Skip to content

docs: fix the tutorial's checksum, install line and add a scale note #6

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

Merged
KalinNonchev merged 1 commit into from
Aug 3, 2026
Merged

docs: fix the tutorial's checksum, install line and add a scale note #6

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
20 changes: 15 additions & 5 deletions examples/predict_tcga_skcm.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,15 +30,20 @@
"tiles with mean pixel above 220, predicts the three genes for each tissue tile, and\n",
"overlays the predictions on the H&E.\n",
"\n",
"To give a sense of scale, this slide is 99,960 x 65,991 px, which is a 446 x 294\n",
"tile grid: 45,751 candidate tiles from the thumbnail pass, of which 36,392 survive\n",
"the exact background filter and are scored by the model.\n",
"\n",
"> The TCGA atlas predictions for melanoma were produced with a DeepSpot-M model\n",
"> finetuned on melanoma spatial transcriptomics. This tutorial uses the base\n",
"> released model in zero-shot mode, so the values are illustrative and the maps\n",
"> will look softer than the finetuned atlas.\n",
"\n",
"Install with `pip install deepspotm pyvips matplotlib`. pyvips needs the system\n",
"libvips with OpenSlide support. A GPU is recommended. The model weights are gated on\n",
"the Hugging Face Hub, so request access and log in first with `huggingface-cli login`.\n",
"Note that pyvips must be imported before torch."
"Install with `pip install \"deepspotm[wsi]\" matplotlib`, matching the `wsi` extra\n",
"documented in the README. pyvips needs the system libvips with OpenSlide support.\n",
"A GPU is recommended. The model weights are gated on the Hugging Face Hub, so\n",
"request access and log in first with `huggingface-cli login`. Note that pyvips\n",
"must be imported before torch."
]
},
{
Expand Down Expand Up @@ -107,8 +112,13 @@
" print(\"downloading ~1 GB slide from GDC ...\")\n",
" urllib.request.urlretrieve(url, SLIDE_PATH)\n",
"\n",
"# Hash in 8 MB chunks; reading the whole ~1 GB slide into memory to checksum it\n",
"# would spike RSS by more than the model itself uses.\n",
"digest = hashlib.md5()\n",
"with open(SLIDE_PATH, \"rb\") as fh:\n",
" md5 = hashlib.md5(fh.read()).hexdigest()\n",
" for chunk in iter(lambda: fh.read(8 << 20), b\"\"):\n",
" digest.update(chunk)\n",
"md5 = digest.hexdigest()\n",
"assert md5 == SLIDE_MD5, f\"checksum mismatch: {md5}\"\n",
"print(\"slide ready:\", SLIDE_PATH)"
]
Expand Down