task IDs follow the format `AZ-<number>`
+- Issue types: Epic, Story, Task, Bug, Subtask
diff --git a/.cursor/skills/autopilot/state.md b/.cursor/skills/autopilot/state.md
index 022ecda..33dd76f 100644
--- a/.cursor/skills/autopilot/state.md
+++ b/.cursor/skills/autopilot/state.md
@@ -41,7 +41,7 @@ retry_count: 3
 ### State File Rules
 
 1. **Create** on the first autopilot invocation (after state detection determines Step 1)
-2. **Update** after every step completion, session boundary, or failed retry
+2. **Update** after every change — this includes: batch completion, sub-step progress, step completion, session boundary, failed retry, or any meaningful state transition. The state file must always reflect the current reality.
 3. **Read** as the first action on every invocation — before folder scanning
 4. **Cross-check**: verify against actual `_docs/` folder contents. If they disagree, trust the folder structure and update the state file
 5. **Never delete** the state file
diff --git a/.cursor/skills/implement/SKILL.md b/.cursor/skills/implement/SKILL.md
index 1039d01..e6e62a2 100644
--- a/.cursor/skills/implement/SKILL.md
+++ b/.cursor/skills/implement/SKILL.md
@@ -128,31 +128,31 @@ Auto-fix loop with bounded retries (max 2 attempts) before escalating to user:
 
 Track `auto_fix_attempts` count in the batch report for retrospective analysis.
 
-### 10. Test
-
-- Read and execute `.cursor/skills/test-run/SKILL.md` (detect runner, run suite, diagnose failures, present blocking choices)
-- Test failures are a **blocking gate** — do not proceed to commit until the test-run skill completes with a user decision
-- Note: the autopilot also runs a separate full test suite after all implementation batches complete (greenfield Step 7, existing-code Steps 6/10). This is intentional — per-batch tests are regression checks, the post-implement run is final validation.
-
-### 11. Commit and Push
+### 10. Commit and Push
 
 - After user confirms the batch (explicitly for FAIL, implicitly for PASS/PASS_WITH_WARNINGS):
   - `git add` all changed files from the batch
   - `git commit` with a message that includes ALL task IDs (tracker IDs or numeric prefixes) of tasks implemented in the batch, followed by a summary of what was implemented. Format: `[TASK-ID-1] [TASK-ID-2] ... Summary of changes`
   - `git push` to the remote branch
 
-### 12. Update Tracker Status → In Testing
+### 11. Update Tracker Status → In Testing
 
 After the batch is committed and pushed, transition the ticket status of each task in the batch to **In Testing** via the configured work item tracker. If `tracker: local`, skip this step.
 
-### 13. Archive Completed Tasks
+### 12. Archive Completed Tasks
 
 Move each completed task file from `TASKS_DIR/todo/` to `TASKS_DIR/done/`.
 
-### 14. Loop
+### 13. Loop
 
 - Go back to step 2 until all tasks in `todo/` are done
-- When all tasks are complete, report final summary
+
+### 14. Final Test Run
+
+- After all batches are complete, run the full test suite once
+- Read and execute `.cursor/skills/test-run/SKILL.md` (detect runner, run suite, diagnose failures, present blocking choices)
+- Test failures are a **blocking gate** — do not proceed until the test-run skill completes with a user decision
+- When tests pass, report final summary
 
 ## Batch Report Persistence
 
@@ -195,7 +195,7 @@ After each batch, produce a structured report:
 | Implementer fails same approach 3+ times | Stop it, escalate to user |
 | Task blocked on external dependency (not in task list) | Report and skip |
 | File ownership conflict unresolvable | ASK user |
-| Any test failure after a batch | Delegate to test-run skill — blocking gate |
+| Test failure after final test run | Delegate to test-run skill — blocking gate |
 | All tasks complete | Report final summary, suggest final commit |
 | `_dependencies_table.md` missing | STOP — run `/decompose` first |
 
@@ -203,7 +203,7 @@ After each batch, produce a structured report:
 
 Each batch commit serves as a rollback checkpoint. If recovery is needed:
 
-- **Tests fail after a batch commit**: `git revert <batch-commit-hash>` using the hash from the batch report in `_docs/03_implementation/`
+- **Tests fail after final test run**: `git revert <batch-commit-hash>` using hashes from the batch reports in `_docs/03_implementation/`
 - **Resuming after interruption**: Read `_docs/03_implementation/batch_*_report.md` files to determine which batches completed, then continue from the next batch
 - **Multiple consecutive batches fail**: Stop and escalate to user with links to batch reports and commit hashes
 
@@ -212,4 +212,4 @@ Each batch commit serves as a rollback checkpoint. If recovery is needed:
 - Never launch tasks whose dependencies are not yet completed
 - Never allow two parallel agents to write to the same file
 - If a subagent fails or is flagged as stuck, stop it and report — do not let it loop indefinitely
-- Always run tests after each batch completes
+- Always run the full test suite after all batches complete (step 14)
diff --git a/.gitignore b/.gitignore
index 02f0c7a..717cf1f 100644
--- a/.gitignore
+++ b/.gitignore
@@ -24,4 +24,8 @@ venv
 *.png
 
 # Test results
-tests/test-results/
\ No newline at end of file
+tests/test-results/
+
+# Test fixture data — override global ignores for committed test data
+!tests/root/
+!tests/root/**
\ No newline at end of file
diff --git a/_docs/02_document/tests/blackbox-tests.md b/_docs/02_document/tests/blackbox-tests.md
index 6a78f16..172a8bf 100644
--- a/_docs/02_document/tests/blackbox-tests.md
+++ b/_docs/02_document/tests/blackbox-tests.md
@@ -1,61 +1,9 @@
 # Blackbox Test Scenarios
 
-## BT-AUG: Augmentation Pipeline
-
-### BT-AUG-01: Single image produces 8 outputs
-- **Input**: 1 image + 1 valid label from fixture dataset
-- **Action**: Run `Augmentator.augment_inner()` on the image
-- **Expected**: Returns list of exactly 8 ImageLabel objects
-- **Traces**: AC: 8× augmentation ratio
-
-### BT-AUG-02: Augmented filenames follow naming convention
-- **Input**: Image with stem "test_image"
-- **Action**: Run `augment_inner()`
-- **Expected**: Output filenames: `test_image.jpg`, `test_image_1.jpg` through `test_image_7.jpg`; matching `.txt` labels
-- **Traces**: AC: Augmentation output format
-
-### BT-AUG-03: All output bounding boxes in valid range
-- **Input**: 1 image + label with multiple bboxes
-- **Action**: Run `augment_inner()`
-- **Expected**: Every bbox coordinate in every output label is in [0.0, 1.0]
-- **Traces**: AC: Bounding boxes clipped to [0, 1]
-
-### BT-AUG-04: Bounding box correction clips edge bboxes
-- **Input**: Label with bbox near edge: `0 0.99 0.5 0.2 0.1`
-- **Action**: Run `correct_bboxes()`
-- **Expected**: Width reduced so bbox fits within [margin, 1-margin]; no coordinate exceeds bounds
-- **Traces**: AC: Bounding boxes clipped to [0, 1]
-
-### BT-AUG-05: Tiny bounding boxes removed after correction
-- **Input**: Label with tiny bbox that becomes < 0.01 after clipping
-- **Action**: Run `correct_bboxes()`
-- **Expected**: Bbox removed from output (area < correct_min_bbox_size)
-- **Traces**: AC: Bounding boxes with area < 0.01% discarded
-
-### BT-AUG-06: Empty label produces 8 outputs with empty labels
-- **Input**: 1 image + empty label file
-- **Action**: Run `augment_inner()`
-- **Expected**: 8 ImageLabel objects returned; all have empty labels lists
-- **Traces**: AC: Augmentation handles empty annotations
-
-### BT-AUG-07: Full augmentation pipeline (filesystem integration)
-- **Input**: 5 images + labels copied to data/ directory in tmp_path
-- **Action**: Run `augment_annotations()` with patched paths
-- **Expected**: 40 images (5 × 8) in processed images dir; 40 matching labels in processed labels dir
-- **Traces**: AC: 8× augmentation, filesystem output
-
-### BT-AUG-08: Augmentation skips already-processed images
-- **Input**: 5 images in data/; 3 already present in processed/ dir
-- **Action**: Run `augment_annotations()`
-- **Expected**: Only 2 new images processed (16 new outputs); existing 3 untouched
-- **Traces**: AC: Augmentation processes only unprocessed images
-
----
-
 ## BT-DSF: Dataset Formation
 
 ### BT-DSF-01: 70/20/10 split ratio
-- **Input**: 100 images + labels in processed/ dir
+- **Input**: 100 images + labels in data/ dir
 - **Action**: Run `form_dataset()` with patched paths
 - **Expected**: train: 70 images+labels, valid: 20, test: 10
 - **Traces**: AC: Dataset split 70/20/10
diff --git a/_docs/02_document/tests/performance-tests.md b/_docs/02_document/tests/performance-tests.md
index 9c75a48..cef73f3 100644
--- a/_docs/02_document/tests/performance-tests.md
+++ b/_docs/02_document/tests/performance-tests.md
@@ -1,18 +1,5 @@
 # Performance Test Scenarios
 
-## PT-AUG-01: Augmentation throughput
-- **Input**: 10 images from fixture dataset
-- **Action**: Run `augment_annotations()`, measure wall time
-- **Expected**: Completes within 60 seconds (10 images × 8 outputs = 80 files)
-- **Traces**: Restriction: Augmentation runs continuously
-- **Note**: Threshold is generous; actual performance depends on CPU
-
-## PT-AUG-02: Parallel augmentation speedup
-- **Input**: 10 images from fixture dataset
-- **Action**: Run with ThreadPoolExecutor vs sequential, compare times
-- **Expected**: Parallel is ≥ 1.5× faster than sequential
-- **Traces**: AC: Parallelized per-image processing
-
 ## PT-DSF-01: Dataset formation throughput
 - **Input**: 100 images + labels
 - **Action**: Run `form_dataset()`, measure wall time
diff --git a/_docs/02_document/tests/resilience-tests.md b/_docs/02_document/tests/resilience-tests.md
index d36d555..65343ee 100644
--- a/_docs/02_document/tests/resilience-tests.md
+++ b/_docs/02_document/tests/resilience-tests.md
@@ -1,25 +1,7 @@
 # Resilience Test Scenarios
 
-## RT-AUG-01: Augmentation handles corrupted image gracefully
-- **Input**: 1 valid image + 1 corrupted image file (truncated JPEG) in data/ dir
-- **Action**: Run `augment_annotations()`
-- **Expected**: Valid image produces 8 outputs; corrupted image skipped without crashing pipeline; total output: 8 files
-- **Traces**: Restriction: Augmentation exception handling per-image
-
-## RT-AUG-02: Augmentation handles missing label file
-- **Input**: 1 image with no matching label file
-- **Action**: Run `augment_annotation()` on the image
-- **Expected**: Exception caught per-thread; does not crash pipeline
-- **Traces**: Restriction: Augmentation exception handling
-
-## RT-AUG-03: Augmentation transform failure produces fewer variants
-- **Input**: 1 image + label that causes some transforms to fail (extremely narrow bbox)
-- **Action**: Run `augment_inner()`
-- **Expected**: Returns 1-8 ImageLabel objects (original always present; failed variants skipped); no crash
-- **Traces**: Restriction: Transform failure handling
-
-## RT-DSF-01: Dataset formation with empty processed directory
-- **Input**: Empty processed images dir
+## RT-DSF-01: Dataset formation with empty data directory
+- **Input**: Empty data images dir
 - **Action**: Run `form_dataset()`
 - **Expected**: Creates empty train/valid/test directories; no crash
 - **Traces**: Restriction: Edge case handling
diff --git a/_docs/02_document/tests/resource-limit-tests.md b/_docs/02_document/tests/resource-limit-tests.md
index f167a6e..e4624d0 100644
--- a/_docs/02_document/tests/resource-limit-tests.md
+++ b/_docs/02_document/tests/resource-limit-tests.md
@@ -1,11 +1,5 @@
 # Resource Limit Test Scenarios
 
