Skip to content

Fix flipped (height,width) in batched inference notebook #246

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Vision-Kek wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Fix flipped (height,width) in batched inference notebook #246

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
965a2c8
fix a) inference mode context b) datapoint.images[0].size is h,w not w,h
Nov 27, 2025
6abb7f8
del \n
Nov 27, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 6 additions & 5 deletions examples/sam3_image_batched_inference.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -81,7 +81,8 @@
"torch.autocast(\"cuda\", dtype=torch.bfloat16).__enter__()\n",
"\n",
"# inference mode for the whole notebook. Disable if you need gradients\n",
"torch.inference_mode().__enter__()\n"
"ctx = torch.inference_mode()\n",
"ctx.__enter__()"
]
},
{
Expand Down Expand Up @@ -165,7 +166,7 @@
" # In practice you're free to set any size you want, just edit the rest of the function\n",
" assert len(datapoint.images) == 1, \"please set the image first\"\n",
"\n",
" w, h = datapoint.images[0].size\n",
" h, w = datapoint.images[0].size\n",
" datapoint.find_queries.append(\n",
" FindQueryLoaded(\n",
" query_text=text_query,\n",
Expand All @@ -177,7 +178,7 @@
" coco_image_id=GLOBAL_COUNTER,\n",
" original_image_id=GLOBAL_COUNTER,\n",
" original_category_id=1,\n",
" original_size=[w, h],\n",
" original_size=[h, w],\n",
" object_id=0,\n",
" frame_index=0,\n",
" )\n",
Expand Down Expand Up @@ -208,7 +209,7 @@
" labels = torch.tensor(labels, dtype=torch.bool).view(-1)\n",
" if not labels.any().item() and text_prompt==\"visual\":\n",
" print(\"Warning: you provided no positive box, nor any text prompt. The prompt is ambiguous and the results will be undefined\")\n",
" w, h = datapoint.images[0].size\n",
" h, w = datapoint.images[0].size\n",
" datapoint.find_queries.append(\n",
" FindQueryLoaded(\n",
" query_text=text_prompt,\n",
Expand All @@ -222,7 +223,7 @@
" coco_image_id=GLOBAL_COUNTER,\n",
" original_image_id=GLOBAL_COUNTER,\n",
" original_category_id=1,\n",
" original_size=[w, h],\n",
" original_size=[h, w],\n",
" object_id=0,\n",
" frame_index=0,\n",
" )\n",
Expand Down