monster/corridorkey-mrp-mlx
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
20 Commits 1 Branch 0 Tags 400 KiB
6c711e85bd
Commit Graph

20 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
cmoyates
6c711e85bd
fix: remove double normalize_rgb in engine preprocessing 
preprocess() already calls normalize_rgb() internally; calling it
before passing to preprocess() applied ImageNet normalization twice.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 06:47:22 -03:30
cmoyates
4bad586112
feat: add CorridorKeyMLXEngine integration surface 
Drop-in MLX backend for main CorridorKey repo. Engine class wraps
existing model/inference with process_frame() API returning alpha,
fg, comp, processed as uint8 numpy arrays. Lowers Python to >=3.11,
removes unused deps, adds smoke script and 16 contract tests.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 06:43:07 -03:30
Cristopher Yates
b0c7d8086c
Merge pull request #5 from cmoyates/feat/phase6-optimization 
feat(phase6): optimization, benchmarking, and tiled inference
 2026-03-01 06:27:57 -03:30
cmoyates
3d330d0e0b
fix: guard against overlap >= tile_size in tiled inference 
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 06:27:31 -03:30
cmoyates
57c3aff913
feat(phase6): optimization, benchmarking, and tiled inference 
- Cache nn.Upsample instances in DecoderHead/GreenFormer __init__
  (eliminated ~7 allocations per forward pass)
- Add mx.compile() support via load_model(compile=True)
- Benchmark harness: eager vs compiled, multi-resolution, parity checks
- Tiled inference with overlap blending for large images
- Profiling utilities with forced mx.eval for accurate timing
- Reference comparison script (scripts/compare_reference.py)
- 12 new tests (compiled consistency + tiling)
- README performance section with Apple Silicon guidance

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 06:24:10 -03:30
Cristopher Yates
a14f8a74e8
Merge pull request #4 from cmoyates/feat/phase5-assembly-parity 
feat(phase5): full model assembly + e2e parity
 2026-03-01 06:12:04 -03:30
cmoyates
329db3fc43
fix: resize inputs to model img_size, use infer_and_save in CLI 
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 06:11:57 -03:30
cmoyates
0481955cfd
feat(phase5): full model assembly + e2e parity 
Wire GreenFormer (backbone + decoders + refiner), add image I/O,
inference pipeline, CLI entry point, and 3 test files (23 new tests).

E2e parity vs PyTorch: max abs err < 1.5e-4 across all outputs.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 06:07:43 -03:30
Cristopher Yates
fd6ca78935
Merge pull request #3 from cmoyates/feat/phase4a-hiera-backbone 
feat(phase4): MLX Hiera backbone port
 2026-03-01 05:59:19 -03:30
Cristopher Yates
c2e7967538
feat(phase4): MLX Hiera backbone port (#2) 
* feat(phase4a): MLX Hiera backbone with unroll/reroll parity

Port timm hiera_base_plus_224 to MLX: PatchEmbed, MaskUnitAttention,
HieraBlock, unroll/reroll, HieraBackbone. 4 NHWC feature maps at
strides 4/8/16/32. backbone.py now re-exports from hiera.py.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat(phase4c): weight loading + pos_embed bicubic interpolation

HieraBackbone.load_checkpoint() loads safetensors, strips encoder.model.
prefix, bicubic-interpolates pos_embed from 512x512 to 128x128 tokens.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat(phase4d): shape + parity tests, fix attention transpose

Fix MaskUnitAttention output transpose (0,2,3,1,4 → 0,3,2,1,4) to
match PyTorch transpose(1,3) token ordering for windowed attention.

Parity results (4 stages):
  Stage 0: max_abs 2.9e-4, mean 1.6e-5
  Stage 1: max_abs 1.3e-4, mean 1.3e-5
  Stage 2: max_abs 1.1e-2, mean 5.2e-5 (16 blocks, expected drift)
  Stage 3: max_abs 5.8e-4, mean 2.6e-5

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* docs: clarify pos_embed parameter intent in HieraBackbone

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 05:55:24 -03:30
cmoyates
ac4393fcb4
docs: clarify pos_embed parameter intent in HieraBackbone 
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 05:55:01 -03:30
cmoyates
cc2d31b435
feat(phase4d): shape + parity tests, fix attention transpose 
Fix MaskUnitAttention output transpose (0,2,3,1,4 → 0,3,2,1,4) to
match PyTorch transpose(1,3) token ordering for windowed attention.

Parity results (4 stages):
  Stage 0: max_abs 2.9e-4, mean 1.6e-5
  Stage 1: max_abs 1.3e-4, mean 1.3e-5
  Stage 2: max_abs 1.1e-2, mean 5.2e-5 (16 blocks, expected drift)
  Stage 3: max_abs 5.8e-4, mean 2.6e-5

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 05:51:36 -03:30
cmoyates
0d6d41cb45
feat(phase4c): weight loading + pos_embed bicubic interpolation 
HieraBackbone.load_checkpoint() loads safetensors, strips encoder.model.
prefix, bicubic-interpolates pos_embed from 512x512 to 128x128 tokens.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 05:48:40 -03:30
cmoyates
853efbe2cf
feat(phase4a): MLX Hiera backbone with unroll/reroll parity 
Port timm hiera_base_plus_224 to MLX: PatchEmbed, MaskUnitAttention,
HieraBlock, unroll/reroll, HieraBackbone. 4 NHWC feature maps at
strides 4/8/16/32. backbone.py now re-exports from hiera.py.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 05:45:43 -03:30
cmoyates
7315a082df
chore: add phase 4-6 prompts, fix naming, note zoxide cd issue 
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 05:20:40 -03:30
Cristopher Yates
150c974eed
feat(phase3): PyTorch→MLX weight converter (#1) 
* feat(phase3): PyTorch→MLX weight converter + safetensors output

365 keys mapped (367 source - 2 num_batches_tracked skipped).
Conv weights transposed (O,I,H,W→O,H,W,I), refiner stem remapped,
4ch patch embed preserved. 12 conversion tests + diagnostic report.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(phase3): conv allowlist, unused var, test fixture

- Replace _is_conv_weight heuristic with explicit CONV_WEIGHT_KEYS frozenset (15 keys)
- Remove unused `skipped` list from convert_state_dict
- Module-scoped pytest fixture eliminates 11 redundant checkpoint loads

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 05:16:56 -03:30
cmoyates
091261ccff
feat(phase2): MLX decoder + refiner with parity tests 
- MLP, DecoderHead (SegFormer-style), RefinerBlock, CNNRefinerModule
- All NHWC native, GroupNorm with pytorch_compatible=True
- Layout utils: NCHW<->NHWC, conv weight transpose (OIHW->OHWI)
- Dump script extended to export decoder/refiner weights
- Parity: decoder max_abs ~5e-6, refiner max_abs ~1e-5

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 05:00:00 -03:30
cmoyates
393dd0d5a7
chore: add remaining test stubs and script placeholders 
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 04:49:09 -03:30
cmoyates
9b07961619
feat(phase1): PyTorch reference harness + fixture tests 
Dump script loads CorridorKey checkpoint, runs deterministic forward
pass (seed=42, 512x512), saves 14 intermediate tensors to golden.npz.
Includes pos_embed bicubic interpolation for resolution mismatch.
32 shape/dtype/range validation tests all passing.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 04:49:01 -03:30
cmoyates
897242d200
init: repo scaffolding, deps, prompts, plan 
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-01 04:48:51 -03:30