-## RL-AUG-01: Augmentation output count bounded
-- **Input**: 1 image
-- **Action**: Run `augment_inner()`
-- **Expected**: Returns exactly 8 outputs (never more, even with retries)
-- **Traces**: AC: 8× augmentation ratio (1 original + 7 augmented)
-
 ## RL-DSF-01: Dataset split ratios sum to 100%
 - **Input**: Any number of images
 - **Action**: Check `train_set + valid_set + test_set`
diff --git a/_docs/02_document/tests/test-data.md b/_docs/02_document/tests/test-data.md
index 392f07e..b98f79b 100644
--- a/_docs/02_document/tests/test-data.md
+++ b/_docs/02_document/tests/test-data.md
@@ -4,8 +4,8 @@
 
 | ID | Data Item | Source | Format | Preparation |
 |----|-----------|--------|--------|-------------|
-| FD-01 | Annotated images (100) | `_docs/00_problem/input_data/dataset/images/` | JPEG | Copy subset to tmp_path at test start |
-| FD-02 | YOLO labels (100) | `_docs/00_problem/input_data/dataset/labels/` | TXT | Copy subset to tmp_path at test start |
+| FD-01 | Annotated images (20) | `tests/data/images/` | JPEG | Copy subset to tmp_path at test start |
+| FD-02 | YOLO labels (20) | `tests/data/labels/` | TXT | Copy subset to tmp_path at test start |
 | FD-03 | ONNX model | `_docs/00_problem/input_data/azaion.onnx` | ONNX | Read bytes at test start |
 | FD-04 | Class definitions | `classes.json` (project root) | JSON | Copy to tmp_path at test start |
 | FD-05 | Corrupted labels | Generated at test time | TXT | Create labels with coords > 1.0 |
diff --git a/_docs/02_document/tests/traceability-matrix.md b/_docs/02_document/tests/traceability-matrix.md
index 63ce3d1..de8a381 100644
--- a/_docs/02_document/tests/traceability-matrix.md
+++ b/_docs/02_document/tests/traceability-matrix.md
@@ -4,15 +4,6 @@
 
 | AC / Restriction | Test IDs | Coverage |
 |------------------|----------|----------|
-| 8× augmentation ratio | BT-AUG-01, BT-AUG-06, BT-AUG-07, RL-AUG-01 | Full |
-| Augmentation naming convention | BT-AUG-02 | Full |
-| Bounding boxes clipped to [0,1] | BT-AUG-03, BT-AUG-04 | Full |
-| Tiny bboxes (< 0.01) discarded | BT-AUG-05 | Full |
-| Augmentation skips already-processed | BT-AUG-08 | Full |
-| Augmentation parallelized | PT-AUG-02 | Full |
-| Augmentation handles corrupted images | RT-AUG-01 | Full |
-| Augmentation handles missing labels | RT-AUG-02 | Full |
-| Transform failure graceful | RT-AUG-03 | Full |
 | Dataset split 70/20/10 | BT-DSF-01, RL-DSF-01 | Full |
 | Dataset directory structure | BT-DSF-02 | Full |
 | Dataset integrity (no data loss) | BT-DSF-03, RL-DSF-02 | Full |
@@ -34,6 +25,17 @@
 | Static model encryption key | ST-ENC-03 | Full |
 | Random IV per encryption | ST-ENC-01 | Full |
 
+## Removed (augmentation now built into YOLO training)
+
+The following tests were removed because external augmentation (`augmentation.py`) is no longer part of the training pipeline. YOLO's built-in augmentation replaces it.
+
+| Removed Test IDs | Reason |
+|-------------------|--------|
+| BT-AUG-01 to BT-AUG-08 | External augmentation replaced by YOLO built-in |
+| PT-AUG-01, PT-AUG-02 | Augmentation performance no longer relevant |
+| RT-AUG-01 to RT-AUG-03 | Augmentation resilience no longer relevant |
+| RL-AUG-01 | Augmentation resource limits no longer relevant |
+
 ## Uncovered (Require External Services)
 
 | AC / Restriction | Reason |
@@ -50,18 +52,18 @@
 
 | Metric | Value |
 |--------|-------|
-| Total AC + Restrictions | 36 |
-| Covered by tests | 29 |
+| Total AC + Restrictions | 27 |
+| Covered by tests | 20 |
 | Uncovered (external deps) | 7 |
-| **Coverage** | **80.6%** |
+| **Coverage** | **74.1%** |
 
