6ea9bd33f5
patches
2026-02-24 22:19:59 -05:00
e021fb32cd
Release 2.0.4-beta.10: Audio PyPI fix (tiktoken workaround complete)
...
Audio/Whisper works with pip install - no Git workaround needed.
See CHANGELOG.md for details.
Tested: 647 passed, 11 skipped (Python 3.10-3.12)
2026-02-05 10:42:50 +01:00
bf7480d042
Release 2.0.4-beta.9: Audio transcription via mlx-audio
...
Major Features:
- Audio transcription via mlx-audio backend (Whisper, >10min duration)
- OpenAI /v1/audio/transcriptions endpoint
- Memory Gate System (Vision: 8GB, Audio: 4GB)
- Config-based backend routing (ADR-020)
- Benchmark toolchain (memmon/memplot, Schema v0.2.2)
Key Fixes:
- EuroLLM tokenizer decoding
- Vision-model text-only routing regression
- Multimodal model context length detection
- Memory cleanup bug (mx.metal.clear_cache)
- Orphan process bug
Test Results:
- Unit tests: 647 passed, 11 skipped (Python 3.10-3.12)
- wet-umbrella: 171 passed total
See CHANGELOG.md for complete details and known issues.
2026-02-04 03:10:30 +01:00
e8b10ea10b
Release 2.0.4-beta.8: Audio transcription support (experimental)
...
Audio input via --audio flag (CLI) and input_audio content type (Server API).
Uses mlx-vlm native audio processing. ~30s duration limit (model constraint).
Currently only Gemma-3n tested (requires --repair-index fix).
Also includes:
- SERVER-HANDBOOK compliance (image limits, validation error envelopes)
- Dependency updates: mlx>=0.30.0, mlx-lm>=0.30.0, huggingface-hub>=1.0.0
- Audio E2E test suite + ADR-019
2026-01-23 20:20:59 +01:00
5751545b8b
Release 2.0.4-beta.7: Server robustness + Vision per-chunk streaming
...
- Server: exit codes, /v1/models crash fix, vision routing, MLXK2_MAX_TOKENS
- Vision: true SSE streaming, hallucination fix (local numbering)
- Workspace: list prefix-match, push ambiguous pattern handling
- Docs: SERVER-HANDBOOK accuracy updates
See CHANGELOG.md for details.
2026-01-18 16:57:32 +01:00
53d9cca82d
Release 2.0.4-beta.6: Local workspace workflow + Vision batch processing
...
- Complete local development cycle: clone → repair → run/show/server on
workspace paths without HuggingFace round-trips
- Vision processing now defaults to safe chunking (one image at a time,
prevents OOM + hallucination)
- Resumable clone with --force-resume and deterministic temp cache naming
- Improved test infrastructure (umbrella marker convention)
- 161 Wet Umbrella tests passing including new Vision→Geo pipe integration tests
See CHANGELOG.md for complete details.
2026-01-07 17:11:07 +01:00
25609e4dcb
Release 2.0.4-beta.5: Community repair tool + OS-agnostic benchmarking
...
Closes #49 (Mistral Tokenizer Bug)
Major features:
- Workspace Infrastructure (ADR-018 Phase 0a): Managed workspace detection,
provenance metadata, backward compatible with unmanaged workspaces
- Convert Operation (ADR-018 Phase 1): `mlxk convert --repair-index` fixes
mlx-vlm #624 affected models (7+ models including Qwen2.5-VL, gemma-3)
- Resumable Pull: Auto-detect partial downloads with `--force-resume`
- Wet Umbrella Test Integration: Single entry point for all real model tests
Fixes:
- #49 : BPE space markers now correctly converted (Mistral-family models)
- Vision Portfolio Discovery: Filter by capabilities instead of model_type
- Memory Cleanup Hook: Triggers for both live_e2e and wet markers
Test suite: 528 passed, 60 skipped (Python 3.9-3.14)
2025-12-31 16:05:18 +01:00
d3f7d091bc
Release 2.0.4-beta.3: Dependency compatibility + Documentation
...
Bugfixes and compatibility improvements. No new features.
Core fixes:
- Framework detection for web API models (Issue #48 )
- Video-only model filtering from vision capability
- Page size detection for memory metrics (macOS)
- Model switch log timing (after load completion)
Compatibility:
- hub 1.x + transformers 5.0 support
- Python 3.9-3.14 verified (494 tests passing)
Testing infrastructure:
- Benchmark schema v0.2.0 (hardware profiling, system health)
- Benchmark template v1.0 (automated JSONL→Markdown reports)
- Memory timeline visualization (memplot.py)
- Unified model filter (build_model_object single source)
Documentation:
- Multi-Modal Support section in README (Vision subsection)
- JSON API 0.1.5-0.1.6 marked Stable
- Vision promoted from alpha to beta status
- Removed conceptual drift and outdated references
See CHANGELOG.md for complete details.
2025-12-23 12:19:04 +01:00
86f669dc82
Release 2.0.4-beta.1: Vision + Pipes + Memory
...
- Vision Support (Issue #45 ): CLI + Server with OpenAI-compatible image API, EXIF metadata
- Unix Pipes (ADR-014): stdin support, isatty detection, SIGPIPE handling
- Memory-Aware Loading (ADR-016): Pre-load checks with >70% RAM warnings
- Python 3.9-3.14: Full compatibility verified (476-485 tests passing)
- Fixed: --log-json regression (Issue #44 ), Vision multimodal history filtering
See CHANGELOG.md for complete details.
2025-12-16 19:35:30 +01:00
05f1c30486
Release 2.0.3: Foundation for pipes
...
Foundation release for Unix pipe integration with stderr separation,
benchmark infrastructure, and reasoning control improvements.
Breaking Changes:
- stdout/stderr separation (Issue #43 ) - errors to stderr in human mode
- JSON mode unchanged (all output to stdout)
Features:
- Benchmark reporting infrastructure (ADR-013 Phase 0)
- --no-reasoning flag (Issue #40 partial - GPT-OSS/QwQ only)
- Interactive mode reasoning control (review_report.md fixes)
Bug Fixes:
- huggingface-hub 1.x incompatibility (critical dependency fix)
- Streaming parity tests refactored (Portfolio Discovery)
Testing:
- 308 tests passing (Python 3.9-3.13)
- 35 skipped (opt-in live tests)
- 79/91 E2E tests passing with HF_HOME
See CHANGELOG.md for complete details and migration guide.
2025-11-17 22:54:06 +01:00
fb54f59cd4
Release 2.0.0-beta.6: Stop token & compatibility bug fixes
...
Fixes Issue #32 (generic multi-EOS detection) and Issue #37 (model detection)
- Generic stop token detection: Multi-EOS models (MXFP4, Qwen, Llama) now use eos_token_ids Set instead of
model-specific workarounds
- Private/org MLX model detection: `mlxk run` now works outside `mlx-community/*` namespace
- Commit-pinned compatibility checks: Models with `@commit_hash` validated before inference
- Packaging dependencies: Fixed `pip install -e .` requirements
- ADR-009: Stop Token Detection Fix (generic approach + test strategy)
- ADR-011: E2E Live Test Architecture (planned)
See CHANGELOG.md and TESTING.md for details.
2025-10-24 15:46:42 +02:00
57bf6d86be
2.0.0-beta.3: Feature Complete - Full 1.1.1 Parity Achieved
...
Major Features Added:
• Complete run command implementation with interactive/single-shot modes
• MLXRunner core engine ported from 1.x with modular architecture
• OpenAI-compatible server with SIGINT-robust supervisor mode
• Experimental push feature properly isolated behind environment variable
Key Improvements:
- Full feature parity with 1.1.1 stable releases
- Enhanced human output formatting across all commands
- Clean separation of stable (184 tests) vs experimental features
- Updated demo GIF showcasing improved 2.0 interface
Fixes:
- Pull operation cache pollution (Issue #30 ) with preflight access checks
- Test stability improvements across all environments
Architecture:
- Modular runner design with focused helper modules
- Thread-safe model loading and memory management
- stable testing across Python 3.9-3.13
Ready for use as comprehensive 1.x alternative.
2025-09-14 18:04:18 +02:00
3f57248121
2.0.0-alpha.3: lenient MLX detection + push branch handling
...
- Detect MLX/chat via README front‑matter + tokenizer; unify list/show; human list filters aligned (Refs #31 )
- Push: create missing branch with --create and retry once on “Invalid rev id”; tolerate missing branches
offline; no‑op still creates branch with --create
- Tests: add offline retry test; detection/human coverage; live list (opt‑in); 98/98 passing
- Docs/Meta: CHANGELOG/TESTING/README/SECURITY/CLAUDE updated; hard split 1.x from this branch; Apache‑2.0 + NOTICE
2025-09-08 01:14:01 +02:00
eedb91b75c
Feat: add experimental push (2.0.0-alpha.2)
...
- Push (upload-only): quiet JSON by default; capture hub logs in data.hf_logs
- No-op detection aligned to hub signal; clear commit fields; uploaded_files_count=0
- Add --dry-run (plan vs remote) and --check-only (offline preflight); merge .hfignore; extend
default ignores
- Human output: concise; --verbose shows commit URL; JSON shape unchanged
- Tests: add offline dry-run cases; live push remains opt-in (wet/live_push)
- Docs: README push section updated; TESTING.md reference + mini-matrix;
- Changelog: add 2.0.0-alpha.2; note Issue #31 under 1.1.1 pending
- Spec: keep schema stable (0.1.3); CLI/version docs consistent
2025-09-05 22:42:39 +02:00
cbd25c658d
Release MLX Knife 1.0.2 - HF_HOME Cache Consistency & Corruption Fixes │
...
│ │
│ Major bug fixes addressing cache path inconsistencies and silent failures: │
│ - Fix Issue #11 : HF_HOME environment variable handling - unified cache logic ensures consistent │
│ /hub subdirectory usage │
│ - Fix Issue #9 : Silent failure on corrupted models with empty snapshots directories │
│ - Enhanced download throttling with adaptive delays (512KB chunks, 2-3s for large files) │
│ - Added migration warnings for legacy cache locations with clear user guidance │
│ - Improved corruption detection and deletion workflow consistency │
│ │
│ Technical improvements: │
│ - Unified cache architecture: CACHE_ROOT/hub for both default and HF_HOME scenarios │
│ - Exception-safe memory management with enhanced baseline tracking │
│ - Updated dependencies to latest tested versions (Python 3.9-3.13 support) │
│ - All 105 tests passing with real MLX model verification
2025-08-18 14:02:30 +02:00
6b04448d1c
Initial commit: MLX Knife 1.0-rc1
2025-08-12 23:00:55 +02:00
Nikesh Parajuli