kaiko-eva 0.3.3__py3-none-any.whl → 0.4.0__py3-none-any.whl

eva/vision/data/transforms/spatial/resize.py

@@ -0,0 +1,62 @@
+"""Image resize transforms."""
+
+import functools
+from typing import Any, Dict
+
+from torchvision import tv_tensors
+from torchvision.transforms import v2
+from typing_extensions import override
+
+from eva.vision.data.transforms.spatial import functional
+
+
+class Resize(v2.Transform):
+    """Resize transform for images with spatial or byte-based constraints.
+
+    This transform provides two mutually exclusive modes of resizing:
+    1. Spatial resizing: Resize to a specific (height, width) dimension
+    2. Byte-based resizing: Resize to fit within a maximum byte size
+
+    The latter is particularly useful for API models (e.g. Claude 3.7) that
+    have strict byte size limits for image inputs.
+    """
+
+    def __init__(self, size: tuple[int, int] | None = None, max_bytes: int | None = None) -> None:
+        """Initializes the transform.
+
+        Args:
+            size: Target size as (height, width) tuple for spatial resizing.
+                If provided, max_bytes must be None.
+            max_bytes: Maximum allowed byte size for the image.
+                If provided, size must be None. Must be a positive integer.
+
+        Raises:
+            ValueError: If both size and max_bytes are provided, or if max_bytes
+                is not a positive integer.
+        """
+        if size is not None and max_bytes is not None:
+            raise ValueError("Cannot provide both 'size' and 'max_bytes' parameters.")
+        if max_bytes is not None and max_bytes <= 0:
+            raise ValueError("'max_bytes' must be a positive integer.")
+
+        super().__init__()
+
+        self.size = size
+        self.max_bytes = max_bytes
+        self.resize_fn = None
+
+        if size is not None:
+            self.resize_fn = v2.Resize(size=size)
+        elif max_bytes is not None:
+            self.resize_fn = functools.partial(functional.resize_to_max_bytes, max_bytes=max_bytes)
+
+    @functools.singledispatchmethod
+    @override
+    def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
+        return inpt
+
+    @_transform.register(tv_tensors.Image)
+    @_transform.register(tv_tensors.Mask)
+    def _(self, inpt: Any, params: Dict[str, Any]) -> Any:
+        inpt_resized = self.resize_fn(inpt) if self.resize_fn is not None else inpt
+        return tv_tensors.wrap(inpt_resized, like=inpt)

eva/vision/models/wrappers/from_registry.py

@@ -3,6 +3,7 @@
 from typing import Any, Callable, Dict
 
 import torch
+from torch import nn
 from typing_extensions import override
 
 from eva.core.models.wrappers import base
@@ -40,14 +41,14 @@ class ModelFromRegistry(base.BaseModel[torch.Tensor, torch.Tensor]):
         self._model_kwargs = model_kwargs or {}
         self._model_extra_kwargs = model_extra_kwargs or {}
 
-        self.load_model()
+        self.model = self.load_model()
 
     @override
-    def load_model(self) -> None:
-        self._model = factory.ModuleFactory(
+    def load_model(self) -> nn.Module:
+        ModelFromRegistry.__name__ = self._model_name
+
+        return factory.ModuleFactory(
             registry=backbone_registry,
             name=self._model_name,
             init_args=self._model_kwargs | self._model_extra_kwargs,
         )
-
-        ModelFromRegistry.__name__ = self._model_name

eva/vision/models/wrappers/from_timm.py

@@ -5,6 +5,7 @@ from urllib import parse
 
 import timm
 import torch
+from torch import nn
 from typing_extensions import override
 
 from eva.core.models.wrappers import base
@@ -46,12 +47,14 @@ class TimmModel(base.BaseModel[torch.Tensor, torch.Tensor]):
         self._out_indices = out_indices
         self._model_kwargs = model_kwargs or {}
 
-        self.load_model()
+        self.model = self.load_model()
 
     @override
-    def load_model(self) -> None:
+    def load_model(self) -> nn.Module:
         """Builds and loads the timm model as feature extractor."""
-        self._model = timm.create_model(
+        TimmModel.__name__ = self._model_name
+
+        return timm.create_model(
             model_name=self._model_name,
             pretrained=True if self._checkpoint_path else self._pretrained,
             pretrained_cfg=self._pretrained_cfg,
@@ -59,7 +62,6 @@ class TimmModel(base.BaseModel[torch.Tensor, torch.Tensor]):
             features_only=self._out_indices is not None,
             **self._model_kwargs,
         )
-        TimmModel.__name__ = self._model_name
 
     @property
     def _pretrained_cfg(self) -> Dict[str, Any]:

{kaiko_eva-0.3.3.dist-info → kaiko_eva-0.4.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: kaiko-eva
-Version: 0.3.3
+Version: 0.4.0
 Summary: Evaluation Framework for oncology foundation models.
 Keywords: machine-learning,evaluation-framework,oncology,foundation-models
 Author-Email: Ioannis Gatopoulos <ioannis@kaiko.ai>, =?utf-8?q?Nicolas_K=C3=A4nzig?= <nicolas@kaiko.ai>, Roman Moser <roman@kaiko.ai>
@@ -245,6 +245,10 @@ Requires-Dist: einops>=0.8.1; extra == "vision"
 Provides-Extra: language
 Requires-Dist: datasets<4.0.0,>=2.19.0; extra == "language"
 Requires-Dist: litellm>=1.61.8; extra == "language"
+Requires-Dist: backoff>=2.2.1; extra == "language"
+Provides-Extra: multimodal
+Requires-Dist: litellm>=1.61.8; extra == "multimodal"
+Requires-Dist: backoff>=2.2.1; extra == "multimodal"
 Provides-Extra: all
 Requires-Dist: h5py>=3.10.0; extra == "all"
 Requires-Dist: nibabel>=4.0.1; extra == "all"
@@ -260,6 +264,7 @@ Requires-Dist: monai>=1.3.2; extra == "all"
 Requires-Dist: einops>=0.8.1; extra == "all"
 Requires-Dist: datasets<4.0.0,>=2.19.0; extra == "all"
 Requires-Dist: litellm>=1.61.8; extra == "all"
+Requires-Dist: backoff>=2.2.1; extra == "all"
 Description-Content-Type: text/markdown
 