 ## Test Count Summary
 
 | Category | Count |
 |----------|-------|
-| Blackbox tests | 32 |
-| Performance tests | 5 |
-| Resilience tests | 6 |
+| Blackbox tests | 21 |
+| Performance tests | 3 |
+| Resilience tests | 3 |
 | Security tests | 7 |
-| Resource limit tests | 5 |
-| **Total** | **55** |
+| Resource limit tests | 4 |
+| **Total** | **38** |
diff --git a/_docs/02_tasks/_dependencies_table.md b/_docs/02_tasks/_dependencies_table.md
index 210aa4c..e8fea72 100644
--- a/_docs/02_tasks/_dependencies_table.md
+++ b/_docs/02_tasks/_dependencies_table.md
@@ -38,8 +38,36 @@ AZ-151 (Epic: Blackbox Tests)
     └── AZ-163 test_annotation_queue
 ```
 
-## Implementation Strategy
+---
 
-- **Batch 1**: AZ-152 (test infrastructure) — must be implemented first
-- **Batch 2**: AZ-153 to AZ-163 (all test tasks) — can be implemented in parallel after infrastructure is ready
-- **Estimated batches**: 2
+## Refactoring Tasks (Epic: AZ-164)
+
+**Date**: 2026-03-28
+**Total Tasks**: 5
+**Total Complexity Points**: 13
+
+| Task | Name | Complexity | Dependencies | Epic |
+|------|------|-----------|-------------|------|
+| AZ-165 | refactor_unify_config | 3 | None | AZ-164 |
+| AZ-166 | refactor_yolo_model | 2 | None | AZ-164 |
+| AZ-167 | refactor_builtin_augmentation | 3 | AZ-166 | AZ-164 |
+| AZ-168 | refactor_remove_processed_dir | 3 | AZ-167 | AZ-164 |
+| AZ-169 | refactor_hard_symlinks | 2 | AZ-168 | AZ-164 |
+
+### Dependency Graph
+
+```
+AZ-164 (Epic: Code Improvements Refactoring)
+├── AZ-165 refactor_unify_config (independent)
+└── AZ-166 refactor_yolo_model
+    └── AZ-167 refactor_builtin_augmentation
+        └── AZ-168 refactor_remove_processed_dir
+            └── AZ-169 refactor_hard_symlinks
+```
+
+### Implementation Strategy
+
+- **Batch 1**: AZ-165 (unify config) + AZ-166 (YOLO model) — independent, can be parallel
+- **Batch 2**: AZ-167 (built-in aug) + AZ-168 (remove processed dir) — sequential chain
+- **Batch 3**: AZ-169 (hard symlinks) — depends on batch 2
+- **Estimated batches**: 3
diff --git a/_docs/02_tasks/todo/AZ-165_refactor_unify_config.md b/_docs/02_tasks/todo/AZ-165_refactor_unify_config.md
new file mode 100644
index 0000000..00c36fd
--- /dev/null
+++ b/_docs/02_tasks/todo/AZ-165_refactor_unify_config.md
@@ -0,0 +1,54 @@
+# Unify Configuration
+
+**Task**: AZ-165_refactor_unify_config
+**Name**: Unify configuration — remove annotation-queue/config.yaml
+**Description**: Consolidate two config files into one shared Config model
+**Complexity**: 3 points
+**Dependencies**: None
+**Component**: Configuration
+**Tracker**: AZ-165
+**Epic**: AZ-164
+
+## Problem
+
+Two separate `config.yaml` files exist (root and `src/annotation-queue/`) with overlapping content but different `dirs` values. The annotation queue handler parses YAML manually instead of using the shared `Config` Pydantic model, creating drift risk.
+
+## Outcome
+
+- Single `Config` model in `constants.py` covers all configuration including queue settings
+- `annotation_queue_handler.py` uses the shared `Config` instead of parsing its own YAML
+- `src/annotation-queue/config.yaml` is deleted
+
+## Scope
+
+### Included
+- Add Pydantic models for `ApiConfig`, `QueueConfig`; extend `DirsConfig` with all directory fields (data, data_seed, data_processed, data_deleted, images, labels)
+- Add these to the `Config` Pydantic model in `constants.py`
+- Refactor `annotation_queue_handler.py` constructor to accept/import the shared Pydantic `Config`
+- Delete `src/annotation-queue/config.yaml`
+
+### Excluded
+- Changing queue connection logic or message handling
+- Modifying root `config.yaml` structure beyond adding queue section (it already has it)
+
+## Acceptance Criteria
+
+**AC-1: Single config source**
+Given the root `config.yaml` contains queue and dirs settings
+When `annotation_queue_handler.py` initializes
+Then it reads configuration from the shared `Config` model, not a local YAML file
+
+**AC-2: No duplicate config file**
+Given the refactoring is complete
+When listing `src/annotation-queue/`
+Then `config.yaml` does not exist
+
+**AC-3: Annotation queue behavior preserved**
+Given the unified configuration
+When the annotation queue handler processes messages
+Then it uses the correct directory paths from configuration
+
+## Constraints
+
+- Root `config.yaml` already has the `queue` section — reuse it
+- `annotation_queue_handler.py` runs as a separate process — config import path must work from its working directory
diff --git a/_docs/02_tasks/todo/AZ-166_refactor_yolo_model.md b/_docs/02_tasks/todo/AZ-166_refactor_yolo_model.md
new file mode 100644
index 0000000..e7cbcf2
--- /dev/null
+++ b/_docs/02_tasks/todo/AZ-166_refactor_yolo_model.md
@@ -0,0 +1,56 @@
+# Update YOLO Model
+
+**Task**: AZ-166_refactor_yolo_model
+**Name**: Update YOLO model to 26m variant (supports both from-scratch and pretrained)
+**Description**: Update model references from YOLO11m to YOLO26m; support both training from scratch (`.yaml`) and from pretrained weights (`.pt`)
+**Complexity**: 2 points
+**Dependencies**: None
+**Component**: Training Pipeline
+**Tracker**: AZ-166
+**Epic**: AZ-164
+
+## Problem
+
+Current `TrainingConfig.model` is set to `yolo11m.yaml` which defines a YOLO11 architecture. YOLO26m is the latest model variant. The system should support both training modes:
+1. **From scratch** — using `yolo26m.yaml` (architecture definition, trains from random weights)
+2. **From pretrained** — using `yolo26m.pt` (pretrained weights, faster convergence)
+
+## Outcome
+
+- `TrainingConfig` default model updated to `yolo26m.pt` (pretrained, recommended default)
+- `config.yaml` updated to `yolo26m.pt`
+- Both `yolo26m.pt` and `yolo26m.yaml` work when set in `config.yaml`
+- `train_dataset()` and `resume_training()` work with either model reference
+
+## Scope
+
+### Included
+- Update `TrainingConfig.model` default from `yolo11m.yaml` to `yolo26m.pt`
+- Update `config.yaml` training.model from `yolo11m.yaml` to `yolo26m.pt`
+- Verify `train_dataset()` works with both `.pt` and `.yaml` model values
+
+### Excluded
+- Changing training hyperparameters (epochs, batch, imgsz)
+- Updating ultralytics library version
+
+## Acceptance Criteria
+
+**AC-1: Default model config updated**
+Given the training configuration
+When reading `TrainingConfig.model`
+Then the default value is `yolo26m.pt`
+
+**AC-2: config.yaml updated**
+Given the root `config.yaml`
+When reading `training.model`
+Then the value is `yolo26m.pt`
+
+**AC-3: From-scratch training supported**
+Given `config.yaml` sets `training.model: yolo26m.yaml`
+When `YOLO(constants.config.training.model)` is called
+Then a YOLO26m model is built from the architecture definition
+
+**AC-4: Pretrained training supported**
+Given `config.yaml` sets `training.model: yolo26m.pt`
+When `YOLO(constants.config.training.model)` is called
+Then a YOLO26m model is loaded from pretrained weights
diff --git a/_docs/02_tasks/todo/AZ-167_refactor_builtin_augmentation.md b/_docs/02_tasks/todo/AZ-167_refactor_builtin_augmentation.md
new file mode 100644
index 0000000..166e653
--- /dev/null
+++ b/_docs/02_tasks/todo/AZ-167_refactor_builtin_augmentation.md
@@ -0,0 +1,55 @@
+# Replace External Augmentation with YOLO Built-in
+
+**Task**: AZ-167_refactor_builtin_augmentation
+**Name**: Replace external augmentation with YOLO built-in
+**Description**: Remove albumentations pipeline and use YOLO model.train() built-in augmentation parameters
+**Complexity**: 3 points
+**Dependencies**: AZ-166_refactor_yolo_model
+**Component**: Training Pipeline
+**Tracker**: AZ-167
+**Epic**: AZ-164
+
+## Problem
+
+`augmentation.py` uses the `albumentations` library to augment images into a `processed_dir` before training. This creates a separate processing step, uses extra disk space (8x original), and adds complexity. YOLO's built-in augmentation applies on-the-fly during training.
+
+## Outcome
+
+- `train_dataset()` passes augmentation parameters directly to `model.train()`
+- Each augmentation parameter is on its own line with a descriptive comment
+- The external augmentation step is removed from the training pipeline
+- `augmentation.py` is no longer called during training
+
+## Scope
+
+### Included
+- Add YOLO built-in augmentation parameters to `model.train()` call in `train_dataset()`
+- Parameters to add: hsv_h, hsv_s, hsv_v, degrees, translate, scale, shear, fliplr, mosaic (each with comment)
+- Remove call to augmentation from training flow
+
+### Excluded
+- Deleting `augmentation.py` file (may still be useful standalone)
+- Changing training hyperparameters unrelated to augmentation
+
+## Acceptance Criteria
+
+**AC-1: Built-in augmentation parameters with comments**
+Given the `train_dataset()` function
+When `model.train()` is called
+Then every parameter (including augmentation: hsv_h, hsv_s, hsv_v, degrees, scale, shear, fliplr, mosaic, and training: data, epochs, batch, imgsz, etc.) is on its own line with an inline comment explaining what the parameter controls
+
+**AC-2: No external augmentation in training flow**
+Given the training pipeline
+When `train_dataset()` runs
+Then it does not call `augment_annotations()` or any albumentations-based augmentation
+
+## Constraints
+
+- Every parameter row in the `model.train()` call MUST have an inline comment describing what it does (e.g. `hsv_h=0.015,  # hue shift fraction of the color wheel`)
+- This applies to ALL parameters, not just augmentation — training params (data, epochs, batch, imgsz, save_period, workers) also need comments
+- Augmentation parameter values should approximate the current albumentations settings:
+  - fliplr=0.6 (was HorizontalFlip p=0.6)
+  - degrees=35.0 (was Affine rotate=(-35,35))
+  - shear=10.0 (was Affine shear=(-10,10))
+  - hsv_h=0.015, hsv_s=0.7, hsv_v=0.4 (approximate HSV shifts)
+  - mosaic=1.0 (new YOLO built-in, recommended default)
diff --git a/_docs/02_tasks/todo/AZ-168_refactor_remove_processed_dir.md b/_docs/02_tasks/todo/AZ-168_refactor_remove_processed_dir.md
new file mode 100644
index 0000000..3045f4d
--- /dev/null
+++ b/_docs/02_tasks/todo/AZ-168_refactor_remove_processed_dir.md
@@ -0,0 +1,60 @@
+# Remove Processed Directory
+
+**Task**: AZ-168_refactor_remove_processed_dir
+**Name**: Remove processed directory — use data dir directly
+**Description**: Eliminate processed_dir concept from Config and all consumers; read from data dir directly; update e2e test fixture
+**Complexity**: 3 points
+**Dependencies**: AZ-167_refactor_builtin_augmentation
+**Component**: Training Pipeline, Data Utilities
+**Tracker**: AZ-168
+**Epic**: AZ-164
+
+## Problem
+
+`Config` exposes `processed_dir`, `processed_images_dir`, `processed_labels_dir` properties. Multiple files read from the processed directory: `train.py::form_dataset()`, `exports.py::form_data_sample()`, `dataset-visualiser.py::visualise_processed_folder()`. With built-in augmentation, the processed directory is no longer populated.
+
+The e2e test fixture (`tests/test_training_e2e.py`) currently copies images to both `data_images_dir` and `processed_images_dir` as a workaround — this needs cleanup once `form_dataset()` reads from data dirs.
+
+## Outcome
+
+- `Config` no longer has `processed_dir`/`processed_images_dir`/`processed_labels_dir` properties
+- `form_dataset()` reads images/labels from `data_images_dir`/`data_labels_dir`
+- `form_data_sample()` reads from `data_images_dir`
+- `visualise_processed_folder()` reads from `data_images_dir`/`data_labels_dir`
+- E2e test fixture copies images only to `data_images_dir`/`data_labels_dir` (no more processed dir population)
+
+## Scope
+
+### Included
+- Remove `processed_dir`, `processed_images_dir`, `processed_labels_dir` from `Config`
+- Update `form_dataset()` in `train.py` to use `data_images_dir` and `data_labels_dir`
+- Update `copy_annotations()` in `train.py` to look up labels from `data_labels_dir` instead of `processed_labels_dir`
+- Update `form_data_sample()` in `exports.py` to use `data_images_dir`
+- Update `visualise_processed_folder()` in `dataset-visualiser.py`
+- Update `tests/test_training_e2e.py` e2e fixture: remove processed dir population (only copy to data dirs)
+
+### Excluded
+- Removing `augmentation.py` file
+- Changing `corrupted_dir` handling
+
+## Acceptance Criteria
+
+**AC-1: No processed dir in Config**
+Given the `Config` class
+When inspecting its properties
+Then `processed_dir`, `processed_images_dir`, `processed_labels_dir` do not exist
+
+**AC-2: Dataset formation reads data dir**
+Given images and labels in `data_images_dir` / `data_labels_dir`
+When `form_dataset()` runs
+Then it reads from `data_images_dir` and validates labels from `data_labels_dir`
+
+**AC-3: Data sample reads data dir**
+Given images in `data_images_dir`
+When `form_data_sample()` runs
+Then it reads from `data_images_dir`
+
+**AC-4: E2e test uses data dirs only**
+Given the e2e test fixture
+When setting up test data
+Then it copies images/labels only to `data_images_dir`/`data_labels_dir` (no processed dir)
diff --git a/_docs/02_tasks/todo/AZ-169_refactor_hard_symlinks.md b/_docs/02_tasks/todo/AZ-169_refactor_hard_symlinks.md
new file mode 100644
index 0000000..88054f4
--- /dev/null
+++ b/_docs/02_tasks/todo/AZ-169_refactor_hard_symlinks.md
@@ -0,0 +1,42 @@
+# Use Hard Symlinks for Dataset
+
+**Task**: AZ-169_refactor_hard_symlinks
+**Name**: Use hard symlinks instead of file copies for dataset formation
+**Description**: Replace shutil.copy() with os.link() in dataset split creation to save disk space
+**Complexity**: 2 points
+**Dependencies**: AZ-168_refactor_remove_processed_dir
+**Component**: Training Pipeline
+**Tracker**: AZ-169
+**Epic**: AZ-164
+
+## Problem
+
+`copy_annotations()` in `train.py` uses `shutil.copy()` to duplicate images and labels into train/valid/test splits. For large datasets this wastes significant disk space.
+
+## Outcome
+
+- Dataset formation uses `os.link()` (hard links) instead of `shutil.copy()`
+- Fallback to `shutil.copy()` when hard links fail (cross-filesystem)
+- No change in training behavior — YOLO reads hard-linked files identically
+
+## Scope
+
+### Included
+- Replace `shutil.copy()` with `os.link()` in `copy_annotations()` inner `copy_image()` function
+- Add try/except fallback to `shutil.copy()` for `OSError` (cross-filesystem)
+
+### Excluded
+- Changing `form_data_sample()` in exports.py (separate utility, lower priority)
+- Changing corrupted file handling
+
+## Acceptance Criteria
+
+**AC-1: Hard links used**
+Given images and labels in the data directory
+When `copy_annotations()` creates train/valid/test splits
+Then files are hard-linked via `os.link()`, not copied
+
+**AC-2: Fallback on failure**
+Given a cross-filesystem scenario where `os.link()` raises `OSError`
+When `copy_annotations()` encounters the error
+Then it falls back to `shutil.copy()` transparently
diff --git a/_docs/04_refactoring/01-code-improvements/analysis/refactoring_roadmap.md b/_docs/04_refactoring/01-code-improvements/analysis/refactoring_roadmap.md
new file mode 100644
index 0000000..27291da
--- /dev/null
+++ b/_docs/04_refactoring/01-code-improvements/analysis/refactoring_roadmap.md
@@ -0,0 +1,33 @@
+# Refactoring Roadmap
+
+**Run**: 01-code-improvements
+**Date**: 2026-03-28
+
+## Execution Order
+
+All 5 changes are grouped into a single phase (straightforward, low-to-medium risk).
+
+| Priority | Change | Risk | Effort |
+|----------|--------|------|--------|
+| 1 | C05: Unify configuration | medium | 3 pts |
+| 2 | C01: Update YOLO model | medium | 2 pts |
+| 3 | C02: Replace external augmentation | medium | 3 pts |
+| 4 | C03: Remove processed directory | medium | 3 pts |
+| 5 | C04: Hard symlinks | low | 2 pts |
+
+**Total estimated effort**: 13 points across 5 tasks
+
+## Dependency Graph
+
+```
+C05 (config unification) ─── independent
+C01 (YOLO update) ← C02 (built-in aug) ← C03 (remove processed dir) ← C04 (hard symlinks)
+```
+
+C05 can be done in parallel with the C01→C04 chain.
+
+## Risk Mitigation
+
+- Existing test suite (83 tests) provides safety net
+- Each change committed separately for easy rollback
+- C02 is the highest-risk change (training pipeline behavior change) — validate by running a short training sanity check after implementation
diff --git a/_docs/04_refactoring/01-code-improvements/analysis/research_findings.md b/_docs/04_refactoring/01-code-improvements/analysis/research_findings.md
new file mode 100644
index 0000000..b037765
--- /dev/null
+++ b/_docs/04_refactoring/01-code-improvements/analysis/research_findings.md
@@ -0,0 +1,50 @@
+# Research Findings
+
+**Run**: 01-code-improvements
+**Date**: 2026-03-28
+
+## Current State Analysis
+
+### Training Pipeline
+- Uses `yolo11m.yaml` (architecture-only config, trains from scratch)
+- External augmentation via `albumentations` library in `src/augmentation.py`
+- Two-step process: augment → form dataset → train
+- Dataset formation copies files with `shutil.copy()`, duplicating ~8x storage
+
+### Configuration
+- Two config files: `config.yaml` (root) and `src/annotation-queue/config.yaml`
+- Annotation queue handler parses YAML manually instead of using shared `Config` model
+- Config drift risk between the two files
+
+## YOLO 26 Model Update
+
+Ultralytics YOLO26 is the latest model family. The medium variant `yolo26m.pt` replaces `yolo11m.yaml`:
+- Uses pretrained weights (`.pt`) rather than architecture-only (`.yaml`)
+- Faster convergence with transfer learning
+- Improved accuracy on detection benchmarks
+
+## Built-in Augmentation Parameters
+
+YOLO's `model.train()` supports the following augmentation parameters that replace the external `albumentations` pipeline:
+
+| Parameter | Default | Equivalent to current external aug |
+|-----------|---------|-----------------------------------|
+| `hsv_h` | 0.015 | HueSaturationValue(hue_shift_limit=10) |
+| `hsv_s` | 0.7 | HueSaturationValue(sat_shift_limit=10) |
+| `hsv_v` | 0.4 | RandomBrightnessContrast + HSV |
+| `degrees` | 0.0 | Affine(rotate=(-35,35)) → set to 35.0 |
+| `translate` | 0.1 | Default is sufficient |
+| `scale` | 0.5 | Affine(scale=(0.8,1.2)) → default covers this |
+| `shear` | 0.0 | Affine(shear=(-10,10)) → set to 10.0 |
+| `fliplr` | 0.5 | HorizontalFlip(p=0.6) → set to 0.6 |
+| `flipud` | 0.0 | Not used currently |
+| `mosaic` | 1.0 | New — YOLO built-in |
+| `mixup` | 0.0 | New — optional |
+
+## Hard Symlinks
+
+`os.link()` creates hard links sharing the same inode. Benefits:
+- Zero additional disk usage for dataset splits
+- Same read performance as regular files
+- Works on same filesystem (which is the case here — all under `/azaion/`)
+- Fallback to `shutil.copy()` for cross-filesystem edge cases
diff --git a/_docs/04_refactoring/01-code-improvements/baseline_metrics.md b/_docs/04_refactoring/01-code-improvements/baseline_metrics.md
new file mode 100644
index 0000000..314d7e2
--- /dev/null
+++ b/_docs/04_refactoring/01-code-improvements/baseline_metrics.md
@@ -0,0 +1,66 @@
+# Baseline Metrics
+
+**Run**: 01-code-improvements
+**Date**: 2026-03-28
+**Mode**: Guided
+**Source**: `_docs/02_document/refactoring_notes.md`
+
+## Goals
+
+Apply 5 improvements identified during documentation:
+1. Update YOLO to v26m version
+2. Replace external augmentation with YOLO built-in augmentation
+3. Remove processed folder — use data dir directly
+4. Use hard symlinks instead of file copies for dataset formation
+5. Unify constants directories — remove `src/annotation-queue/config.yaml`
+
+## Code Metrics
+
+| Metric | Value |
+|--------|-------|
+| Source files (src/) | 24 Python files |
+| Source LOC | 2,945 |
+| Test files | 21 Python files |
+| Test LOC | 1,646 |
+| Total tests | 83 (77 blackbox/unit + 6 performance) |
+| Test execution time | ~130s (120s unit + 10s perf) |
+| Python version | 3.10.8 |
+| Ultralytics version | 8.4.30 |
+| Pip packages | ~76 |
+
+## Files Affected by Refactoring
+
+| File | LOC | Refactoring Items |
+|------|-----|-------------------|
+| `src/constants.py` | 118 | #3 (remove processed_dir), #5 (unify config) |
+| `src/train.py` | 178 | #1 (YOLO version), #2 (built-in aug), #3 (data dir), #4 (symlinks) |
+| `src/augmentation.py` | 152 | #2 (replace with YOLO built-in), #3 (processed dir) |
+| `src/exports.py` | 118 | #3 (processed dir references) |
+| `src/convert-annotations.py` | 119 | #3 (processed dir references) |
+| `src/dataset-visualiser.py` | 52 | #3 (processed dir references) |
+| `src/annotation-queue/annotation_queue_handler.py` | 173 | #5 (remove separate config.yaml) |
+| `src/annotation-queue/config.yaml` | 21 | #5 (delete — duplicated config) |
+| `config.yaml` | 30 | #5 (single source of truth) |
+
+## Test Suite Baseline
+
+```
+77 passed, 0 failed, 0 skipped (blackbox/unit)
+6 passed, 0 failed, 0 skipped (performance)
+Total: 83 passed in ~130s
+```
+
+## Functionality Inventory
+
+| Feature | Status | Affected by Refactoring |
+|---------|--------|------------------------|
+| Augmentation pipeline | Working | Yes (#2, #3) |
+| Dataset formation | Working | Yes (#3, #4) |
+| Training | Working | Yes (#1, #2) |
+| Model export (ONNX) | Working | No |
+| Inference (ONNX/TensorRT) | Working | No |
+| Annotation queue | Working | Yes (#5) |
+| API client | Working | No |
+| CDN manager | Working | No |
+| Security/encryption | Working | No |
+| Label validation | Working | No |
diff --git a/_docs/04_refactoring/01-code-improvements/discovery/components/01_training_pipeline.md b/_docs/04_refactoring/01-code-improvements/discovery/components/01_training_pipeline.md
new file mode 100644
index 0000000..8acc8b6
--- /dev/null
+++ b/_docs/04_refactoring/01-code-improvements/discovery/components/01_training_pipeline.md
@@ -0,0 +1,26 @@
+# Training Pipeline
+
+## Files
+- `src/train.py` (178 LOC)
+- `src/augmentation.py` (152 LOC)
+- `src/constants.py` (118 LOC)
+
+## Current Flow
+
+```mermaid
+graph TD
+    A[augmentation.py] -->|reads from| B[data_dir]
+    A -->|writes to| C[processed_dir]
+    D[train.py::form_dataset] -->|reads from| C
+    D -->|shutil.copy to| E[datasets_dir/today/train,valid,test]
+    F[train.py::train_dataset] -->|YOLO.train| E
+```
+
+## Issues
+- External augmentation (albumentations) runs as separate step, writing to `processed_dir`
+- `form_dataset()` copies files from `processed_dir` to dataset splits using `shutil.copy`
+- YOLO has built-in augmentation that runs during training (mosaic, mixup, flips, etc.)
+- Using built-in aug eliminates need for `processed_dir` and the full `augmentation.py` pipeline
+- `copy_annotations()` uses `shutil.copy` — wasteful for large datasets
+- Global mutable `total_files_copied` variable in `copy_annotations`
+- Model config `yolo11m.yaml` trains from scratch; likely should use pretrained weights or updated variant
diff --git a/_docs/04_refactoring/01-code-improvements/discovery/components/02_config_system.md b/_docs/04_refactoring/01-code-improvements/discovery/components/02_config_system.md
new file mode 100644
index 0000000..2e8dfd9
--- /dev/null
+++ b/_docs/04_refactoring/01-code-improvements/discovery/components/02_config_system.md
@@ -0,0 +1,18 @@
+# Configuration System
+
+## Files
+- `src/constants.py` (118 LOC)
+- `config.yaml` (root, 30 lines)
+- `src/annotation-queue/config.yaml` (21 lines)
+- `src/annotation-queue/annotation_queue_handler.py` (173 LOC)
+
+## Current State
+- `constants.py` defines `Config` (Pydantic model) loaded from root `config.yaml`
+- `annotation_queue_handler.py` reads its own `config.yaml` with raw `yaml.safe_load`
+- Both config files share `api`, `queue`, `dirs` sections but with different `dirs` values
+- Annotation queue config has `data: 'data-test'` vs root `data: 'data'`
+
+## Issues
+- Two config files with overlapping content — drift risk
+- `annotation_queue_handler.py` parses config manually instead of using `Config` model
+- `constants.py` still has `processed_dir` properties that become obsolete after removing external augmentation
diff --git a/_docs/04_refactoring/01-code-improvements/discovery/components/03_data_utilities.md b/_docs/04_refactoring/01-code-improvements/discovery/components/03_data_utilities.md
new file mode 100644
index 0000000..3e0138a
--- /dev/null
+++ b/_docs/04_refactoring/01-code-improvements/discovery/components/03_data_utilities.md
@@ -0,0 +1,10 @@
+# Data Utilities
+
+## Files
+- `src/exports.py` — `form_data_sample()` reads from `processed_images_dir`
+- `src/dataset-visualiser.py` — `visualise_processed_folder()` reads from `processed_images_dir`/`processed_labels_dir`
+
+## Impact
+- Both files reference `processed_dir` via `constants.config`
+- After removing `processed_dir`, these must switch to `data_images_dir`/`data_labels_dir`
+- `form_data_sample()` also uses `shutil.copy` — candidate for hard links
diff --git a/_docs/04_refactoring/01-code-improvements/list-of-changes.md b/_docs/04_refactoring/01-code-improvements/list-of-changes.md
new file mode 100644
index 0000000..70aa976
--- /dev/null
+++ b/_docs/04_refactoring/01-code-improvements/list-of-changes.md
@@ -0,0 +1,52 @@
+# List of Changes
+
+**Run**: 01-code-improvements
+**Mode**: guided
+**Source**: `_docs/02_document/refactoring_notes.md`
+**Date**: 2026-03-28
+
+## Summary
+
+Apply 5 improvements from documentation review: update YOLO model, switch to built-in augmentation, remove processed directory, use hard symlinks for dataset formation, and unify configuration files.
+
+## Changes
+
+### C01: Update YOLO model to 26m variant
+- **File(s)**: `src/constants.py`, `src/train.py`
+- **Problem**: Current model config uses `yolo11m.yaml` which trains from a YAML architecture definition
+- **Change**: Update `TrainingConfig.model` to the YOLO 26m variant; ensure `train_dataset()` uses the updated model reference
+- **Rationale**: Use updated model version as requested; pretrained weights improve convergence
+- **Risk**: medium
+- **Dependencies**: None
+
+### C02: Replace external augmentation with YOLO built-in
+- **File(s)**: `src/train.py`, `src/augmentation.py`
+- **Problem**: `augmentation.py` uses albumentations to augment images into a separate `processed_dir` before training — adds complexity, disk usage, and a separate processing step
+- **Change**: Remove the `augment_annotations()` call from the training pipeline; add YOLO built-in augmentation parameters (hsv_h, hsv_s, hsv_v, degrees, translate, scale, shear, flipud, fliplr, mosaic, mixup) to the `model.train()` call in `train_dataset()`, each on its own line with a descriptive comment; `augmentation.py` remains in codebase but is no longer called during training
+- **Rationale**: YOLO's built-in augmentation applies on-the-fly during training, eliminating the pre-processing step and processed directory
+- **Risk**: medium
+- **Dependencies**: C01
+
+### C03: Remove processed directory — use data dir directly
+- **File(s)**: `src/constants.py`, `src/train.py`, `src/exports.py`, `src/dataset-visualiser.py`
+- **Problem**: `processed_dir`, `processed_images_dir`, `processed_labels_dir` properties in `Config` are no longer needed when built-in augmentation is used; `form_dataset()` reads from processed dir; `form_data_sample()` reads from processed dir; `visualise_processed_folder()` reads from processed dir
+- **Change**: Remove `processed_dir`/`processed_images_dir`/`processed_labels_dir` properties from `Config`; update `form_dataset()` to read from `data_images_dir`/`data_labels_dir`; update `form_data_sample()` similarly; update `visualise_processed_folder()` similarly
+- **Rationale**: Processed directory is unnecessary without external augmentation step
+- **Risk**: medium
+- **Dependencies**: C02
+
+### C04: Use hard symlinks instead of file copies for dataset
+- **File(s)**: `src/train.py`
+- **Problem**: `copy_annotations()` uses `shutil.copy()` to duplicate images and labels into train/valid/test splits — wastes disk space on large datasets
+- **Change**: Replace `shutil.copy()` with `os.link()` to create hard links; add fallback to `shutil.copy()` for cross-filesystem scenarios
+- **Rationale**: Hard links share the same inode, saving disk space while maintaining independent directory entries
+- **Risk**: low
+- **Dependencies**: C03
+
+### C05: Unify configuration — remove annotation-queue/config.yaml
+- **File(s)**: `src/constants.py`, `src/annotation-queue/annotation_queue_handler.py`, `src/annotation-queue/config.yaml`
+- **Problem**: `src/annotation-queue/config.yaml` duplicates root `config.yaml` with different `dirs` values; `annotation_queue_handler.py` parses config manually via `yaml.safe_load` instead of using the shared `Config` model
+- **Change**: Extend `Config` in `constants.py` to include queue and annotation-queue directory settings; refactor `annotation_queue_handler.py` to accept a `Config` instance (or import from constants); delete `src/annotation-queue/config.yaml`
+- **Rationale**: Single source of truth for configuration eliminates drift risk and inconsistency
+- **Risk**: medium
+- **Dependencies**: None
diff --git a/_docs/_autopilot_state.md b/_docs/_autopilot_state.md
index 5bd50d2..87b3c26 100644
--- a/_docs/_autopilot_state.md
+++ b/_docs/_autopilot_state.md
@@ -2,8 +2,8 @@
 
 ## Current Step
 flow: existing-code
-step: 6
-name: Run Tests
+step: 7
+name: Refactor
 status: in_progress
-sub_step: 0
-retry_count: 1
+sub_step: 4 — Execution (Batch 1 done: AZ-165, AZ-166, AZ-167; next: Batch 2 AZ-168)
