Initial commit

README.md

@@ -1,4 +1,19 @@
-📢 [2026-04-10] The pretrained SPECTRE can now be loaded directly from the transformers library. Check below for the details.
+---
+license: cc-by-nc-sa-4.0
+language:
+- en
+tags:
+- medical-imaging
+- ct-scan
+- 3d
+- vision-transformer
+- self-supervised-learning
+- foundation-model
+- radiology
+library_name: transformers
+pipeline_tag: feature-extraction---
+
+📢 [2026-05-20] The pretrained SPECTRE can now be loaded directly from the `transformers` library. Check below for the details.
 
 📢 [2026-04-10] SPECTRE is now an official baseline for the [**CVPR 2026 Workshop Competition: Foundation Models for General CT Image Diagnosis**](https://www.codabench.org/competitions/12650/)! See `experiments/cvpr26_fm_for_ct_diag_task_1` for scripts and additional details.  
 
@@ -29,14 +44,25 @@ SPECTRE has been trained on a large cohort of **open-source CT scans** of the **
 This repository provides pretrained SPECTRE models together with tools for fine-tuning and evaluation.
 
 ## 🧠 Pretrained Models
-The pretrained SPECTRE model can easily be imported as follows:
+The pretrained SPECTRE model can easily be imported using `AutoModel` from the `transformers` library
 
 ```python
-from spectre import SpectreImageFeatureExtractor, MODEL_CONFIGS
-import torch
+from transformers import AutoModel
+model = AutoModel.from_pretrained('cclaess/SPECTRE-Large', trust_remote_code=True)
+```
 
+or by using the `spectre-fm` package as follows:
+
+```python
+from spectre import SpectreImageFeatureExtractor, MODEL_CONFIGS
 config = MODEL_CONFIGS['spectre-large-pretrained']
 model = SpectreImageFeatureExtractor.from_config(config)
+```
+
+A simple forward pass would look like:
+```python
+import torch
+
 model.eval()
 
 # Dummy input: (batch, crops, channels, height, width, depth)

metadata.yaml

@@ -0,0 +1,13 @@
+license: cc-by-nc-sa-4.0
+language:
+- en
+tags:
+- medical-imaging
+- ct-scan
+- 3d
+- vision-transformer
+- self-supervised-learning
+- foundation-model
+- radiology
+library_name: transformers
+pipeline_tag: feature-extraction

modeling_spectre.py

@@ -30,7 +30,7 @@ class SpectreModel(PreTrainedModel):
         self,
         pixel_values: torch.Tensor,
         grid_size=None,
-        return_dict=True,
+        return_dict=False,
         **kwargs,
     ):
         outputs = self.model(pixel_values, grid_size=grid_size)