From 67e4b12246f4024b0022227e3c7ecd06152cf2fd Mon Sep 17 00:00:00 2001 From: Daniel Maddern Date: Wed, 19 Aug 2026 22:04:12 +0700 Subject: [PATCH] Make process_image return 2D grid_thw [[t,h,w]] for multi-image support --- src/h3_blackwell_runtime/qwen3vl_vision.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/h3_blackwell_runtime/qwen3vl_vision.py b/src/h3_blackwell_runtime/qwen3vl_vision.py index 7fd2fcb..97cd465 100644 --- a/src/h3_blackwell_runtime/qwen3vl_vision.py +++ b/src/h3_blackwell_runtime/qwen3vl_vision.py @@ -103,7 +103,7 @@ def process_image( grid_h = h_bar // patch_size grid_w = w_bar // patch_size - grid_thw = torch.tensor([1, grid_h, grid_w], device=device, dtype=torch.long) + grid_thw = torch.tensor([[1, grid_h, grid_w]], device=device, dtype=torch.long) pixel_values = normalized channel = pixel_values.shape[0]