+retry_count: 0
diff --git a/config.yaml b/config.yaml
index 9b6ba6f..b7fff65 100644
--- a/config.yaml
+++ b/config.yaml
@@ -20,7 +20,7 @@ dirs:
   labels: 'labels'
 
 training:
-  model: 'yolo11m.yaml'
+  model: 'yolo26m.pt'
   epochs: 120
   batch: 11
   imgsz: 1280
diff --git a/src/annotation-queue/annotation_queue_handler.py b/src/annotation-queue/annotation_queue_handler.py
index 248dd39..543ffc4 100644
--- a/src/annotation-queue/annotation_queue_handler.py
+++ b/src/annotation-queue/annotation_queue_handler.py
@@ -1,12 +1,13 @@
 import os
 import shutil
 import sys
+from os import path, makedirs
 
+import constants
 import yaml
 import asyncio
 from rstream import Consumer, AMQPMessage, ConsumerOffsetSpecification, OffsetType, MessageContext
 from rstream.amqp import amqp_decoder
-from os import path, makedirs
 from datetime import datetime, timedelta
 import logging
 from logging.handlers import TimedRotatingFileHandler
@@ -30,7 +31,6 @@ logger.addHandler(logging.StreamHandler(sys.stdout))
 
 
 class AnnotationQueueHandler:
-    CONFIG_FILE = 'config.yaml'
     OFFSET_FILE = 'offset.yaml'
     QUEUE_ANNOTATION_STATUS_RECORD = b'AnnotationStatus'
     QUEUE_EMAIL_RECORD = b'Email'
@@ -46,18 +46,14 @@ class AnnotationQueueHandler:
             self.lbl_seed = path.join(h.data_seed_dir, h.labels_dir, f"{name}{h.TXT_EXT}")
 
     def __init__(self):
-        with open(self.CONFIG_FILE, "r") as f:
-            config_dict = yaml.safe_load(f)
+        cfg = constants.config
+        self.data_dir = path.join(cfg.dirs.root, cfg.dirs.data)
+        self.data_seed_dir = path.join(cfg.dirs.root, cfg.dirs.data_seed)
+        self.images_dir = cfg.dirs.images
+        self.labels_dir = cfg.dirs.labels
 
-        root     = config_dict['dirs']['root']
-
-        self.data_dir = path.join(root, config_dict['dirs']['data'])
-        self.data_seed_dir = path.join(root, config_dict['dirs']['data_seed'])
-        self.images_dir = config_dict['dirs']['images']
-        self.labels_dir = config_dict['dirs']['labels']
-
-        self.del_img_dir = path.join(root, config_dict['dirs']['data_deleted'], self.images_dir)
-        self.del_lbl_dir = path.join(root, config_dict['dirs']['data_deleted'], self.labels_dir)
+        self.del_img_dir = path.join(cfg.dirs.root, cfg.dirs.data_deleted, self.images_dir)
+        self.del_lbl_dir = path.join(cfg.dirs.root, cfg.dirs.data_deleted, self.labels_dir)
 
         makedirs(path.join(self.data_dir, self.images_dir), exist_ok=True)
         makedirs(path.join(self.data_dir, self.labels_dir), exist_ok=True)
@@ -68,12 +64,12 @@ class AnnotationQueueHandler:
         makedirs(self.del_lbl_dir, exist_ok=True)
 
         self.consumer = Consumer(
-            host=config_dict['queue']['host'],
-            port=config_dict['queue']['port'],
-            username=config_dict['queue']['consumer_user'],
-            password=config_dict['queue']['consumer_pw']
+            host=cfg.queue.host,
+            port=cfg.queue.port,
+            username=cfg.queue.consumer_user,
+            password=cfg.queue.consumer_pw
         )
-        self.queue_name = config_dict['queue']['name']
+        self.queue_name = cfg.queue.name
 
         try:
             with open(self.OFFSET_FILE, 'r') as f:
diff --git a/src/annotation-queue/config.yaml b/src/annotation-queue/config.yaml
deleted file mode 100644
index febb818..0000000
--- a/src/annotation-queue/config.yaml
+++ /dev/null
@@ -1,20 +0,0 @@
-api:
-  url: 'https://api.azaion.com'
-  email: 'uploader@azaion.com'
-  password: 'Az@1on_10Upl0@der'
-
-queue:
-  host: '188.245.120.247'
-  port: 5552
-  consumer_user: 'azaion_receiver'
-  consumer_pw: 'Az1onRecce777ve2r'
-  name: 'azaion-annotations'
-
-dirs:
-  root: '/azaion'
-  data: 'data-test'
-  data_seed: 'data-test-seed'
-  data_processed: 'data-test-processed'
-  data_deleted: 'data-test_deleted'
-  images: 'images'
-  labels: 'labels'
diff --git a/src/constants.py b/src/constants.py
index f83c298..e883be6 100644
--- a/src/constants.py
+++ b/src/constants.py
@@ -3,13 +3,35 @@ from os import path
 import yaml
 from pydantic import BaseModel
 
+_PROJECT_ROOT = path.dirname(path.dirname(path.abspath(__file__)))
+
+
+class ApiConfig(BaseModel):
+    url: str = ''
+    email: str = ''
+    password: str = ''
+
+
+class QueueConfig(BaseModel):
+    host: str = 'localhost'
+    port: int = 5552
+    consumer_user: str = ''
+    consumer_pw: str = ''
+    name: str = ''
+
 
 class DirsConfig(BaseModel):
     root: str = '/azaion'
+    data: str = 'data'
+    data_seed: str = 'data-seed'
+    data_processed: str = 'data-processed'
+    data_deleted: str = 'data_deleted'
+    images: str = 'images'
+    labels: str = 'labels'
 
 
 class TrainingConfig(BaseModel):
-    model: str = 'yolo11m.yaml'
+    model: str = 'yolo26m.pt'
     epochs: int = 120
     batch: int = 11
     imgsz: int = 1280
@@ -24,36 +46,38 @@ class ExportConfig(BaseModel):
 
 class Config(BaseModel):
     dirs: DirsConfig = DirsConfig()
+    api: ApiConfig = ApiConfig()
+    queue: QueueConfig = QueueConfig()
     training: TrainingConfig = TrainingConfig()
     export: ExportConfig = ExportConfig()
 
     @property
-    def azaion(self) -> str:
+    def root(self) -> str:
         return self.dirs.root
 
     @property
     def data_dir(self) -> str:
-        return path.join(self.dirs.root, 'data')
+        return path.join(self.dirs.root, self.dirs.data)
 
     @property
     def data_images_dir(self) -> str:
-        return path.join(self.data_dir, 'images')
+        return path.join(self.data_dir, self.dirs.images)
 
     @property
     def data_labels_dir(self) -> str:
-        return path.join(self.data_dir, 'labels')
+        return path.join(self.data_dir, self.dirs.labels)
 
     @property
     def processed_dir(self) -> str:
-        return path.join(self.dirs.root, 'data-processed')
+        return path.join(self.dirs.root, self.dirs.data_processed)
 
     @property
     def processed_images_dir(self) -> str:
-        return path.join(self.processed_dir, 'images')
+        return path.join(self.processed_dir, self.dirs.images)
 
     @property
     def processed_labels_dir(self) -> str:
-        return path.join(self.processed_dir, 'labels')
+        return path.join(self.processed_dir, self.dirs.labels)
 
     @property
     def corrupted_dir(self) -> str:
@@ -61,11 +85,11 @@ class Config(BaseModel):
 
     @property
     def corrupted_images_dir(self) -> str:
-        return path.join(self.corrupted_dir, 'images')
+        return path.join(self.corrupted_dir, self.dirs.images)
 
     @property
     def corrupted_labels_dir(self) -> str:
-        return path.join(self.corrupted_dir, 'labels')
+        return path.join(self.corrupted_dir, self.dirs.labels)
 
     @property
     def sample_dir(self) -> str:
@@ -115,4 +139,4 @@ SMALL_SIZE_KB = 3
 CDN_CONFIG = 'cdn.yaml'
 MODELS_FOLDER = 'models'
 
-config: Config = Config.from_yaml(CONFIG_FILE)
+config: Config = Config.from_yaml(path.join(_PROJECT_ROOT, CONFIG_FILE))
diff --git a/src/train.py b/src/train.py
index 486e86a..8c0dc4a 100644
--- a/src/train.py
+++ b/src/train.py
@@ -132,13 +132,23 @@ def create_yaml():
 
 def resume_training(last_pt_path):
     model = YOLO(last_pt_path)
-    results = model.train(data=yaml,
-                          resume=True,
-                          epochs=constants.config.training.epochs,
-                          batch=constants.config.training.batch,
-                          imgsz=constants.config.training.imgsz,
-                          save_period=constants.config.training.save_period,
-                          workers=constants.config.training.workers)
+    results = model.train(data=yaml,                          # path to dataset YAML
+                          resume=True,                        # continue from last checkpoint
+                          epochs=constants.config.training.epochs,  # total training epochs
+                          batch=constants.config.training.batch,    # batch size per GPU
+                          imgsz=constants.config.training.imgsz,    # input image size in pixels
+                          save_period=constants.config.training.save_period,  # save checkpoint every N epochs
+                          workers=constants.config.training.workers,  # dataloader worker threads
+                          hsv_h=0.015,                        # hue shift fraction of the color wheel
+                          hsv_s=0.7,                          # saturation shift fraction
+                          hsv_v=0.4,                          # value (brightness) shift fraction
+                          degrees=35.0,                       # max rotation degrees (+/- range)
+                          translate=0.1,                      # max translation as fraction of image size
+                          scale=0.5,                          # max scale gain (1 +/- scale)
+                          shear=10.0,                         # max shear degrees (+/- range)
+                          fliplr=0.6,                         # probability of horizontal flip
+                          mosaic=1.0,                         # mosaic augmentation probability
+                          mixup=0.0)                          # mixup augmentation probability (disabled)
 
 
 def train_dataset():
@@ -147,12 +157,22 @@ def train_dataset():
     model = YOLO(constants.config.training.model)
 
     today_dataset = path.join(constants.config.datasets_dir, today_folder)
-    results = model.train(data=abspath(path.join(today_dataset, 'data.yaml')),
-                           epochs=constants.config.training.epochs,
-                           batch=constants.config.training.batch,
-                           imgsz=constants.config.training.imgsz,
-                           save_period=constants.config.training.save_period,
-                           workers=constants.config.training.workers)
+    results = model.train(data=abspath(path.join(today_dataset, 'data.yaml')),  # path to dataset YAML
+                           epochs=constants.config.training.epochs,             # total training epochs
+                           batch=constants.config.training.batch,               # batch size per GPU
+                           imgsz=constants.config.training.imgsz,               # input image size in pixels
+                           save_period=constants.config.training.save_period,   # save checkpoint every N epochs
+                           workers=constants.config.training.workers,           # dataloader worker threads
+                           hsv_h=0.015,                                         # hue shift fraction of the color wheel
+                           hsv_s=0.7,                                           # saturation shift fraction
+                           hsv_v=0.4,                                           # value (brightness) shift fraction
+                           degrees=35.0,                                        # max rotation degrees (+/- range)
+                           translate=0.1,                                       # max translation as fraction of image size
+                           scale=0.5,                                           # max scale gain (1 +/- scale)
+                           shear=10.0,                                        # max shear degrees (+/- range)
+                           fliplr=0.6,                                          # probability of horizontal flip
+                           mosaic=1.0,                                          # mosaic augmentation probability
+                           mixup=0.0)                                           # mixup augmentation probability (disabled)
 
     model_dir = path.join(constants.config.models_dir, today_folder)
 
diff --git a/tests/conftest.py b/tests/conftest.py
index 0738ee1..45452b5 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -4,8 +4,10 @@ from pathlib import Path
 import pytest
 
 _PROJECT_ROOT = Path(__file__).resolve().parent.parent
-_DATASET_IMAGES = _PROJECT_ROOT / "_docs/00_problem/input_data/dataset/images"
-_DATASET_LABELS = _PROJECT_ROOT / "_docs/00_problem/input_data/dataset/labels"
+_TESTS_DIR = Path(__file__).resolve().parent
+_TEST_ROOT = _TESTS_DIR / "root"
+_DATASET_IMAGES = _TEST_ROOT / "data" / "images"
+_DATASET_LABELS = _TEST_ROOT / "data" / "labels"
 _ONNX_MODEL = _PROJECT_ROOT / "_docs/00_problem/input_data/azaion.onnx"
 _CLASSES_JSON = _PROJECT_ROOT / "src" / "classes.json"
 _CONFIG_TEST = _PROJECT_ROOT / "config.test.yaml"
diff --git a/tests/performance/test_augmentation_perf.py b/tests/performance/test_augmentation_perf.py
deleted file mode 100644
index a7c62cb..0000000
--- a/tests/performance/test_augmentation_perf.py
+++ /dev/null
@@ -1,115 +0,0 @@
-import concurrent.futures
-import random
-import shutil
-import time
-from pathlib import Path
-
-import numpy as np
-import pytest
-
-from tests.conftest import apply_constants_patch
-
-
-def _patch_augmentation_paths(monkeypatch, base: Path):
-    apply_constants_patch(monkeypatch, base)
-
-
-def _augment_annotation_with_total(monkeypatch):
-    import augmentation as aug
-
-    orig = aug.Augmentator.augment_annotation
-
-    def wrapped(self, image_file):
-        self.total_to_process = self.total_images_to_process
-        return orig(self, image_file)
-
-    monkeypatch.setattr(aug.Augmentator, "augment_annotation", wrapped)
-
-
-def _seed():
-    random.seed(42)
-    np.random.seed(42)
-
-
-@pytest.mark.performance
-def test_pt_aug_01_throughput_ten_images_sixty_seconds(
-    tmp_path, monkeypatch, sample_images_labels
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _augment_annotation_with_total(monkeypatch)
-    _seed()
-    import constants as c
-    from augmentation import Augmentator
-
-    img_dir = Path(c.config.data_images_dir)
-    lbl_dir = Path(c.config.data_labels_dir)
-    img_dir.mkdir(parents=True, exist_ok=True)
-    lbl_dir.mkdir(parents=True, exist_ok=True)
-    src_img, src_lbl = sample_images_labels(10)
-    for p in src_img.glob("*.jpg"):
-        shutil.copy2(p, img_dir / p.name)
-    for p in src_lbl.glob("*.txt"):
-        shutil.copy2(p, lbl_dir / p.name)
-    # Act
-    t0 = time.perf_counter()
-    Augmentator().augment_annotations()
-    elapsed = time.perf_counter() - t0
-    # Assert
-    assert elapsed <= 60.0
-
-
-@pytest.mark.performance
-def test_pt_aug_02_parallel_at_least_one_point_five_x_faster(
-    tmp_path, monkeypatch, sample_images_labels
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _augment_annotation_with_total(monkeypatch)
-    _seed()
-    import constants as c
