Set fastMRI AXT1 inference batch_size to 10

Set Calgary inference batch_size to 10

Organize packs into calgary_campinas/ and fastmri_axt1/; drop previous-hosting note

Add CVPR Calgary comparison VarNet weights (iter 8500, avg-best SSIM 5x/10x)

Document Calgary comparison varnet baseline

Add CVPR Calgary comparison VarNet inference YAML (filters=64)

Clarify fastMRI AXT1 packs are from paper appendix (arXiv:2111.09639)

Remove old fastmri_axt1_recurrentvarnet_ablation assets

Document AXT1 RecurrentVarNet t6_h128_nl2 naming

Add renamed AXT1 RecurrentVarNet t6_h128_nl2 checkpoint

Rename AXT1 RecurrentVarNet ablation to t6_h128_nl2 (T=6, 128 ch, nl=2)

Upload fastmri_axt1_varnet.yaml with huggingface_hub

Upload fastmri_axt1_unet.yaml with huggingface_hub

Upload fastmri_axt1_rim.yaml with huggingface_hub

Upload fastmri_axt1_recurrentvarnet_ablation.yaml with huggingface_hub

Upload fastmri_axt1_recurrentvarnet.yaml with huggingface_hub

Upload fastmri_axt1_lpd.yaml with huggingface_hub

Add fastMRI AXT1 CVPR packs (remapped weights; 4x default, commented 8x)

Rewrite CVPR checkpoints to ModConv Unet key layout (sensitivity/unet/xpdnet)

Add crop_outer_slices: true and commented 10x accel to Calgary inference YAMLs

Docs: install from GitHub via conda + editable .[dev], not PyPI

Add DIRECT JOSS citation to model card

Rewrite model card: clearer docs, dataset links, no checkpoint iters

Docs: training protocol and single-R inference masking note

Replace configs with minimal inference-only YAMLs (no training/validation)

Add CVPR 2022 RecurrentVarNet baselines (name.yaml + name.pt) for fastMRI

Fix direct predict usage: output_directory is positional

initial commit