 <div align="center">
@@ -298,7 +303,7 @@ Check out the [documentation](https://kaiko-ai.github.io/eva/) for more informat
 
 ### Highlights:
 - Easy and reliable benchmark of Oncology FMs
-- Supports patch-level classification, slide-level classification, semantic segmentation, and text classification downstream tasks
+- Supports patch-level classification, slide-level classification, semantic segmentation, and (visual) question answering tasks.
 - Automatic embedding inference and evaluation of a downstream task
 - Native support of popular medical [datasets](https://kaiko-ai.github.io/eva/dev/datasets/) and models
 - Produce statistics over multiple evaluation fits and multiple metrics
@@ -316,6 +321,9 @@ pip install 'kaiko-eva[vision]'
 # to install the expanded `language` version
 pip install 'kaiko-eva[language]'
 
+# to install the expanded `multimodal` version
+pip install 'kaiko-eva[multimodal]'
+
 # to install everything
 pip install 'kaiko-eva[all]'
 ```

{kaiko_eva-0.3.3.dist-info → kaiko_eva-0.4.0.dist-info}/RECORD

@@ -3,7 +3,7 @@ eva/__main__.py,sha256=kM5tQ0egTuBWixNLLx9QU-PpS2Bbs3zE3nYE6b2vWa0,282
 eva/__version__.py,sha256=YFR4oOlvPg0sS4Ni7GJ_vU42VTs5WiWp6odK7yH4TBY,611
 eva/core/__init__.py,sha256=AYlMZcH76B7I1lOa-E67u2o9DxsCwI4JMLCYXLk9oDQ,451
 eva/core/callbacks/__init__.py,sha256=_XsS4QGf4r9frzFwEx-krmzJm3kbLmXQFtwYCfRyjrM,281
-eva/core/callbacks/config.py,sha256=-DRt20a2aF9Z9-7nZvbGBcOZ30qNf3ESf25EPRgRL1w,4267
+eva/core/callbacks/config.py,sha256=X5d-Ex_1irOLtyzSbB4t3BjfctqtJXsIdHEbJUISmDE,4337
 eva/core/callbacks/writers/__init__.py,sha256=z8cNVJOanj-yYyIiX-mOkhur1NExuCOKzPjp4mmm3AE,232
 eva/core/callbacks/writers/embeddings/__init__.py,sha256=zMxP4POf1yFFUVSjGcsQgyPYNY6JsZO_F66ngKJZaG8,301
 eva/core/callbacks/writers/embeddings/_manifest.py,sha256=pB9nGv9ofVbnI4HHPrHY4p7jeFiFQTRc09zszto_DTk,2424
@@ -14,11 +14,9 @@ eva/core/callbacks/writers/embeddings/typings.py,sha256=qXZSlasaSKHad6HyJeRTeUv-
 eva/core/cli/__init__.py,sha256=1lGiomn4JINI0DKy41_D4cEyyH-hN6cfTZfMPxLxTCA,68
 eva/core/cli/cli.py,sha256=AZ4B4OP3D2af9H2RYBd5nxoy5I9DlaClZaadSWpPEPI,422
 eva/core/cli/logo.py,sha256=x6-vGWI0s9gza-xxQrBDi2wneb2wFU_mQGHgpAiq2MQ,786
-eva/core/cli/setup.py,sha256=kR-7l4X5Hu8kSLoQZGYGIeLXtn9S_EU52dauDy6fm0w,2663
+eva/core/cli/setup.py,sha256=q5bLSqlHvR0bgF-uh4tL1JOv2aCn5hoRP-XqCaBKTvo,2690
 eva/core/data/__init__.py,sha256=yG3BeOWhp1EjVYMFqx8M_TBWFDyfIwwksQGQmMdSPaI,340
-eva/core/data/dataloaders/__init__.py,sha256=0AvpsPOdh4NX5rwkX9Th1M_rzxZVVzTPTdC5oTGFd5w,194
-eva/core/data/dataloaders/collate_fn/__init__.py,sha256=CfSAVrPD36shpyYAkob2ny05VWymb95MutawQcZkbuo,134
-eva/core/data/dataloaders/collate_fn/collate.py,sha256=oBdxaDCIaXBB6H8LB90Qsi2Inw1tyHGF4kAFBINPOeQ,689
+eva/core/data/dataloaders/__init__.py,sha256=fbNClVZ8J3QoGi4qiPq635ig1j9GdI7six3RhfwDbjY,110
 eva/core/data/dataloaders/dataloader.py,sha256=v_UL5p78EiyLqXUhWsS0EbfW0iO-DAWOMv7-WTclOvA,2750
 eva/core/data/datamodules/__init__.py,sha256=qZchYbgxo9lxYnGoqdk0C6MfS2IbF0WItO0kCdP9Mqc,229
 eva/core/data/datamodules/call.py,sha256=jjj9w3UXYuQB-qyCcw1EZpRJW10OC1I3dvgvsuQWLck,940
@@ -35,7 +33,7 @@ eva/core/data/datasets/typings.py,sha256=KSmckjsU64pGV-8uSLkD1HmvPKYlyypngiRx9yy
 eva/core/data/samplers/__init__.py,sha256=rRrKtg4l6YoziD3M0MkctQvX1NdRxaQa5sm6RHH_jXc,315
 eva/core/data/samplers/classification/__init__.py,sha256=gvv7BH4lG9JlkMaTOnaL0f4k1ghiVBgrH64bh1-rreQ,147
 eva/core/data/samplers/classification/balanced.py,sha256=MGTHt-WQaQKiJ5A1D_P6HJ6YzPTD-ERhc0R7rNMFqfg,3788
-eva/core/data/samplers/random.py,sha256=znl0Z9a-X-3attP-EH9jwwo83n40FXW_JzOLNZAml_c,1252
+eva/core/data/samplers/random.py,sha256=fAlPmlTiXO7QqTh2hLVjgmELsX4P7DrS-9vV2qMAcOQ,1409
 eva/core/data/samplers/sampler.py,sha256=0DOLUzFoweqEubuO1A4bZBRU0AWFoWGWrO3pawRT-eI,877
 eva/core/data/splitting/__init__.py,sha256=VQJ8lfakbv6y2kAk3VDtITAvh7kcZo3H1JwJBc5jT08,198
 eva/core/data/splitting/random.py,sha256=r6iy7j34seRTlyB79_Xy7m6lsKRi8ZM9X5Ln1b-SBjg,1453
@@ -48,7 +46,7 @@ eva/core/data/transforms/padding/pad_2d_tensor.py,sha256=J4maGFmeQf9IHRxt5kU-6eI
 eva/core/data/transforms/sampling/__init__.py,sha256=BFKbvRjlZrwS0GcNrM54ZSWt6PrQARfFlXM1jJ-wpvo,149
 eva/core/data/transforms/sampling/sample_from_axis.py,sha256=Zbhp94lVa70WQKmSOKMTsOMe2c7wLqNZto7JqWhSdtI,1229
 eva/core/interface/__init__.py,sha256=chdpKXipxe1NP-Fgr_d9r6X1gMna0XiEa38waJ6FzTM,98
-eva/core/interface/interface.py,sha256=EeBrIqUlO497BOZGZrtO-fZnc_BhIJrrqyQmRfqWXcw,4591
+eva/core/interface/interface.py,sha256=pni92cLb4KQMSvOoYMWP2qIn8rtzqxj84hEhjsdCgis,5364
 eva/core/loggers/__init__.py,sha256=4YMLNlN9LnuKqhBI1R1keh69dmMD-2lcH3HKwwyn380,266
 eva/core/loggers/dummy.py,sha256=Y7ypH0ecSAIkkZ5LzTmNNEzlKkqeaHfUNMCDKVOg6D4,1204
 eva/core/loggers/experimental_loggers.py,sha256=p5uCK_9QCYufRhE-LZQUJWbhGElyobX_zRM78yX4p2o,230
@@ -78,7 +76,7 @@ eva/core/models/__init__.py,sha256=T6Fo886LxMj-Y58_ylzkPkFSnFR2aISiMIbuO_weC4s,4
 eva/core/models/modules/__init__.py,sha256=_4VjN9Qs4_mftEVgt0KqVxCRAtDcKYbUnBA3Ox1Pzis,350
 eva/core/models/modules/head.py,sha256=bZ45RBPi3N8sjvKyt2_TeKWI1eB6GyBeGzV6J11ERO8,5225
 eva/core/models/modules/inference.py,sha256=ih-0Rr2oNf2N6maiXPOW7XH5KVwUT1_MOxnJKOhJ1uQ,978
-eva/core/models/modules/module.py,sha256=LtjYxTZb7UY0owonmt_yQ5EySw3sX-xD9HLN2io8EK4,6697
+eva/core/models/modules/module.py,sha256=8BHc_NdIXAyA7oUP7LFLYvMjIJSsuR5lvFiHS_wycU4,6695
 eva/core/models/modules/scheduler.py,sha256=orv5TDq9vEItxE7rTj0sQlzdChpx5jHx-3ypnHjUyr4,1650
 eva/core/models/modules/typings.py,sha256=LPR8JdIid2gJZpjMG1FcH5OZ60JlFOj_LupIh__2k_8,803
 eva/core/models/modules/utils/__init__.py,sha256=ScLCHwQfzlg_UsHVi5sf_SavUkh9secwtRn_umC_qA8,325
@@ -93,17 +91,17 @@ eva/core/models/transforms/extract_cls_features.py,sha256=A-oNNiGYBkMLK_E9DrzJUJ
 eva/core/models/transforms/extract_patch_features.py,sha256=5UsHjZnh3Uk2jAfi3GLtxoJ14ALTrfVoWUy8nTspRDY,2218
 eva/core/models/wrappers/__init__.py,sha256=jaiANQdbO-IPgH8U-Y0ftFsuuCAM5i5KuYRHauKw5k8,450
 eva/core/models/wrappers/_utils.py,sha256=ZWe9Ih_0kH5Wg_AQAtAn77LZ_CODAve5u3G12ifLNsc,4902
-eva/core/models/wrappers/base.py,sha256=S3LGbb_wo2ZTeDvoLvSlaymBsRN6UAuKrCmVAGasHWY,1452
-eva/core/models/wrappers/from_function.py,sha256=LVnfW3bL1mRtO8kL1D0MLJOiAiyiHkM7RX6-JIJzI0E,1844
-eva/core/models/wrappers/from_torchhub.py,sha256=-saKMxYq4KvVpaJL6BdHmUcEYHuaNvm2-0O_eQMg3GA,3250
-eva/core/models/wrappers/huggingface.py,sha256=-_fA81YRnoMc7O7SbrnCEj1dM_xArpQ8WdsZk9z_f1E,1449
-eva/core/models/wrappers/onnx.py,sha256=34li_xSwPryN8nJDrFyif_Hve1AEH7Ry9E_lZmf7JJM,1834
+eva/core/models/wrappers/base.py,sha256=MvFZcwXvWsgvi9LqJUcUKjqimyFXvJ7P5ftOaNQpecc,1450
+eva/core/models/wrappers/from_function.py,sha256=ccjCMwkdmyKqCxUbNaUDB1V6KvBgKuoXxIb8gH7Giyo,1855
+eva/core/models/wrappers/from_torchhub.py,sha256=VHOigAIkYJfMfaq8WJwYzbex-1ny_ZjszxGSfaDeVbw,3291
+eva/core/models/wrappers/huggingface.py,sha256=rP0DmdZnJ0Zqjy3o7BqwBa8uG3C24Epgb4h1sRPZOJ0,1459
+eva/core/models/wrappers/onnx.py,sha256=BhXlXE0mF3pZ2R8SZtDUF51ooFwdsAjJ47CiJg-wjNY,1837
 eva/core/trainers/__init__.py,sha256=jhsKJF7HAae7EOiG3gKIAHH_h3dZlTE2JRcCHJmOzJc,208
 eva/core/trainers/_logging.py,sha256=gi4FqPy2GuVmh0WZY6mYwF7zMPvnoFA050B0XdCP6PU,2571
 eva/core/trainers/_recorder.py,sha256=M-BJHLgqGxR_MSV6f_WC7GN2JHYEEinV1-hNLpH667A,8062
 eva/core/trainers/_utils.py,sha256=M3h8lVhUmkeSiEXpX9hRdMvThGFCnTP15gv-hd1CZkc,321
 eva/core/trainers/functional.py,sha256=_Mw-NIPU2tPffxpK5t3sHBmVI6u163phCpoJFiauH7E,4583
-eva/core/trainers/trainer.py,sha256=a3OwLWOZKDqxayrd0ugUmxJKyQx6XDb4GHtdL8-AEV0,4826
+eva/core/trainers/trainer.py,sha256=i_Qj76h6BfwK-5Y3UsiLcWrSnUKumg0XLxBdAR2WRHU,4908
 eva/core/utils/__init__.py,sha256=cndVBvtYxEW7hykH39GCNVI86zkXNn8Lw2A0sUJHS04,237
 eva/core/utils/clone.py,sha256=qcThZOuAs1cs0uV3BL5eKeM2VIBjuRPBe1t-NiUFM5Y,569
 eva/core/utils/factory.py,sha256=upWPWkWW7p3_ZoePAzbkkZvEPiAY4OaOQkwNvh0oa-E,2397
@@ -119,23 +117,79 @@ eva/core/utils/registry.py,sha256=iEDVsQ7DdPXhzfW32z3lWNgwjagR_jh_g4VkvajjRK4,13
 eva/core/utils/requirements.py,sha256=u01QPnBDu0YApncxgvg7xEANubyGSptqja12cycg3k0,909
 eva/core/utils/suppress_logs.py,sha256=pOk1076J0mKWn2lgDqEVC1g65FXhA_2IkC4LBEWhnwQ,902
 eva/core/utils/workers.py,sha256=hfx63M82qNg0Dwhre2tl53MnhtRsV7APaDONM9nhVB8,634
-eva/language/__init__.py,sha256=EQxtGM6zT3y-6EGecGAo2IXKS6J8dyWNOGP2mvD7fZ0,360
+eva/language/__init__.py,sha256=W_u2wedmQvEOWVEw8dmkie7c6ATeVuTVn9_Yw7uc3cU,406
+eva/language/callbacks/__init__.py,sha256=gVTSGjKYvovTgERVJmQLssW6Fz7aJwjYpe6gqUtq0_8,131
+eva/language/callbacks/writers/__init__.py,sha256=FV02MK218j3pxCgmTvMPtgNL8-4d2NbUcnyvsZI0CNs,150
+eva/language/callbacks/writers/prediction.py,sha256=MzU8Jc17fwL2SV0tjLl-ravxN9GPDAWARZdT2cr86FY,6894
 eva/language/data/__init__.py,sha256=ONF-CLBK2HO2p1bMCAbdvaHfSsYjdSE9-O3lWgwQQGk,89
-eva/language/data/datasets/__init__.py,sha256=AeGw3EVyz5w8Ja2m8Hd3FoKkCeG68M0BwtF64VCvOk4,211
+eva/language/data/dataloaders/__init__.py,sha256=6pGDiohDxfNepOyE76XhgQMOYzgL1LYNaR9Qu7xb8fQ,169
+eva/language/data/dataloaders/collate_fn/__init__.py,sha256=QZ22JjL4t8VldudGXTaEPv3rH8XKta_nlip1TszWn9g,171
+eva/language/data/dataloaders/collate_fn/text.py,sha256=BrgbgUOFyaitymVAp2lZNaS2kqU7KzmeMOKfNzGaiYY,1944
+eva/language/data/datasets/__init__.py,sha256=FE5tpQ18z2yRE_q60r2xu7WapLT3oGt5d_iZ6_uJ_DI,308
+eva/language/data/datasets/base.py,sha256=BDpyT_mnYuP5qMlGP_vfRejO-gp-bsQYNMBqM0EmXyY,309
 eva/language/data/datasets/classification/__init__.py,sha256=HXtgJpMt9CgqvdeWc147bOn-9YeUMcL8v-0UHHkY_A4,144
-eva/language/data/datasets/classification/base.py,sha256=53Tbu8dRSMuNfRB_xEBk_3wSe0jRqd_S-IwE9WIpgW0,1669
-eva/language/data/datasets/classification/pubmedqa.py,sha256=AxqW5OQAtDBLVX-Sr0WIt3E7GBbkGMUPOvhOlKLTLh8,5439
-eva/language/data/datasets/language.py,sha256=Kjy1X9xZBIfjjEYJKMikgKhLaquugRSjXTYM_HWiHmw,305
-eva/language/models/__init__.py,sha256=ggL8uDv_J8vQTr5fc5j8UENoPA24h2Bo3uNSdd_vRU0,606
-eva/language/models/modules/__init__.py,sha256=UWO_yq7kPzTrptl6iWTFy07E_Cyzf-sTu-zAK-sN3CI,112
-eva/language/models/modules/text.py,sha256=wknEnS_t8IToLTo_F9AG4EXkt0_6o4n6dYm-8AVxgOY,2906
+eva/language/data/datasets/classification/base.py,sha256=Xdy8quqx4lSZtS04yJ1WubXREMEnaB7BiCnn-Ao8wIQ,589
+eva/language/data/datasets/classification/pubmedqa.py,sha256=n3CSwVbQAPU7NJ3m8WeYF1coNRLOiolVyvqVZIFfLg4,6510
+eva/language/data/datasets/prediction.py,sha256=4ZyDFK5IWPuVShKy8z2Bn6Uq0dYgolFhf94EUnzYiC8,5880
+eva/language/data/datasets/schemas.py,sha256=x0IBinezADXzQr71SDFSSeRYXW7uuINmzdPQLemWF-Y,405
+eva/language/data/datasets/text.py,sha256=MpVJa5h7Al572UdU-WtfKwbF7nEAt_jd11rJF3ZIuQk,2778
+eva/language/data/datasets/typings.py,sha256=_JZUD6ZlaCQhvWBD851zQZy3czgCxpsAnbnifSvSRCk,904
+eva/language/data/messages.py,sha256=2x7k-rzZkM8Ow-fIm6cBrsJB6W_vlsjExs6IaQkdDa4,1306
+eva/language/models/__init__.py,sha256=QZHt1uz9bB33Ko5mznlLjY0uGCVmnS7SOLb8Z0Bzgm4,725
+eva/language/models/modules/__init__.py,sha256=Dk2pdjkerDA_Omrue38ia0rYmW84jRvyc9pPKziZH9c,172
+eva/language/models/modules/language.py,sha256=ybDHNDtJ_YCKq_zoXHZorF0rOM1l4jz_URZ2cfdSSok,3097
 eva/language/models/modules/typings.py,sha256=dWq3FTbvGZGoIMTEIX1yWBk_VFKIHsvxaTvBXlbJfCA,379
-eva/language/models/wrappers/__init__.py,sha256=syKHBqX3MRD2oq_MjDjvoNC6W0F8TdmGYqhwrGLBahI,402
-eva/language/models/wrappers/huggingface.py,sha256=l3wVOB4QOe2_yeG4xwNm0imP_kWbjM8RG0ISE4ANAyE,2396
-eva/language/models/wrappers/litellm.py,sha256=0GSo3gxUv1Ah6FQ6VqTiiu_0TVx-h4xUkZqdOAGbKTM,2619
-eva/language/models/wrappers/vllm.py,sha256=LXadfveGSHmjWAlOBrJm1QHkKzB_O-s3M8U7bnKfW-E,5684
-eva/language/utils/__init__.py,sha256=jqrU3o80NAwt6HbNDWSM3e3B1-65vw811DqiUWT0Bys,150
-eva/language/utils/str_to_int_tensor.py,sha256=YlNppLlAoX0ZRVfcDX4BStCQPbS1NNParcwptxH0_wE,3143
+eva/language/models/networks/__init__.py,sha256=zrRaiFqcIeFsLm6-yJWUeEgh_1xvBY-5QAdRGjEOysg,374
+eva/language/models/networks/alibaba.py,sha256=kSGPdZWnn0q71Dm03HKmoXWACeEyc8V3HVLuk_fIhno,798
+eva/language/models/networks/api/__init__.py,sha256=NrLiFExfplzbiFqOTgQqSUxeXzlAwCoDcbuBH9hbi58,222
+eva/language/models/networks/api/anthropic.py,sha256=EdIKcFS7RCX_vpiOYMh-tS6rM74orOcNRmhw9pETen8,1194
+eva/language/models/networks/registry.py,sha256=yFA1AhIbgoDLRvn9MUU2FPH25H7BODpN49CN_ya7wrY,106
+eva/language/models/typings.py,sha256=SclZPkxg9T3ADl5SVWXzV4ymJyJ5Wf0VxDrWRRlBLIU,906
+eva/language/models/wrappers/__init__.py,sha256=N5DrkKZs_zE7WHY76bnhViO1hpoO88Sm1BOj12tgwJU,626
+eva/language/models/wrappers/base.py,sha256=0MLM-Ds20hftbO3JUUWd7wuXXRn2axz1ITUplw0u_LI,1682
+eva/language/models/wrappers/from_registry.py,sha256=hdGzLfkgjZLrEb9GyqjF1mNlFMTErQ6sY900_tBJHH8,1793
+eva/language/models/wrappers/huggingface.py,sha256=Qz_GAcIwJWt0Ao4gRSII7FEDUBMqYU5u5lVBejI3j_Y,4025
+eva/language/models/wrappers/litellm.py,sha256=sGyaQHaFEjgEHxTcnlMHa2b7L5eln27Tv8g821eXd_U,3442
+eva/language/models/wrappers/vllm.py,sha256=puJVOFeyFrqBplvkjvH5dGALlZD4gduaX3lY5fScbQU,6514
+eva/language/utils/__init__.py,sha256=Bvw7Z5Z-mBQp7rTycycrgrxxrbItD5IqNU4uHRAnD9s,238
+eva/language/utils/str_to_int_tensor.py,sha256=_D2FH_5S1Sc4ywsJ6EWrTwuaw-NJVHygwAaRSRs-NPs,3471
+eva/language/utils/text/__init__.py,sha256=fxMel4GvRx-SKrxGcwXDNG77yj_deEOWc6mkePS2aWo,143
+eva/language/utils/text/messages.py,sha256=9i4AUNjAo8XOF0_gB_MaoQy4SHI2zohz65qtMoTYtWE,3989
+eva/multimodal/__init__.py,sha256=aRq8KCWg4h7XHS68EWSHDepu9VeACxVOU3B4DiQhphE,132
+eva/multimodal/callbacks/__init__.py,sha256=tIwSH9Z1YJfebSU3-rSCGJgaX3Io6rptLX13gmI6Hck,135
+eva/multimodal/callbacks/writers/__init__.py,sha256=6UYVvH3Hbya0haKQh9Dci_lRji9kz1RO4Z_vnTnY24U,154
+eva/multimodal/callbacks/writers/prediction.py,sha256=OSfInVpz2qUDTkzM3MzwWZiJoa3N4y8amMLV5TEK-k0,1316
+eva/multimodal/data/__init__.py,sha256=wL6XCRfIuDp-7YWaHsWZ90CdII-reSW1jQ2dn8_OC0Q,113
+eva/multimodal/data/dataloaders/__init__.py,sha256=OZ4KnLxwfs1ccA6FItyoFBtgUL_vd5JnOycCOhk8H5M,143
+eva/multimodal/data/dataloaders/collate_fn/__init__.py,sha256=1CHJ7cL9ELlo92LakB5ib0YD-kbTjzBrMBZ5j5kYu74,160
+eva/multimodal/data/dataloaders/collate_fn/text_image.py,sha256=Hwf4GOmYtYcJYMTiZSVrSppo_SQqcAmvvahe_4xoVb0,883
+eva/multimodal/data/datasets/__init__.py,sha256=XisAN7MOBncOucelkVOch_rX7KSHPIL7W6lmKUQVQuQ,236
+eva/multimodal/data/datasets/base.py,sha256=mRkxiVpZJ3Rn_YRHvqGGJiAr-I1qY1A2H53vhn59d2g,315
+eva/multimodal/data/datasets/multiple_choice/__init__.py,sha256=tmMLCUQWBAGvDPaUVL4qNoXTabkagoZTgbiAJ9Sci-E,148
+eva/multimodal/data/datasets/multiple_choice/patch_camelyon.py,sha256=r2ZgjStennqHDK4H2Q980sDObvZDjTA6XWSLeBmbWaQ,3100
+eva/multimodal/data/datasets/schemas.py,sha256=44ll-3B8V8ctkmUJ1EuoUSow38EmrRTx9K1nfLw3AgE,366
+eva/multimodal/data/datasets/text_image.py,sha256=CxT8ej6lmcWDdCeimgBcujfgMrdCMEm5LhYnzW0RbL0,2608
+eva/multimodal/data/datasets/typings.py,sha256=dkxszTFljKf8CUlIxblsIlNR0Ckjrl2_NTBq7cAY-fk,641
+eva/multimodal/models/__init__.py,sha256=AILD2D1zLnPW6nelWTjwIBA4jS0kx2iSw3HEQ1wMJTo,130
+eva/multimodal/models/modules/__init__.py,sha256=E6bhjdMGPBEUNI_Lq8foLQjRTbv26B7vgxOB8q3jeSQ,147
+eva/multimodal/models/modules/vision_language.py,sha256=DaBv3_uzNllUeZcixWiDg1Mf5c0vZnALhwS1oC8uSfw,1844
+eva/multimodal/models/networks/__init__.py,sha256=RjW6MZ4lCAAIbstTqIq_rkaZd-fm1_4XwLMetIry9gw,463
+eva/multimodal/models/networks/alibaba.py,sha256=dsGG7QVFxLeMKSWMPOzJRPtWTyuIwFWQZYbf-5ibpGg,1241
+eva/multimodal/models/networks/api/__init__.py,sha256=eE0nB7XjhuSH1itEyyMtJKxHwHg1IfsHU8S69AvfDo8,224
+eva/multimodal/models/networks/api/anthropic.py,sha256=zlPiML2JL0q-l528x29dez-FofO5ktmOjPLFRh1PwTk,1198
+eva/multimodal/models/networks/others.py,sha256=TMBR1cjzLa4azs2mHamdi1tzeRzE8wPtt8lLdYgp9IM,1535
+eva/multimodal/models/networks/registry.py,sha256=beZsRZUrKdUhMPq7UYFOEF7z50FgxGT1OCwE64xBf9A,108
+eva/multimodal/models/typings.py,sha256=bwQbUFIsXhAxk7YTXqRdBYw0RqPA0a6RGNLEt4mBrxk,693
+eva/multimodal/models/wrappers/__init__.py,sha256=RvLw74VGP3mh1CMYPj4Ij09TNkBml4ohvdirXeVBm3I,421
+eva/multimodal/models/wrappers/base.py,sha256=FXRjAekVIovc3kbd8Q6lkwpWTSML8a8XRFAgOCl2Gsk,1719
+eva/multimodal/models/wrappers/from_registry.py,sha256=JBCck-ecIOC9_jSaEVMbU8R2MOgjHQRxzb0SzO4whIQ,1807
+eva/multimodal/models/wrappers/huggingface.py,sha256=qyyhl1C90JKoOD9-p4TeK-hWWLo8T1iJgx8-yBa13O4,6714
+eva/multimodal/models/wrappers/litellm.py,sha256=EyN180LbyfkkHXNMVdEcgKjv-7COqTvall3YOcvmZ_I,2063
+eva/multimodal/utils/__init__.py,sha256=pRJUXvuJtcKOJc1yqKHdQv9dDAn8fOoP0d6Q_9cR7ek,32
+eva/multimodal/utils/image/__init__.py,sha256=7DCx0FgFC0CwcR40872P6w1bQWgP8HXuwyy47ZQiv-Q,126
+eva/multimodal/utils/image/encode.py,sha256=q_tq6MJtsy6yWu6Fei1j0mMmyKGBIf1CmvJcK4ev3eU,898
+eva/multimodal/utils/text/__init__.py,sha256=Uk1vnstlU7hLa2Pn4cacVhdhQwe2vsBoPsB0_ZWjOdc,37
+eva/multimodal/utils/text/messages.py,sha256=IIYKXE-6FwoKM3zNt3edKFqiJdEpNLnZXR4yhSZdjkM,2541
 eva/vision/__init__.py,sha256=oUZXFYjwtkWzi8An0uS5Xc84pLKintlXe2iti8zW6BQ,480
 eva/vision/callbacks/__init__.py,sha256=su1V73L0dDVYWSyvV_lnWbszDi2KikRraF7OsgeaKl4,139
 eva/vision/callbacks/loggers/__init__.py,sha256=td1JRJbE08nsGIZdO64_yLC3FUuMDp0kma0HjpUdXT4,161
@@ -159,7 +213,7 @@ eva/vision/data/datasets/classification/crc.py,sha256=sv18Lw4iUqGkYXEQN-kyZV_Foc
 eva/vision/data/datasets/classification/gleason_arvaniti.py,sha256=z1OQlxZYx-n3S2wcfu-GuchySRw0E70PURJDsvejFjE,5871
 eva/vision/data/datasets/classification/mhist.py,sha256=I-guWIucQZBHdSx-TWP24NXHf9IA9lU4hyfPZbJop0g,3033
 eva/vision/data/datasets/classification/panda.py,sha256=HVfCvByyajdo5o_waqTpzZWCbQXQqPjvvyS5I0NAvns,7277
-eva/vision/data/datasets/classification/patch_camelyon.py,sha256=1yXkfP680qxkQUFAPKRFbZv0cHAFx23s2vvT9th2nKM,7149
+eva/vision/data/datasets/classification/patch_camelyon.py,sha256=mzgbVGQtaoer9yBXV8pXZx3DjbCy1w6aHNDJfC6CAOI,7187
 eva/vision/data/datasets/classification/unitopatho.py,sha256=IO3msEsuOnmdcYZxF-eBpo0K97y54rWFmCb_KxuF4bk,5129
 eva/vision/data/datasets/classification/wsi.py,sha256=YMGxU8ECjudizt_uXUevuPS8k66HxtEQ7M2IZJmL6kE,4079
 eva/vision/data/datasets/segmentation/__init__.py,sha256=f0q9tzk4ahaZfrw_SgIE_puk_D7qmkSCKX1FP9aJITU,668
@@ -176,7 +230,7 @@ eva/vision/data/datasets/segmentation/msd_task7_pancreas.py,sha256=dTsPD73PAP15V
 eva/vision/data/datasets/structs.py,sha256=RaTDW-B36PumcR5gymhCiX-r8GiKqIFcjqoEEjjFyUE,389
 eva/vision/data/datasets/vision.py,sha256=-_WRiyICMgqABR6Ay_RKBMfsPGwgx9MQfCA7WChHo24,3219
 eva/vision/data/datasets/wsi.py,sha256=dEAT_Si_Qb3qdSovUPeoiWeoPb7m-NGYqq44e3UXHk8,8384
-eva/vision/data/transforms/__init__.py,sha256=39KkoDaIILjgbCEi_WD48a0p6TKV7Woe6744a-CLeOU,863
+eva/vision/data/transforms/__init__.py,sha256=7CpXeo-oSjft1VpzjK-dmozC0Ds-HeBzBU_KYLTnfIo,885
 eva/vision/data/transforms/base/__init__.py,sha256=d3gAnEHx8WZHuUpKLr0YAC1JR60qhRY02yeoADrx1Yw,143
 eva/vision/data/transforms/base/monai.py,sha256=g2U4eNxKiPm6TXjHjKZ4L8-1zxa1tEf-c73kKI_7WwQ,1081
 eva/vision/data/transforms/common/__init__.py,sha256=LWA5u0VsWzEEugdKpjpePr7fgV_XIGdASGwoaOU29Ac,211
@@ -192,8 +246,11 @@ eva/vision/data/transforms/intensity/__init__.py,sha256=mNp6pi0pnHcA24kQuiGHzMb4