-    from augmentation import Augmentator
-
-    img_dir = Path(c.config.data_images_dir)
-    lbl_dir = Path(c.config.data_labels_dir)
-    proc_dir = Path(c.config.processed_dir)
-    img_dir.mkdir(parents=True, exist_ok=True)
-    lbl_dir.mkdir(parents=True, exist_ok=True)
-    src_img, src_lbl = sample_images_labels(10)
-    for p in src_img.glob("*.jpg"):
-        shutil.copy2(p, img_dir / p.name)
-    for p in src_lbl.glob("*.txt"):
-        shutil.copy2(p, lbl_dir / p.name)
-    Path(c.config.processed_images_dir).mkdir(parents=True, exist_ok=True)
-    Path(c.config.processed_labels_dir).mkdir(parents=True, exist_ok=True)
-    names = sorted(p.name for p in img_dir.glob("*.jpg"))
-
-    class _E:
-        __slots__ = ("name",)
-
-        def __init__(self, name):
-            self.name = name
-
-    entries = [_E(n) for n in names]
-
-    # Act
-    aug_seq = Augmentator()
-    aug_seq.total_images_to_process = len(entries)
-    t0 = time.perf_counter()
-    for e in entries:
-        aug_seq.augment_annotation(e)
-    seq_elapsed = time.perf_counter() - t0
-
-    shutil.rmtree(proc_dir)
-    Path(c.config.processed_images_dir).mkdir(parents=True, exist_ok=True)
-    Path(c.config.processed_labels_dir).mkdir(parents=True, exist_ok=True)
-
-    aug_par = Augmentator()
-    aug_par.total_images_to_process = len(entries)
-    t0 = time.perf_counter()
-    with concurrent.futures.ThreadPoolExecutor() as ex:
-        list(ex.map(aug_par.augment_annotation, entries))
-    par_elapsed = time.perf_counter() - t0
-
-    # Assert
-    assert seq_elapsed >= par_elapsed * 1.5
diff --git a/tests/root/data-processed/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.jpg b/tests/root/data-processed/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.jpg
new file mode 100644
index 0000000..87bf488
Binary files /dev/null and b/tests/root/data-processed/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.jpg differ
diff --git a/tests/root/data-processed/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.jpg b/tests/root/data-processed/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.jpg
new file mode 100644
index 0000000..732e64d
Binary files /dev/null and b/tests/root/data-processed/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000106.jpg b/tests/root/data-processed/images/#Азов_з_неба_000106.jpg
new file mode 100644
index 0000000..f316b2b
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000106.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000113.jpg b/tests/root/data-processed/images/#Азов_з_неба_000113.jpg
new file mode 100644
index 0000000..3871e98
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000113.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000148.jpg b/tests/root/data-processed/images/#Азов_з_неба_000148.jpg
new file mode 100644
index 0000000..8c54eba
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000148.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000162.jpg b/tests/root/data-processed/images/#Азов_з_неба_000162.jpg
new file mode 100644
index 0000000..480b9ac
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000162.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000173.jpg b/tests/root/data-processed/images/#Азов_з_неба_000173.jpg
new file mode 100644
index 0000000..ee949db
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000173.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000189.jpg b/tests/root/data-processed/images/#Азов_з_неба_000189.jpg
new file mode 100644
index 0000000..d9e1058
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000189.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000210.jpg b/tests/root/data-processed/images/#Азов_з_неба_000210.jpg
new file mode 100644
index 0000000..85de419
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000210.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000245.jpg b/tests/root/data-processed/images/#Азов_з_неба_000245.jpg
new file mode 100644
index 0000000..664dfc3
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000245.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000269.jpg b/tests/root/data-processed/images/#Азов_з_неба_000269.jpg
new file mode 100644
index 0000000..b765680
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000269.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000307.jpg b/tests/root/data-processed/images/#Азов_з_неба_000307.jpg
new file mode 100644
index 0000000..1ad21e2
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000307.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000314.jpg b/tests/root/data-processed/images/#Азов_з_неба_000314.jpg
new file mode 100644
index 0000000..89ecc54
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000314.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000320.jpg b/tests/root/data-processed/images/#Азов_з_неба_000320.jpg
new file mode 100644
index 0000000..dd66c48
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000320.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000338.jpg b/tests/root/data-processed/images/#Азов_з_неба_000338.jpg
new file mode 100644
index 0000000..14744f5
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000338.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000346.jpg b/tests/root/data-processed/images/#Азов_з_неба_000346.jpg
new file mode 100644
index 0000000..52986da
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000346.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000347.jpg b/tests/root/data-processed/images/#Азов_з_неба_000347.jpg
new file mode 100644
index 0000000..dcedfe7
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000347.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000373.jpg b/tests/root/data-processed/images/#Азов_з_неба_000373.jpg
new file mode 100644
index 0000000..49ac20a
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000373.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000380.jpg b/tests/root/data-processed/images/#Азов_з_неба_000380.jpg
new file mode 100644
index 0000000..27dd8d8
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000380.jpg differ
diff --git a/tests/root/data-processed/images/#Азов_з_неба_000403.jpg b/tests/root/data-processed/images/#Азов_з_неба_000403.jpg
new file mode 100644
index 0000000..4774f3c
Binary files /dev/null and b/tests/root/data-processed/images/#Азов_з_неба_000403.jpg differ
diff --git a/tests/root/data-processed/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.txt b/tests/root/data-processed/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.txt
new file mode 100644
index 0000000..f0231e6
--- /dev/null
+++ b/tests/root/data-processed/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.txt
@@ -0,0 +1 @@
+0 0.63516 0.76054 0.05480 0.09305
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.txt b/tests/root/data-processed/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.txt
new file mode 100644
index 0000000..e03dfe2
--- /dev/null
+++ b/tests/root/data-processed/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.txt
@@ -0,0 +1 @@
+0 0.61853 0.77915 0.05973 0.09305
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000106.txt b/tests/root/data-processed/labels/#Азов_з_неба_000106.txt
new file mode 100644
index 0000000..304e549
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000106.txt
@@ -0,0 +1 @@
+0 0.47200 0.78007 0.26215 0.42338
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000113.txt b/tests/root/data-processed/labels/#Азов_з_неба_000113.txt
new file mode 100644
index 0000000..104c11e
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000113.txt
@@ -0,0 +1,2 @@
+0 0.76062 0.31074 0.08738 0.13238
+0 0.79600 0.20352 0.04985 0.08424
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000148.txt b/tests/root/data-processed/labels/#Азов_з_неба_000148.txt
new file mode 100644
index 0000000..67e97f5
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000148.txt
@@ -0,0 +1,6 @@
+6 0.52646 0.69638 0.04738 0.18161
+6 0.56554 0.69856 0.03077 0.16848
+6 0.59908 0.69311 0.03754 0.17263
+6 0.70000 0.74293 0.04738 0.17176
+6 0.77046 0.69638 0.03077 0.14441
+6 0.73538 0.70190 0.03077 0.14660
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000162.txt b/tests/root/data-processed/labels/#Азов_з_неба_000162.txt
new file mode 100644
index 0000000..2216408
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000162.txt
@@ -0,0 +1,4 @@
+6 0.63569 0.46827 0.04185 0.13675
+6 0.68492 0.38403 0.04800 0.13019
+6 0.61569 0.36161 0.03877 0.10065
+6 0.56708 0.44147 0.04985 0.12253
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000173.txt b/tests/root/data-processed/labels/#Азов_з_неба_000173.txt
new file mode 100644
index 0000000..08ac396
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000173.txt
@@ -0,0 +1 @@
+3 0.74738 0.58588 0.33415 0.49450
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000189.txt b/tests/root/data-processed/labels/#Азов_з_неба_000189.txt
new file mode 100644
index 0000000..546bd29
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000189.txt
@@ -0,0 +1 @@
+0 0.91200 0.50492 0.09846 0.08971
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000210.txt b/tests/root/data-processed/labels/#Азов_з_неба_000210.txt
new file mode 100644
index 0000000..5a5eefa
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000210.txt
@@ -0,0 +1 @@
+0 0.87846 0.50930 0.09785 0.07658
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000245.txt b/tests/root/data-processed/labels/#Азов_з_неба_000245.txt
new file mode 100644
index 0000000..7b9d776
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000245.txt
@@ -0,0 +1 @@
+2 0.53169 0.44475 0.06523 0.12691
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000269.txt b/tests/root/data-processed/labels/#Азов_з_неба_000269.txt
new file mode 100644
index 0000000..a6dd311
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000269.txt
@@ -0,0 +1 @@
+2 0.51323 0.45679 0.08369 0.08752
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000307.txt b/tests/root/data-processed/labels/#Азов_з_неба_000307.txt
new file mode 100644
index 0000000..f6614c0
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000307.txt
@@ -0,0 +1 @@
+0 0.90154 0.37309 0.08246 0.10831
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000314.txt b/tests/root/data-processed/labels/#Азов_з_неба_000314.txt
new file mode 100644
index 0000000..dfe2158
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000314.txt
@@ -0,0 +1,4 @@
+6 0.35723 0.48851 0.03015 0.12253
+6 0.21415 0.44147 0.02954 0.11597
+6 0.24277 0.44147 0.02769 0.11597
+6 0.28892 0.44530 0.02769 0.08205
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000320.txt b/tests/root/data-processed/labels/#Азов_з_неба_000320.txt
new file mode 100644
index 0000000..b9f65f6
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000320.txt
@@ -0,0 +1 @@
+3 0.24338 0.33097 0.16308 0.21443
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000338.txt b/tests/root/data-processed/labels/#Азов_з_неба_000338.txt
new file mode 100644
index 0000000..f15f033
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000338.txt
@@ -0,0 +1 @@
+3 0.53200 0.18055 0.07815 0.08862
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000346.txt b/tests/root/data-processed/labels/#Азов_з_неба_000346.txt
new file mode 100644
index 0000000..cfb31fd
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000346.txt
@@ -0,0 +1 @@
+1 0.63538 0.20571 0.05354 0.11050
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000347.txt b/tests/root/data-processed/labels/#Азов_з_неба_000347.txt
new file mode 100644
index 0000000..bccb5e5
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000347.txt
@@ -0,0 +1 @@
+3 0.70092 0.15046 0.04431 0.06564
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000373.txt b/tests/root/data-processed/labels/#Азов_з_неба_000373.txt
new file mode 100644
index 0000000..5415a80
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000373.txt
@@ -0,0 +1 @@
+1 0.28431 0.81398 0.12677 0.33805
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000380.txt b/tests/root/data-processed/labels/#Азов_з_неба_000380.txt
new file mode 100644
index 0000000..194c6ac
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000380.txt
@@ -0,0 +1,2 @@
+0 0.33108 0.75600 0.23262 0.35556
+6 0.10862 0.51641 0.03385 0.09080
\ No newline at end of file
diff --git a/tests/root/data-processed/labels/#Азов_з_неба_000403.txt b/tests/root/data-processed/labels/#Азов_з_неба_000403.txt
new file mode 100644
index 0000000..035b823
--- /dev/null
+++ b/tests/root/data-processed/labels/#Азов_з_неба_000403.txt
@@ -0,0 +1 @@
+0 0.41138 0.64113 0.29108 0.21552
\ No newline at end of file
diff --git a/tests/root/data/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.jpg b/tests/root/data/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.jpg
new file mode 100644
index 0000000..87bf488
Binary files /dev/null and b/tests/root/data/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.jpg differ
diff --git a/tests/root/data/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.jpg b/tests/root/data/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.jpg
new file mode 100644
index 0000000..732e64d
Binary files /dev/null and b/tests/root/data/images/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000106.jpg b/tests/root/data/images/#Азов_з_неба_000106.jpg
new file mode 100644
index 0000000..f316b2b
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000106.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000113.jpg b/tests/root/data/images/#Азов_з_неба_000113.jpg
new file mode 100644
index 0000000..3871e98
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000113.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000148.jpg b/tests/root/data/images/#Азов_з_неба_000148.jpg
new file mode 100644
index 0000000..8c54eba
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000148.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000162.jpg b/tests/root/data/images/#Азов_з_неба_000162.jpg
new file mode 100644
index 0000000..480b9ac
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000162.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000173.jpg b/tests/root/data/images/#Азов_з_неба_000173.jpg
new file mode 100644
index 0000000..ee949db
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000173.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000189.jpg b/tests/root/data/images/#Азов_з_неба_000189.jpg
new file mode 100644
index 0000000..d9e1058
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000189.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000210.jpg b/tests/root/data/images/#Азов_з_неба_000210.jpg
new file mode 100644
index 0000000..85de419
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000210.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000245.jpg b/tests/root/data/images/#Азов_з_неба_000245.jpg
new file mode 100644
index 0000000..664dfc3
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000245.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000269.jpg b/tests/root/data/images/#Азов_з_неба_000269.jpg
new file mode 100644
index 0000000..b765680
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000269.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000307.jpg b/tests/root/data/images/#Азов_з_неба_000307.jpg
new file mode 100644
index 0000000..1ad21e2
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000307.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000314.jpg b/tests/root/data/images/#Азов_з_неба_000314.jpg
new file mode 100644
index 0000000..89ecc54
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000314.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000320.jpg b/tests/root/data/images/#Азов_з_неба_000320.jpg
new file mode 100644
index 0000000..dd66c48
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000320.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000338.jpg b/tests/root/data/images/#Азов_з_неба_000338.jpg
new file mode 100644
index 0000000..14744f5
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000338.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000346.jpg b/tests/root/data/images/#Азов_з_неба_000346.jpg
new file mode 100644
index 0000000..52986da
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000346.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000347.jpg b/tests/root/data/images/#Азов_з_неба_000347.jpg
new file mode 100644
index 0000000..dcedfe7
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000347.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000373.jpg b/tests/root/data/images/#Азов_з_неба_000373.jpg
new file mode 100644
index 0000000..49ac20a
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000373.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000380.jpg b/tests/root/data/images/#Азов_з_неба_000380.jpg
new file mode 100644
index 0000000..27dd8d8
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000380.jpg differ
diff --git a/tests/root/data/images/#Азов_з_неба_000403.jpg b/tests/root/data/images/#Азов_з_неба_000403.jpg
new file mode 100644
index 0000000..4774f3c
Binary files /dev/null and b/tests/root/data/images/#Азов_з_неба_000403.jpg differ
diff --git a/tests/root/data/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.txt b/tests/root/data/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.txt
new file mode 100644
index 0000000..f0231e6
--- /dev/null
+++ b/tests/root/data/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001226.txt
@@ -0,0 +1 @@
+0 0.63516 0.76054 0.05480 0.09305
\ No newline at end of file
diff --git a/tests/root/data/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.txt b/tests/root/data/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.txt
new file mode 100644
index 0000000..e03dfe2
--- /dev/null
+++ b/tests/root/data/labels/!Працюєпершийсамохіднийартилерійськийдивізіон.К_001278.txt
@@ -0,0 +1 @@
+0 0.61853 0.77915 0.05973 0.09305
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000106.txt b/tests/root/data/labels/#Азов_з_неба_000106.txt
new file mode 100644
index 0000000..304e549
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000106.txt
@@ -0,0 +1 @@
+0 0.47200 0.78007 0.26215 0.42338
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000113.txt b/tests/root/data/labels/#Азов_з_неба_000113.txt
new file mode 100644
index 0000000..104c11e
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000113.txt
@@ -0,0 +1,2 @@
+0 0.76062 0.31074 0.08738 0.13238
+0 0.79600 0.20352 0.04985 0.08424
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000148.txt b/tests/root/data/labels/#Азов_з_неба_000148.txt
new file mode 100644
index 0000000..67e97f5
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000148.txt
@@ -0,0 +1,6 @@
+6 0.52646 0.69638 0.04738 0.18161
+6 0.56554 0.69856 0.03077 0.16848
+6 0.59908 0.69311 0.03754 0.17263
+6 0.70000 0.74293 0.04738 0.17176
+6 0.77046 0.69638 0.03077 0.14441
+6 0.73538 0.70190 0.03077 0.14660
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000162.txt b/tests/root/data/labels/#Азов_з_неба_000162.txt
new file mode 100644
index 0000000..2216408
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000162.txt
@@ -0,0 +1,4 @@
+6 0.63569 0.46827 0.04185 0.13675
+6 0.68492 0.38403 0.04800 0.13019
+6 0.61569 0.36161 0.03877 0.10065
+6 0.56708 0.44147 0.04985 0.12253
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000173.txt b/tests/root/data/labels/#Азов_з_неба_000173.txt
new file mode 100644
index 0000000..08ac396
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000173.txt
@@ -0,0 +1 @@
+3 0.74738 0.58588 0.33415 0.49450
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000189.txt b/tests/root/data/labels/#Азов_з_неба_000189.txt
new file mode 100644
index 0000000..546bd29
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000189.txt
@@ -0,0 +1 @@
+0 0.91200 0.50492 0.09846 0.08971
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000210.txt b/tests/root/data/labels/#Азов_з_неба_000210.txt
new file mode 100644
index 0000000..5a5eefa
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000210.txt
@@ -0,0 +1 @@
+0 0.87846 0.50930 0.09785 0.07658
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000245.txt b/tests/root/data/labels/#Азов_з_неба_000245.txt
new file mode 100644
index 0000000..7b9d776
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000245.txt
@@ -0,0 +1 @@
+2 0.53169 0.44475 0.06523 0.12691
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000269.txt b/tests/root/data/labels/#Азов_з_неба_000269.txt
new file mode 100644
index 0000000..a6dd311
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000269.txt
@@ -0,0 +1 @@
+2 0.51323 0.45679 0.08369 0.08752
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000307.txt b/tests/root/data/labels/#Азов_з_неба_000307.txt
new file mode 100644
index 0000000..f6614c0
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000307.txt
@@ -0,0 +1 @@
+0 0.90154 0.37309 0.08246 0.10831
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000314.txt b/tests/root/data/labels/#Азов_з_неба_000314.txt
new file mode 100644
index 0000000..dfe2158
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000314.txt
@@ -0,0 +1,4 @@
+6 0.35723 0.48851 0.03015 0.12253
+6 0.21415 0.44147 0.02954 0.11597
+6 0.24277 0.44147 0.02769 0.11597
+6 0.28892 0.44530 0.02769 0.08205
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000320.txt b/tests/root/data/labels/#Азов_з_неба_000320.txt
new file mode 100644
index 0000000..b9f65f6
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000320.txt
@@ -0,0 +1 @@
+3 0.24338 0.33097 0.16308 0.21443
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000338.txt b/tests/root/data/labels/#Азов_з_неба_000338.txt
new file mode 100644
index 0000000..f15f033
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000338.txt
@@ -0,0 +1 @@
+3 0.53200 0.18055 0.07815 0.08862
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000346.txt b/tests/root/data/labels/#Азов_з_неба_000346.txt
new file mode 100644
index 0000000..cfb31fd
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000346.txt
@@ -0,0 +1 @@
+1 0.63538 0.20571 0.05354 0.11050
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000347.txt b/tests/root/data/labels/#Азов_з_неба_000347.txt
new file mode 100644
index 0000000..bccb5e5
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000347.txt
@@ -0,0 +1 @@
+3 0.70092 0.15046 0.04431 0.06564
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000373.txt b/tests/root/data/labels/#Азов_з_неба_000373.txt
new file mode 100644
index 0000000..5415a80
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000373.txt
@@ -0,0 +1 @@
+1 0.28431 0.81398 0.12677 0.33805
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000380.txt b/tests/root/data/labels/#Азов_з_неба_000380.txt
new file mode 100644
index 0000000..194c6ac
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000380.txt
@@ -0,0 +1,2 @@
+0 0.33108 0.75600 0.23262 0.35556
+6 0.10862 0.51641 0.03385 0.09080
\ No newline at end of file
diff --git a/tests/root/data/labels/#Азов_з_неба_000403.txt b/tests/root/data/labels/#Азов_з_неба_000403.txt
new file mode 100644
index 0000000..035b823
--- /dev/null
+++ b/tests/root/data/labels/#Азов_з_неба_000403.txt
@@ -0,0 +1 @@
+0 0.41138 0.64113 0.29108 0.21552
\ No newline at end of file
diff --git a/tests/test_augmentation.py b/tests/test_augmentation.py
deleted file mode 100644
index 88ca3c9..0000000
--- a/tests/test_augmentation.py
+++ /dev/null
@@ -1,267 +0,0 @@
-import random
-import shutil
-from pathlib import Path
-
-import cv2
-import numpy as np
-
-from tests.conftest import apply_constants_patch
-
-
-def _patch_augmentation_paths(monkeypatch, base: Path):
-    apply_constants_patch(monkeypatch, base)
-
-
-def _seed():
-    random.seed(42)
-    np.random.seed(42)
-
-
-def _augment_annotation_with_total(monkeypatch):
-    import augmentation as aug
-
-    orig = aug.Augmentator.augment_annotation
-
-    def wrapped(self, image_file):
-        self.total_to_process = self.total_images_to_process
-        return orig(self, image_file)
-
-    monkeypatch.setattr(aug.Augmentator, "augment_annotation", wrapped)
-
-
-def test_bt_aug_01_augment_inner_returns_eight_image_labels(
-    tmp_path, monkeypatch, fixture_images_dir, fixture_labels_dir
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _seed()
-    from augmentation import Augmentator
-
-    stem = sorted(fixture_images_dir.glob("*.jpg"))[0].stem
-    img_path = fixture_images_dir / f"{stem}.jpg"
-    lbl_path = fixture_labels_dir / f"{stem}.txt"
-    img = cv2.imdecode(np.fromfile(str(img_path), dtype=np.uint8), cv2.IMREAD_COLOR)
-    aug = Augmentator()
-    labels = aug.read_labels(lbl_path)
-    proc_img = Path(tmp_path) / "azaion" / "data-processed" / "images" / f"{stem}.jpg"
-    proc_lbl = Path(tmp_path) / "azaion" / "data-processed" / "labels" / f"{stem}.txt"
-    proc_img.parent.mkdir(parents=True, exist_ok=True)
-    proc_lbl.parent.mkdir(parents=True, exist_ok=True)
-    from dto.imageLabel import ImageLabel
-
-    img_ann = ImageLabel(
-        image_path=str(proc_img),
-        image=img,
-        labels_path=str(proc_lbl),
-        labels=labels,
-    )
-    # Act
-    out = aug.augment_inner(img_ann)
-    # Assert
-    assert len(out) == 8
-
-
-def test_bt_aug_02_naming_convention(tmp_path, monkeypatch, fixture_images_dir, fixture_labels_dir):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _seed()
-    from augmentation import Augmentator
-    from dto.imageLabel import ImageLabel
-
-    stem = "test_image"
-    proc_img = Path(tmp_path) / "azaion" / "data-processed" / "images" / f"{stem}.jpg"
-    proc_lbl = Path(tmp_path) / "azaion" / "data-processed" / "labels" / f"{stem}.txt"
-    proc_img.parent.mkdir(parents=True, exist_ok=True)
-    proc_lbl.parent.mkdir(parents=True, exist_ok=True)
-    src_img = sorted(fixture_images_dir.glob("*.jpg"))[0]
-    img = cv2.imdecode(np.fromfile(str(src_img), dtype=np.uint8), cv2.IMREAD_COLOR)
-    lbl_path = fixture_labels_dir / f"{src_img.stem}.txt"
-    labels = Augmentator().read_labels(lbl_path)
-    aug = Augmentator()
-    img_ann = ImageLabel(
-        image_path=str(proc_img),
-        image=img,
-        labels_path=str(proc_lbl),
-        labels=labels,
-    )
-    # Act
-    out = aug.augment_inner(img_ann)
-    # Assert
-    names = [Path(o.image_path).name for o in out]
-    expected = [f"{stem}.jpg"] + [f"{stem}_{i}.jpg" for i in range(1, 8)]
-    assert names == expected
-    lbl_names = [Path(o.labels_path).name for o in out]
-    expected_lbl = [f"{stem}.txt"] + [f"{stem}_{i}.txt" for i in range(1, 8)]
-    assert lbl_names == expected_lbl
-
-
-def _all_coords_in_unit(labels_list):
-    for row in labels_list:
-        for j in range(4):
-            v = float(row[j])
-            if v < 0.0 or v > 1.0:
-                return False
-    return True
-
-
-def test_bt_aug_03_all_bbox_coords_in_zero_one(
-    tmp_path, monkeypatch, fixture_images_dir, fixture_labels_dir
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _seed()
-    from augmentation import Augmentator
-    from dto.imageLabel import ImageLabel
-
-    stem = sorted(fixture_images_dir.glob("*.jpg"))[0].stem
-    proc_img = Path(tmp_path) / "azaion" / "data-processed" / "images" / f"{stem}.jpg"
-    proc_lbl = Path(tmp_path) / "azaion" / "data-processed" / "labels" / f"{stem}.txt"
-    proc_img.parent.mkdir(parents=True, exist_ok=True)
-    proc_lbl.parent.mkdir(parents=True, exist_ok=True)
-    img_path = fixture_images_dir / f"{stem}.jpg"
-    lbl_path = fixture_labels_dir / f"{stem}.txt"
-    img = cv2.imdecode(np.fromfile(str(img_path), dtype=np.uint8), cv2.IMREAD_COLOR)
-    aug = Augmentator()
-    labels = aug.read_labels(lbl_path)
-    img_ann = ImageLabel(
-        image_path=str(proc_img),
-        image=img,
-        labels_path=str(proc_lbl),
-        labels=labels,
-    )
-    # Act
-    out = aug.augment_inner(img_ann)
-    # Assert
-    for o in out:
-        for row in o.labels:
-            assert len(row) >= 5
-        assert _all_coords_in_unit(o.labels)
-
-
-def test_bt_aug_04_correct_bboxes_clips_edge(tmp_path, monkeypatch):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    from augmentation import Augmentator
-
-    aug = Augmentator()
-    m = aug.correct_margin
-    inp = [[0.99, 0.5, 0.2, 0.1, 0]]
-    # Act
-    res = aug.correct_bboxes(inp)
-    # Assert
-    assert len(res) == 1
-    x, y, w, h, _ = res[0]
-    hw, hh = 0.5 * w, 0.5 * h
-    assert x - hw >= m - 1e-9
-    assert x + hw <= 1.0 - m + 1e-9
-    assert y - hh >= m - 1e-9
-    assert y + hh <= 1.0 - m + 1e-9
-
-
-def test_bt_aug_05_tiny_bbox_removed_after_clipping(tmp_path, monkeypatch):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    from augmentation import Augmentator
-
-    aug = Augmentator()
-    inp = [[0.995, 0.5, 0.01, 0.5, 0]]
-    # Act
-    res = aug.correct_bboxes(inp)
-    # Assert
-    assert res == []
-
-
-def test_bt_aug_06_empty_label_eight_outputs_empty_labels(
-    tmp_path, monkeypatch, fixture_images_dir
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _seed()
-    from augmentation import Augmentator
-    from dto.imageLabel import ImageLabel
-
-    stem = "empty_case"
-    proc_img = Path(tmp_path) / "azaion" / "data-processed" / "images" / f"{stem}.jpg"
-    proc_lbl = Path(tmp_path) / "azaion" / "data-processed" / "labels" / f"{stem}.txt"
-    proc_img.parent.mkdir(parents=True, exist_ok=True)
