From 66af02f1e5c50e041ced3e8b85659709a1c1f4f3 Mon Sep 17 00:00:00 2001 From: YurenHao0426 Date: Sat, 1 Aug 2026 16:23:26 -0500 Subject: Record the gate correction in the concept document The user-facing statement still carried the retired correlation threshold. Adds the correction, the joint condition that replaces it, the width diagnosis, and the closure of the shrink-N route. Co-Authored-By: Claude --- logs/obj_base_gpu.log | 4 +++- logs/obj_large_gpu.log | 3 +++ logs/rank_gate.log | 0 logs/rho_full_width.log | 1 + 4 files changed, 7 insertions(+), 1 deletion(-) create mode 100644 logs/obj_large_gpu.log create mode 100644 logs/rank_gate.log (limited to 'logs') diff --git a/logs/obj_base_gpu.log b/logs/obj_base_gpu.log index c3ea786..3fccb88 100644 --- a/logs/obj_base_gpu.log +++ b/logs/obj_base_gpu.log @@ -1,3 +1,5 @@ Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`. `torch_dtype` is deprecated! Use `dtype` instead! - segment: 0%| | 0/250 [00:00