Merge pull request #118 from mwalmsley/inigo_dev

Inigo dev

docs/guides/finetuning.rst

@@ -60,16 +60,16 @@ These files are called checkpoints (like video game save files - computer scient
     model = finetune.FinetuneableZoobotClassifier(
       name='hf_hub:mwalmsley/zoobot-encoder-convnext_nano',  # which pretrained model to download
       num_classes=2,
-      n_layers=0
+      n_blocks=0
     )
 
 You can see the list of pretrained models at :doc:`/pretrained_models`.
 
 What about the other arguments?
 When loading the checkpoint, FinetuneableZoobotClassifier will automatically change the head layer to suit a classification problem (hence, ``Classifier``).
 ``num_classes=2`` specifies how many classes we have, Here, two classes (a.k.a. binary classification).
-``n_layers=0`` specifies how many layers (other than the output layer) we want to finetune.
-0 indicates no other layers, so we will only be changing the weights of the output layer.
+``n_blocks=0`` specifies how many inner blocks (groups of layers, excluding the output layer) we want to finetune.
+0 indicates no other blocks, so we will only be changing the weights of the output layer.
 
 
 Prepare Galaxy Data

setup.py

@@ -76,8 +76,7 @@
             'albumentations',
             'pyro-ppl>=1.8.0',
             'torchmetrics==0.11.0',
-            'timm >= 0.9.10',
-            'galaxy_datasets == 0.0.17'
+            'timm >= 0.9.10'
         ],
         # TODO may add narval/Digital Research Canada config
         'tensorflow': [ # WARNING now deprecated
@@ -117,6 +116,6 @@
         'webdataset',  # for reading webdataset files
         'huggingface_hub',  # login may be required
         'setuptools',  # no longer pinned
-        'galaxy-datasets>=0.0.18'  # for dataset loading in both TF and Torch (see github/mwalmsley/galaxy-datasets)
+        'galaxy-datasets>=0.0.21'  # for dataset loading in both TF and Torch (see github/mwalmsley/galaxy-datasets)
     ]
 )