-    proc_lbl.parent.mkdir(parents=True, exist_ok=True)
-    src_img = sorted(fixture_images_dir.glob("*.jpg"))[0]
-    img = cv2.imdecode(np.fromfile(str(src_img), dtype=np.uint8), cv2.IMREAD_COLOR)
-    aug = Augmentator()
-    img_ann = ImageLabel(
-        image_path=str(proc_img),
-        image=img,
-        labels_path=str(proc_lbl),
-        labels=[],
-    )
-    # Act
-    out = aug.augment_inner(img_ann)
-    # Assert
-    assert len(out) == 8
-    for o in out:
-        assert o.labels == []
-
-
-def test_bt_aug_07_full_pipeline_five_images_forty_outputs(
-    tmp_path, monkeypatch, sample_images_labels
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _augment_annotation_with_total(monkeypatch)
-    _seed()
-    import constants as c
-    from augmentation import Augmentator
-
-    img_dir = Path(c.config.data_images_dir)
-    lbl_dir = Path(c.config.data_labels_dir)
-    img_dir.mkdir(parents=True, exist_ok=True)
-    lbl_dir.mkdir(parents=True, exist_ok=True)
-    src_img, src_lbl = sample_images_labels(5)
-    for p in src_img.glob("*.jpg"):
-        shutil.copy2(p, img_dir / p.name)
-    for p in src_lbl.glob("*.txt"):
-        shutil.copy2(p, lbl_dir / p.name)
-    # Act
-    Augmentator().augment_annotations()
-    # Assert
-    proc_img = Path(c.config.processed_images_dir)
-    proc_lbl = Path(c.config.processed_labels_dir)
-    assert len(list(proc_img.glob("*.jpg"))) == 40
-    assert len(list(proc_lbl.glob("*.txt"))) == 40
-
-
-def test_bt_aug_08_skips_already_processed(tmp_path, monkeypatch, sample_images_labels):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _augment_annotation_with_total(monkeypatch)
-    _seed()
-    import constants as c
-    from augmentation import Augmentator
-
-    img_dir = Path(c.config.data_images_dir)
-    lbl_dir = Path(c.config.data_labels_dir)
-    proc_img = Path(c.config.processed_images_dir)
-    proc_lbl = Path(c.config.processed_labels_dir)
-    img_dir.mkdir(parents=True, exist_ok=True)
-    lbl_dir.mkdir(parents=True, exist_ok=True)
-    proc_img.mkdir(parents=True, exist_ok=True)
-    proc_lbl.mkdir(parents=True, exist_ok=True)
-    src_img, src_lbl = sample_images_labels(5)
-    jpgs = sorted(src_img.glob("*.jpg"))
-    for p in jpgs:
-        shutil.copy2(p, img_dir / p.name)
-    for p in src_lbl.glob("*.txt"):
-        shutil.copy2(p, lbl_dir / p.name)
-    markers = []
-    for p in jpgs[:3]:
-        dst = proc_img / p.name
-        shutil.copy2(p, dst)
-        markers.append(dst.read_bytes())
-    # Act
-    Augmentator().augment_annotations()
-    # Assert
-    after_jpgs = list(proc_img.glob("*.jpg"))
-    assert len(after_jpgs) == 19
-    assert len(list(proc_lbl.glob("*.txt"))) == 16
-    for i, p in enumerate(jpgs[:3]):
-        assert (proc_img / p.name).read_bytes() == markers[i]
diff --git a/tests/test_augmentation_nonfunc.py b/tests/test_augmentation_nonfunc.py
deleted file mode 100644
index b2bf7a6..0000000
--- a/tests/test_augmentation_nonfunc.py
+++ /dev/null
@@ -1,143 +0,0 @@
-import random
-import shutil
-from pathlib import Path
-from types import SimpleNamespace
-
-import cv2
-import numpy as np
-import pytest
-
-from tests.conftest import apply_constants_patch
-
-
-def _patch_augmentation_paths(monkeypatch, base: Path):
-    apply_constants_patch(monkeypatch, base)
-
-
-def _augment_annotation_with_total(monkeypatch):
-    import augmentation as aug
-
-    orig = aug.Augmentator.augment_annotation
-
-    def wrapped(self, image_file):
-        self.total_to_process = self.total_images_to_process
-        return orig(self, image_file)
-
-    monkeypatch.setattr(aug.Augmentator, "augment_annotation", wrapped)
-
-
-def _seed():
-    random.seed(42)
-    np.random.seed(42)
-
-
-@pytest.mark.resilience
-def test_rt_aug_01_corrupted_image_skipped(
-    tmp_path, monkeypatch, fixture_images_dir, fixture_labels_dir
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _augment_annotation_with_total(monkeypatch)
-    _seed()
-    import constants as c
-    from augmentation import Augmentator
-
-    img_dir = Path(c.config.data_images_dir)
-    lbl_dir = Path(c.config.data_labels_dir)
-    img_dir.mkdir(parents=True, exist_ok=True)
-    lbl_dir.mkdir(parents=True, exist_ok=True)
-    stem = sorted(fixture_images_dir.glob("*.jpg"))[0].stem
-    shutil.copy2(fixture_images_dir / f"{stem}.jpg", img_dir / f"{stem}.jpg")
-    shutil.copy2(fixture_labels_dir / f"{stem}.txt", lbl_dir / f"{stem}.txt")
-    raw = (fixture_images_dir / f"{stem}.jpg").read_bytes()[:200]
-    (img_dir / "corrupted_trunc.jpg").write_bytes(raw)
-    # Act
-    Augmentator().augment_annotations()
-    # Assert
-    proc_img = Path(c.config.processed_images_dir)
-    assert len(list(proc_img.glob("*.jpg"))) == 8
-
-
-@pytest.mark.resilience
-def test_rt_aug_02_missing_label_no_crash(tmp_path, monkeypatch, fixture_images_dir):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _augment_annotation_with_total(monkeypatch)
-    import constants as c
-    from augmentation import Augmentator
-
-    img_dir = Path(c.config.data_images_dir)
-    lbl_dir = Path(c.config.data_labels_dir)
-    img_dir.mkdir(parents=True, exist_ok=True)
-    lbl_dir.mkdir(parents=True, exist_ok=True)
-    stem = "no_label_here"
-    shutil.copy2(sorted(fixture_images_dir.glob("*.jpg"))[0], img_dir / f"{stem}.jpg")
-    aug = Augmentator()
-    aug.total_images_to_process = 1
-    # Act
-    aug.augment_annotation(SimpleNamespace(name=f"{stem}.jpg"))
-    # Assert
-    assert len(list(Path(c.config.processed_images_dir).glob("*.jpg"))) == 0
-
-
-@pytest.mark.resilience
-def test_rt_aug_03_narrow_bbox_fewer_or_eight_variants(
-    tmp_path, monkeypatch, fixture_images_dir
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _seed()
-    from augmentation import Augmentator
-    from dto.imageLabel import ImageLabel
-
-    stem = "narrow_bbox"
-    proc_img = Path(tmp_path) / "azaion" / "data-processed" / "images" / f"{stem}.jpg"
-    proc_lbl = Path(tmp_path) / "azaion" / "data-processed" / "labels" / f"{stem}.txt"
-    proc_img.parent.mkdir(parents=True, exist_ok=True)
-    proc_lbl.parent.mkdir(parents=True, exist_ok=True)
-    src_img = sorted(fixture_images_dir.glob("*.jpg"))[0]
-    img = cv2.imdecode(np.fromfile(str(src_img), dtype=np.uint8), cv2.IMREAD_COLOR)
-    aug = Augmentator()
-    labels = [[0.5, 0.5, 0.0005, 0.0005, 0]]
-    img_ann = ImageLabel(
-        image_path=str(proc_img),
-        image=img,
-        labels_path=str(proc_lbl),
-        labels=labels,
-    )
-    # Act
-    out = aug.augment_inner(img_ann)
-    # Assert
-    assert 1 <= len(out) <= 8
-
-
-@pytest.mark.resource_limit
-def test_rl_aug_01_augment_inner_exactly_eight_outputs(
-    tmp_path, monkeypatch, fixture_images_dir, fixture_labels_dir
-):
-    # Arrange
-    _patch_augmentation_paths(monkeypatch, tmp_path)
-    _seed()
-    from augmentation import Augmentator
-    from dto.imageLabel import ImageLabel
-
-    stem = sorted(fixture_images_dir.glob("*.jpg"))[0].stem
-    img_path = fixture_images_dir / f"{stem}.jpg"
-    lbl_path = fixture_labels_dir / f"{stem}.txt"
-    img = cv2.imdecode(np.fromfile(str(img_path), dtype=np.uint8), cv2.IMREAD_COLOR)
-    aug = Augmentator()
-    labels = aug.read_labels(lbl_path)
-    proc_img = Path(tmp_path) / "azaion" / "data-processed" / "images" / f"{stem}.jpg"
-    proc_lbl = Path(tmp_path) / "azaion" / "data-processed" / "labels" / f"{stem}.txt"
-    proc_img.parent.mkdir(parents=True, exist_ok=True)
-    proc_lbl.parent.mkdir(parents=True, exist_ok=True)
-    img_ann = ImageLabel(
-        image_path=str(proc_img),
-        image=img,
-        labels_path=str(proc_lbl),
-        labels=labels,
-    )
-    # Act
-    out = aug.augment_inner(img_ann)
-    # Assert
-    assert len(out) == 8
diff --git a/tests/test_dataset_formation.py b/tests/test_dataset_formation.py
index 46ecb49..4f4ae20 100644
--- a/tests/test_dataset_formation.py
+++ b/tests/test_dataset_formation.py
@@ -55,15 +55,15 @@ def test_bt_dsf_01_split_ratio_70_20_10(
         constants_patch,
         fixture_images_dir,
         fixture_labels_dir,
-        100,
+        20,
         set(),
     )
     # Act
     train.form_dataset()
     # Assert
-    assert _count_jpg(Path(today_ds, "train", "images")) == 70
-    assert _count_jpg(Path(today_ds, "valid", "images")) == 20
-    assert _count_jpg(Path(today_ds, "test", "images")) == 10
+    assert _count_jpg(Path(today_ds, "train", "images")) == 14
+    assert _count_jpg(Path(today_ds, "valid", "images")) == 4
+    assert _count_jpg(Path(today_ds, "test", "images")) == 2
 
 
 def test_bt_dsf_02_six_subdirectories(
@@ -80,7 +80,7 @@ def test_bt_dsf_02_six_subdirectories(
         constants_patch,
         fixture_images_dir,
         fixture_labels_dir,
-        100,
+        20,
         set(),
     )
     # Act
@@ -95,7 +95,7 @@ def test_bt_dsf_02_six_subdirectories(
     assert (base / "test" / "labels").is_dir()
 
 
-def test_bt_dsf_03_total_files_one_hundred(
+def test_bt_dsf_03_total_files_twenty(
     monkeypatch,
     tmp_path,
     constants_patch,
@@ -109,7 +109,7 @@ def test_bt_dsf_03_total_files_one_hundred(
         constants_patch,
         fixture_images_dir,
         fixture_labels_dir,
-        100,
+        20,
         set(),
     )
     # Act
@@ -120,7 +120,7 @@ def test_bt_dsf_03_total_files_one_hundred(
         + _count_jpg(Path(today_ds, "valid", "images"))
         + _count_jpg(Path(today_ds, "test", "images"))
     )
-    assert n == 100
+    assert n == 20
 
 
 def test_bt_dsf_04_corrupted_labels_quarantined(
@@ -131,7 +131,7 @@ def test_bt_dsf_04_corrupted_labels_quarantined(
     fixture_labels_dir,
 ):
     # Arrange
-    stems = [p.stem for p in sorted(fixture_images_dir.glob("*.jpg"))[:100]]
+    stems = [p.stem for p in sorted(fixture_images_dir.glob("*.jpg"))[:20]]
     corrupt = set(stems[:5])
     train, today_ds = _prepare_form_dataset(
         monkeypatch,
@@ -139,7 +139,7 @@ def test_bt_dsf_04_corrupted_labels_quarantined(
         constants_patch,
         fixture_images_dir,
         fixture_labels_dir,
-        100,
+        20,
         corrupt,
     )
     # Act
@@ -150,7 +150,7 @@ def test_bt_dsf_04_corrupted_labels_quarantined(
         + _count_jpg(Path(today_ds, "valid", "images"))
         + _count_jpg(Path(today_ds, "test", "images"))
     )
-    assert split_total == 95
+    assert split_total == 15
     assert _count_jpg(c_mod.config.corrupted_images_dir) == 5
     assert len(list(Path(c_mod.config.corrupted_labels_dir).glob("*.txt"))) == 5
 
@@ -202,7 +202,7 @@ def test_rl_dsf_02_no_filename_duplication_across_splits(
         constants_patch,
         fixture_images_dir,
         fixture_labels_dir,
-        100,
+        20,
         set(),
     )
     # Act
@@ -214,4 +214,4 @@ def test_rl_dsf_02_no_filename_duplication_across_splits(
         for f in (base / split / "images").glob("*.jpg"):
             names.append(f.name)
     assert len(names) == len(set(names))
-    assert len(names) == 100
+    assert len(names) == 20
diff --git a/tests/test_infrastructure.py b/tests/test_infrastructure.py
index 1a7a43b..bb22ec5 100644
--- a/tests/test_infrastructure.py
+++ b/tests/test_infrastructure.py
@@ -3,12 +3,12 @@ import constants as c
 
 def test_fixture_images_dir_has_jpegs(fixture_images_dir):
     jpgs = list(fixture_images_dir.glob("*.jpg"))
-    assert len(jpgs) == 100
+    assert len(jpgs) == 20
 
 
 def test_fixture_labels_dir_has_yolo_labels(fixture_labels_dir):
     txts = list(fixture_labels_dir.glob("*.txt"))
-    assert len(txts) == 100
+    assert len(txts) == 20
 
 
 def test_fixture_onnx_model_bytes(fixture_onnx_model):
@@ -54,6 +54,6 @@ def test_empty_label_file(empty_label):
 
 def test_constants_patch_uses_tmp(constants_patch, tmp_path):
     constants_patch(tmp_path)
-    assert c.config.azaion.startswith(str(tmp_path))
+    assert c.config.root.startswith(str(tmp_path))
     assert c.config.data_dir.startswith(str(tmp_path))
     assert c.config.current_onnx_model.startswith(str(tmp_path))
diff --git a/tests/test_training_e2e.py b/tests/test_training_e2e.py
index 40a6a61..21ba7e4 100644
--- a/tests/test_training_e2e.py
+++ b/tests/test_training_e2e.py
@@ -9,33 +9,18 @@ import constants as c
 import train as train_mod
 import exports as exports_mod
 
-_PROJECT_ROOT = Path(__file__).resolve().parent.parent
-_DATASET_IMAGES = _PROJECT_ROOT / "_docs/00_problem/input_data/dataset/images"
-_DATASET_LABELS = _PROJECT_ROOT / "_docs/00_problem/input_data/dataset/labels"
-_CONFIG_TEST = _PROJECT_ROOT / "config.test.yaml"
+_TESTS_DIR = Path(__file__).resolve().parent
+_TEST_ROOT = _TESTS_DIR / "root"
+_DATASET_IMAGES = _TEST_ROOT / "data" / "images"
+_CONFIG_TEST = _TESTS_DIR.parent / "config.test.yaml"
 
 
 @pytest.fixture(scope="module")
-def e2e_result(tmp_path_factory):
-    base = tmp_path_factory.mktemp("e2e")
-
+def e2e_result():
     old_config = c.config
-    c.config = c.Config.from_yaml(str(_CONFIG_TEST), root=str(base / "azaion"))
+    c.config = c.Config.from_yaml(str(_CONFIG_TEST), root=str(_TEST_ROOT))
 
-    data_img = Path(c.config.data_images_dir)
-    data_lbl = Path(c.config.data_labels_dir)
-    data_img.mkdir(parents=True)
-    data_lbl.mkdir(parents=True)
-    Path(c.config.models_dir).mkdir(parents=True)
-
-    for img in sorted(_DATASET_IMAGES.glob("*.jpg")):
-        shutil.copy2(img, data_img / img.name)
-        lbl = _DATASET_LABELS / f"{img.stem}.txt"
-        if lbl.exists():
-            shutil.copy2(lbl, data_lbl / lbl.name)
-
-    from augmentation import Augmentator
-    Augmentator().augment_annotations()
+    Path(c.config.models_dir).mkdir(parents=True, exist_ok=True)
 
     train_mod.train_dataset()
 
@@ -48,15 +33,14 @@ def e2e_result(tmp_path_factory):
         "today_dataset": today_ds,
     }
 
+    shutil.rmtree(c.config.datasets_dir, ignore_errors=True)
+    shutil.rmtree(c.config.models_dir, ignore_errors=True)
+    shutil.rmtree(c.config.corrupted_dir, ignore_errors=True)
     c.config = old_config
 
 
 @pytest.mark.e2e
 class TestTrainingPipeline:
-    def test_augmentation_produced_output(self, e2e_result):
-        proc = Path(c.config.processed_images_dir)
-        assert len(list(proc.glob("*.jpg"))) == 800
-
     def test_dataset_formed(self, e2e_result):
         base = Path(e2e_result["today_dataset"])
         for split in ("train", "valid", "test"):
@@ -66,7 +50,7 @@ class TestTrainingPipeline:
             len(list((base / s / "images").glob("*.jpg")))
             for s in ("train", "valid", "test")
         )
-        assert total == 800
+        assert total == 20
 
     def test_data_yaml_created(self, e2e_result):
         yaml_path = Path(e2e_result["today_dataset"]) / "data.yaml"