 eva/vision/data/transforms/intensity/rand_scale_intensity.py,sha256=x3gYRRAFbZXFWgqvoUjT0sDFWT6zbhja3PJLXbOBtTg,2184
 eva/vision/data/transforms/intensity/rand_shift_intensity.py,sha256=5ngSHfHgJOR9KQykNKJhQMWhQYj781Mj_JKJ_ADchbM,2117
 eva/vision/data/transforms/intensity/scale_intensity_ranged.py,sha256=VLvYZYG6jQCuR5poJsAlhIFjw6VjPEpcDPKBlJTjYBM,1873
-eva/vision/data/transforms/spatial/__init__.py,sha256=k7C_p4fMZd7A00ikldAMsprYDedKrlMjKQB6BLA5InA,284
+eva/vision/data/transforms/spatial/__init__.py,sha256=S2JwmBm98oPWLN497aYM-h1kyPU15Yfg9xPVTvzq6lg,355
 eva/vision/data/transforms/spatial/flip.py,sha256=jfRc-wPBvG58OtCNU3GrOkb57kcRddRqpwcAdCB0_No,2553
+eva/vision/data/transforms/spatial/functional/__init__.py,sha256=OzJ_Hv3CJAvvTQsk6mo_IQS0MqPW2WQ6fwtWE3UQ_Xk,166
+eva/vision/data/transforms/spatial/functional/resize.py,sha256=NBswLluIRs1cJP-dx5Kd3Z01gF57NjPXbtPGKNa0td0,949
+eva/vision/data/transforms/spatial/resize.py,sha256=O7igA-x7jCutw5BE_5MXOEfCeLsZPUz9K2VEIekMCi4,2305
 eva/vision/data/transforms/spatial/rotate.py,sha256=FpMTAPWtgrG10yQ3R1_Ii6obPcn3boNWOuLhsblxUbQ,1793
 eva/vision/data/transforms/spatial/spacing.py,sha256=T1UhqK-OhhbLQxzejMyI8BQzYRF44PNc02Qap4nk1hY,2695
 eva/vision/data/transforms/utility/__init__.py,sha256=TjncS2aOgRJwjjRuIvmr4eRz2nKVg6b76tThp4UlzII,163
@@ -265,8 +322,8 @@ eva/vision/models/networks/decoders/segmentation/semantic/swin_unetr.py,sha256=e
 eva/vision/models/networks/decoders/segmentation/semantic/with_image.py,sha256=I5PyGKKo8DcXYcw4xlCFzuavRJNRrzGT-szpDidMPXI,3516
 eva/vision/models/networks/decoders/segmentation/typings.py,sha256=rY4CXp0MNF16SHnx9TgGjXI_r8bVGSqAWdR835hXndg,537
 eva/vision/models/wrappers/__init__.py,sha256=ogmr-eeVuGaOCcsuxSp6PGyauP2QqWTb8dGTtbC7lRU,210
-eva/vision/models/wrappers/from_registry.py,sha256=2vpKTkoa_CGjVLbhnUclstFzDfkXCqMxjdtxTk212HQ,1729
-eva/vision/models/wrappers/from_timm.py,sha256=Lkdibly4H_XNjl863YFBj10ZbSTN1onT2Bc6yV072m0,2334
+eva/vision/models/wrappers/from_registry.py,sha256=aZTbPqlfbxvx5vmh2aUym9njTlEaFwmTL5fU2zPW0To,1761
+eva/vision/models/wrappers/from_timm.py,sha256=vf5ZoEdh-QJW-wGD6PgpqkkgQzXgQ4lT1YIaI_kwnfU,2367
 eva/vision/utils/__init__.py,sha256=vaUovprE743SmyFH8l6uk4pYSWpI4zxn7lN0EwePTJI,96
 eva/vision/utils/colormap.py,sha256=sP1F0JCX3abZfFgdxEjLJO-LhNYKjXZvXxs03ZgrEvI,2876
 eva/vision/utils/convert.py,sha256=fqGmKrg5-JJLrTkTXB4YDcWTudXPrO1gGjsckVRUesU,1881
@@ -276,8 +333,8 @@ eva/vision/utils/io/image.py,sha256=IdOkr5MYqhYHz8U9drZ7wULTM3YHwCWSjZlu_Qdl4GQ,
 eva/vision/utils/io/mat.py,sha256=qpGifyjmpE0Xhv567Si7-zxKrgkgE0sywP70cHiLFGU,808
 eva/vision/utils/io/nifti.py,sha256=TFMgNhLqIK3sl3RjIRXEABM7FmSQjqVOwk1vXkuvX2w,4983
 eva/vision/utils/io/text.py,sha256=qYgfo_ZaDZWfG02NkVVYzo5QFySqdCCz5uLA9d-zXtI,701
-kaiko_eva-0.3.3.dist-info/METADATA,sha256=3-qTJLt0hRJswZbPVixj9e9Bt0gFXDjOp6YbNf6Ohd4,25704
-kaiko_eva-0.3.3.dist-info/WHEEL,sha256=9P2ygRxDrTJz3gsagc0Z96ukrxjr-LFBGOgv3AuKlCA,90
-kaiko_eva-0.3.3.dist-info/entry_points.txt,sha256=6CSLu9bmQYJSXEg8gbOzRhxH0AGs75BB-vPm3VvfcNE,88
-kaiko_eva-0.3.3.dist-info/licenses/LICENSE,sha256=e6AEzr7j_R-PYr2qLO-JwLn8y70jbVD3U2mxbRmwcI4,11338
-kaiko_eva-0.3.3.dist-info/RECORD,,
+kaiko_eva-0.4.0.dist-info/METADATA,sha256=tw3KrpA3CqmTlDJ28SxkI_4t32IHcDE0YluzC3-xbq4,26017
+kaiko_eva-0.4.0.dist-info/WHEEL,sha256=9P2ygRxDrTJz3gsagc0Z96ukrxjr-LFBGOgv3AuKlCA,90
+kaiko_eva-0.4.0.dist-info/entry_points.txt,sha256=6CSLu9bmQYJSXEg8gbOzRhxH0AGs75BB-vPm3VvfcNE,88
+kaiko_eva-0.4.0.dist-info/licenses/LICENSE,sha256=e6AEzr7j_R-PYr2qLO-JwLn8y70jbVD3U2mxbRmwcI4,11338
+kaiko_eva-0.4.0.dist-info/RECORD,,

eva/core/data/dataloaders/collate_fn/__init__.py

@@ -1,5 +0,0 @@
-"""Collate functions API."""
-
-from eva.core.data.dataloaders.collate_fn.collate import text_collate_fn
-
-__all__ = ["text_collate_fn"]

eva/core/data/dataloaders/collate_fn/collate.py

