Migrate to safetensors

README.md

@@ -33,7 +33,7 @@ FastNet has an encode-process-decode structure with a series of graph neural net
 ### Model Sources
 
 - **Inference-only repository:** [in preparation]
-- **Paper:** [main publication](https://arxiv.org/abs/2509.17601) and [technical paper](https://arxiv.org/abs/2509.17658)
+- **Paper:** [main publication](https://journals.ametsoc.org/view/journals/aies/5/3/AIES-D-25-0090.1.xml) and [technical paper](https://arxiv.org/abs/2509.17658)
 
 ## Uses
 
@@ -99,7 +99,7 @@ We also evaluate against the Met Office Global Model operational
 ## Computational Requirements
 
 ### Hardware
-For inference, FastNet-global runs on a single NVIDIA A100 GPU, though CPU inference is also supported. When using the GPU model file, we recommend a GPU with compute capability ≥ 8.0 (i.e. A100 or equivalent/higher).
+For inference, FastNet-global runs on a single NVIDIA A100 GPU, though CPU inference is also supported. When using the GPU, we recommend a GPU with compute capability ≥ 8.0 (i.e. A100 or equivalent/higher).
 
 ### Software
 

model_file → model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:338861dee35d800fef84c54cff282b7bfb7f1c73537ec49edad9aef83b0e4c79
-size 256322507
+oid sha256:73a8b701252004fb197a07b64cd26f718382c60448a290c8c5d158314ae62fef
+size 255860640

model_file_cpu

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:338861dee35d800fef84c54cff282b7bfb7f1c73537ec49edad9aef83b0e4c79
-size 256322507