Spaces:
Paused
Paused
Commit History
fix: Downgrade transformers to 4.51.3 to fix grounding tool 37bdbfa
fix: Load grounding model without dtype parameter to avoid errors e7a5afc
fix: Update grounding tool for transformers 4.56.0 and fix DICOM segmentation 99d91bc
fix: Don't invert processed DICOM files for segmentation 8bb9754
fix: Upgrade transformers to 4.56.0 to fix NV-Reason-CXR dtype error 6a7c30f
fix: Load NV-Reason-CXR without dtype parameter to avoid JSON error 7a6a9a6
fix: Use 'dtype' instead of deprecated 'torch_dtype' for NV-Reason-CXR f49ba8b
fix: Invert PNG/JPEG X-rays before HU conversion for accurate segmentation 85e170b
debug: Add verbose logging to NV-Reason-CXR loading e88cf5f
fix: Match NVIDIA's official NV-Reason-CXR loading approach c1c4dbf
fix: Remove quantization parameters from NV-Reason-CXR to fix dict.to_dict() error 90107a8
refactor: Simplify segmentation to single overlay panel 0e36bfb
fix: Use correct Hounsfield Unit range for torchxrayvision segmentation f6108da
debug: Add comprehensive logging for segmentation pipeline 940e3ae
fix: Add 3-panel diagnostic visualization for segmentation 6d677d9
fix: Lower segmentation threshold and add probability debugging 7f44ee5
fix: Enhanced segmentation overlay with contours and better alignment 89eb957
fix: Improve segmentation overlay and add DICOM support 3a18164
Fix NV-Reason-CXR quantization config error f1877c8
Fix MAIRA-2 grounding tool - remove BitsAndBytesConfig 39ab112
Fix remaining tool loading issues 55abc5d
Fix multiple tool loading issues 4df7c2f
Fix tool loading errors in MedRAX2 b8918e3
Add NVIDIA NV-Reason-CXR tool for expert chest X-ray analysis 27f1dea
Use backwards-compatible langchain.chains import for RetrievalQA d32145b
Fix RetrievalQA import to use langchain_community.chains f65bee4
Fix all deprecated langchain imports to use new module structure 24627d1
Merge branch 'main' into tool-changes e4e9fae
VictorLJZ commited on
fixes d38594f
Junzhe Li commited on
tool changes b8b11e0
Junzhe Li commited on
CLI/env/SLURM-aware resolver 0ad1012
Emily Xie commited on
Configurable MedGemma API URL and auto-pick free loopback IP/port c4d7934
Emily Xie commited on
Fix merge conflicts 128b355
Adibvafa commited on
Improve style 7393de8
Adibvafa commited on
Fix merge conflicts 5b96cf3
Adibvafa commited on
Add parallel tool calling 77eeeb6
Adibvafa commited on
fixed rexvqa benchmark and added handling for image norm for tools 16278b5
victorli commited on
MedGemma fixes b2aba7d
Emily Xie commited on
Fix env loading of MedGemma c34de72
Adibvafa commited on
Add env loading for HF d280bf3
Adibvafa commited on
fixed medgemma 7b3e756
victorli commited on
fixed up some things and corrected agent init logic fbcbf94
victorli commited on
Merge branch 'main' into emily/medgemma 4c987e5
VictorLJZ commited on
Delete medrax/tools/vqa/medgemma/medgemma_requirements.txt c94d452
Emily Xie commited on
standard req medgemma env 547624d
Emily Xie commited on
Update medgemma.py 1dd8f9f
Emily Xie commited on
Update medgemma_client.py b3d2946
Emily Xie commited on
fix bug d51b788
Emily Xie commited on
medgemma fastapi tool integration 205758b
Emily Xie commited on