@@ -1,24 +0,0 @@
-"""Collate functions for text data."""
-
-from typing import Dict, List, Tuple
-
-import torch
-
-
-def text_collate_fn(
-    batch: List[Tuple[str, torch.Tensor, Dict]],
-) -> Tuple[List[str], torch.Tensor, List[Dict]]:
-    """Collate function for text data that keeps texts as separate strings.
-
-    Args:
-        batch: List of tuples containing (text, target, metadata) from the dataset
-
-    Returns:
-        Tuple containing:
-            - List of text strings
-            - Batched tensor of targets
-            - List of metadata dictionaries
-    """
-    texts, targets, metadata = zip(*batch, strict=False)
-    targets = torch.stack(targets)
-    return list(texts), targets, list(metadata)

eva/language/models/modules/text.py

@@ -1,85 +0,0 @@
-"""LLM Text Module for Inference."""
-
-from typing import Any, List
-
-from lightning.pytorch.utilities.types import STEP_OUTPUT
-from loguru import logger
-from torch import nn
-from typing_extensions import override
-
-from eva.core.metrics import structs as metrics_lib
-from eva.core.models.modules import module
-from eva.core.models.modules.utils import batch_postprocess
-from eva.language.models.modules.typings import TEXT_BATCH
-
-
-class TextModule(module.ModelModule):
-    """Text-based LLM module for inference.
-
-    Uses LLM wrappers for text generation and supports evaluation using
-    configurable metrics and post-processing transforms.
-    """
-
-    def __init__(
-        self,
-        model: nn.Module,
-        prompt: str,
-        metrics: metrics_lib.MetricsSchema | None = None,
-        postprocess: batch_postprocess.BatchPostProcess | None = None,
-    ) -> None:
-        """Initializes the text inference module.
-
-        Args:
-            model: An LLM wrapper (PyTorch-compatible) for text generation.
-            prompt: The prompt to use for generating text.
-            metrics: Metrics schema for evaluation.
-            postprocess: A helper function to post-process model outputs before evaluation.
-        """
-        super().__init__(metrics=metrics, postprocess=postprocess)
-
-        self.model = model
-        self.prompt = prompt
-
-    @override
-    def forward(self, prompts: List[str], *args: Any, **kwargs: Any) -> List[str]:
-        """Generates text responses for a batch of prompts.
-
-        Args:
-            prompts: List of input texts to generate responses.
-            args: Additional arguments.
-            kwargs: Additional keyword arguments.
-
-        Returns:
-            List of generated responses.
-        """
-        return self.model(prompts)
-
-    @override
-    def validation_step(self, batch: TEXT_BATCH, *args: Any, **kwargs: Any) -> STEP_OUTPUT:
-        """Validation step that runs batch inference and evaluates metrics.
-
-        Args:
-            batch: An input batch.
-            args: Additional arguments.
-            kwargs: Additional keyword arguments.
-
-        Returns:
-            Dictionary with predictions, ground truth, and evaluation metrics.
-        """
-        return self._batch_step(batch)
-
-    def _batch_step(self, batch: TEXT_BATCH) -> STEP_OUTPUT:
-        """Runs inference on a batch and evaluates model predictions.
-
-        Args:
-            batch: Input batch containing data, targets, and metadata.
-
-        Returns:
-            Dictionary with predictions, ground truth, and evaluation metrics.
-        """
-        data, targets, metadata = batch
-        messages = [str(d) + "\n" + self.prompt for d in data]
-        predictions = self(messages)
-        logger.debug(f"Predictions: {predictions}")
-        logger.debug(f"Targets: {targets}")
-        return {"predictions": predictions, "targets": targets, "metadata": metadata}