Upload benchmarks/edit/run_traditional_metrics.py with huggingface_hub

Support held-out metadata val eval samples

Add eval sample builder for valN

Update sharded metric commands for resume script

Restore traditional metrics script with resume flush

Add 70x4 sharded metric commands

Batch GPU metric inference across rows

Add grouped smoke manifest sampler

Add worker prefetch support to traditional metrics

Add multi-candidate filtering for aesthetic RM pipeline

Add aesthetic reward model data pipeline

Resolve eval sources from dataset manifests

Support source maps for renamed source videos

Validate source paths in six-method manifest builder

Keep edited-only metrics when source videos are missing

Document manual metric weight downloads

Document NIQE cache setup

Set metric cache dirs by default

Make optional metric scorer init nonfatal

Add unified pretrained metric scorers

Add isolated metric conda environments

Document qalign dependency fixes

Update pretrained metric download commands

Update edit benchmark evaluation docs

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool