fusion-bench 0.2.9__py3-none-any.whl

fusion_bench/models/modeling_smile_mistral/register.py

@@ -0,0 +1,8 @@
+from transformers import AutoConfig, AutoModel, AutoModelForCausalLM
+
+from .configuration_smile_mistral import SmileMistralConfig
+from .modeling_smile_mistral import SmileMistralForCausalLM, SmileMistralModel
+
+AutoConfig.register("smile_mistral", SmileMistralConfig)
+AutoModel.register(SmileMistralConfig, SmileMistralModel)
+AutoModelForCausalLM.register(SmileMistralConfig, SmileMistralForCausalLM)

fusion_bench/models/nyuv2/aspp.py

@@ -0,0 +1,82 @@
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+
+class DeepLabHead(nn.Sequential):
+    def __init__(self, in_channels: int, num_classes: int):
+        super(DeepLabHead, self).__init__(
+            ASPP(in_channels, [12, 24, 36]),
+            nn.Conv2d(256, 256, 3, padding=1, bias=False),
+            nn.BatchNorm2d(256),
+            nn.ReLU(),
+            nn.Conv2d(256, num_classes, 1),
+        )
+
+
+class ASPPConv(nn.Sequential):
+    def __init__(self, in_channels, out_channels, dilation):
+        modules = [
+            nn.Conv2d(
+                in_channels,
+                out_channels,
+                3,
+                padding=dilation,
+                dilation=dilation,
+                bias=False,
+            ),
+            nn.BatchNorm2d(out_channels),
+            nn.ReLU(),
+        ]
+        super(ASPPConv, self).__init__(*modules)
+
+
+class ASPPPooling(nn.Sequential):
+    def __init__(self, in_channels, out_channels):
+        super(ASPPPooling, self).__init__(
+            nn.AdaptiveAvgPool2d(1),
+            nn.Conv2d(in_channels, out_channels, 1, bias=False),
+            nn.BatchNorm2d(out_channels),
+            nn.ReLU(),
+        )
+
+    def forward(self, x):
+        size = x.shape[-2:]
+        x = super(ASPPPooling, self).forward(x)
+        return F.interpolate(x, size=size, mode="bilinear", align_corners=False)
+
+
+class ASPP(nn.Module):
+    def __init__(self, in_channels: int, atrous_rates):
+        super(ASPP, self).__init__()
+        out_channels = 256
+        modules = []
+        modules.append(
+            nn.Sequential(
+                nn.Conv2d(in_channels, out_channels, 1, bias=False),
+                nn.BatchNorm2d(out_channels),
+                nn.ReLU(),
+            )
+        )
+
+        rate1, rate2, rate3 = tuple(atrous_rates)
+        modules.append(ASPPConv(in_channels, out_channels, rate1))
+        modules.append(ASPPConv(in_channels, out_channels, rate2))
+        modules.append(ASPPConv(in_channels, out_channels, rate3))
+        modules.append(ASPPPooling(in_channels, out_channels))
+
+        self.convs = nn.ModuleList(modules)
+
+        self.project = nn.Sequential(
+            nn.Conv2d(5 * out_channels, out_channels, 1, bias=False),
+            nn.BatchNorm2d(out_channels),
+            nn.ReLU(),
+            nn.Dropout(0.5),
+        )
+
+    def forward(self, x):
+        res = []
+        for conv in self.convs:
+            res.append(conv(x))
+        res = torch.cat(res, dim=1)
+        return self.project(res)

fusion_bench/models/nyuv2/lightning_module.py

@@ -0,0 +1,176 @@
+import itertools
+from typing import Dict, List, Literal, Optional, cast
+
+import lightning as L
+import torch.nn.functional as F
+from torch import Tensor, nn
+from torchmetrics import Metric
+
+from fusion_bench.metrics.nyuv2 import metric_classes
+from fusion_bench.metrics.nyuv2.loss import loss_fn
+
+
+class NYUv2Model(nn.Module):
+    image_size = (288, 384)
+
+    def __init__(
+        self,
+        encoder: nn.Module,
+        decoders: nn.ModuleDict,
+    ):
+        R"""
+        Args:
+            encoder: The encoder module.
+            decoders: A dictionary of the decoder modules.
+        """
+        super().__init__()
+
+        self.encoder = encoder
+        self.decoders = decoders
+
+    def encode(self, images: Tensor) -> Tensor:
+        return self.encoder(images)
+
+    def decode(self, features: Tensor, key: str) -> Tensor:
+        return self.decoders[key](features)
+
+    def model_parameters(self):
+        "parameters of the encoder and the decoders"
+        return itertools.chain(
+            self.encoder.parameters(),
+            self.decoders.parameters(),
+        )
+
+    def _compute_loss(
+        self,
+        outputs: Dict[str, Tensor],
+        targets: Dict[str, Tensor],
+        task: Literal["segmentation", "depth", "normal"] | List[str],
+    ) -> Tensor | List[Tensor]:
+        if isinstance(task, str):
+            return loss_fn[task](outputs[task], targets[task])
+        else:
+            return [loss_fn[t](outputs[t], targets[t]) for t in task]
+
+    def forward(
+        self, images: Tensor, tasks: Optional[List[str]] = None
+    ) -> Dict[str, Tensor]:
+        features = self.encode(images)
+
+        if tasks is None:
+            tasks = self.decoders.keys()
+        outputs = {}
+        for task in tasks:
+            outputs[task] = F.interpolate(
+                self.decode(features, task),
+                self.image_size,
+                mode="bilinear",
+                align_corners=True,
+            )
+
+        return outputs
+
+
+class NYUv2MTLModule(NYUv2Model, L.LightningModule):
+    R"""
+    multi-task learning module for NYUv2 dataset, with weighted loss
+    """
+
+    def __init__(
+        self,
+        encoder: nn.Module,
+        decoders: nn.ModuleDict,
+        tasks: List[str],
+        task_weights: List[float],
+    ):
+        R"""
+        By using this module, you optimize the weighted sum of the losses of the tasks.
+        The joint loss is defined as:
+
+        .. math::
+            \mathcal{L} = \sum_{i=1}^{N} w_i \mathcal{L}_i
+
+        where :math:`N` is the number of tasks, :math:`w_i` is the weight for the task :math:`i`,
+
+        Args:
+            encoder (nn.Module): The encoder module.
+            decoders (nn.ModuleDict): A dictionary of the decoder modules.
+            tasks (List[str]): A list of tasks.
+            task_weights (List[float]): A list of weights for each task.
+
+        Raises:
+            AssertionError: If tasks and task_weights do not have the same length.
+        """
+        assert len(tasks) == len(
+            task_weights
+        ), "tasks and task_weights must have the same length"
+
+        super().__init__(encoder, decoders)
+        self.tasks = tasks
+        self.task_weights = task_weights
+
+        self.metrics = nn.ModuleDict({t: metric_classes[t]() for t in self.tasks})
+
+    # training
+
+    def on_train_epoch_start(self):
+        for t in self.tasks:
+            self.metrics[t].reset()
+
+    def _single_step(self, images, targets):
+        outputs = self(images, self.tasks)
+        losses = self._compute_loss(outputs, targets, self.tasks)
+        weighted_loss = sum([w * l for w, l in zip(self.task_weights, losses)])
+
+        for t in self.metrics:
+            self.metrics[t].update(outputs[t], targets[t])
+
+        return {
+            "outputs": outputs,
+            "losses": losses,
+            "weighted_loss": weighted_loss,
+        }
+
+    def training_step(self, batch, batch_idx: int):
+        images, targets = batch
+        results = self._single_step(images, targets)
+
+        outputs = results["outputs"]
+        losses = results["losses"]
+        weighted_loss = results["weighted_loss"]
+
+        for i, t in enumerate(self.tasks):
+            self.log(f"train/{t}_loss", losses[i])
+        self.log("train/loss", weighted_loss, prog_bar=True)
+
+        return weighted_loss
+
+    def on_train_epoch_end(self) -> None:
+        for t in self.tasks:
+            metrics = cast(Metric, self.metrics[t]).compute()
+            for metric_name, metric_value in zip(self.metrics[t].metric_names, metrics):
+                self.log(f"train/{t}_{metric_name}", metric_value)
+
+    # validation
+
+    def on_validation_epoch_start(self) -> None:
+        for t in self.tasks:
+            self.metrics[t].reset()
+
+    def validation_step(self, batch, batch_idx: int):
+        images, targets = batch
+        results = self._single_step(images, targets)
+
+        outputs = results["outputs"]
+        losses = results["losses"]
+        weighted_loss = results["weighted_loss"]
+
+        for i, t in enumerate(self.tasks):
+            self.log(f"val/{t}_loss", losses[i])
+        self.log("val/loss", weighted_loss, prog_bar=True)
+
+    def on_validation_epoch_end(self) -> None:
+        for t in self.tasks:
+            metrics = cast(Metric, self.metrics[t]).compute()
+            for metric_name, metric_value in zip(self.metrics[t].metric_names, metrics):
+                self.log(f"val/{t}_{metric_name}", metric_value)

fusion_bench/models/nyuv2/resnet.py

@@ -0,0 +1,405 @@
+import torch
+import torch.nn as nn
+from torch.hub import load_state_dict_from_url
+
+# from torchvision.models.utils import load_state_dict_from_url
+
+
+# __all__ = ['ResNet', 'resnet18', 'resnet34', 'resnet50', 'resnet101',
+#            'resnet152', 'resnext50_32x4d', 'resnext101_32x8d',
+#            'wide_resnet50_2', 'wide_resnet101_2']
+
+
+model_urls = {
+    "resnet18": "https://download.pytorch.org/models/resnet18-5c106cde.pth",
+    "resnet34": "https://download.pytorch.org/models/resnet34-333f7ec4.pth",
+    "resnet50": "https://download.pytorch.org/models/resnet50-19c8e357.pth",
+    "resnet101": "https://download.pytorch.org/models/resnet101-5d3b4d8f.pth",
+    "resnet152": "https://download.pytorch.org/models/resnet152-b121ed2d.pth",
+    "resnext50_32x4d": "https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth",
+    "resnext101_32x8d": "https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth",
+    "wide_resnet50_2": "https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth",
+    "wide_resnet101_2": "https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth",
+}
+
+
+def conv3x3(in_planes, out_planes, stride=1, groups=1, dilation=1):
+    """3x3 convolution with padding"""
+    return nn.Conv2d(
+        in_planes,
+        out_planes,
+        kernel_size=3,
+        stride=stride,
+        padding=dilation,
+        groups=groups,
+        bias=False,
+        dilation=dilation,
+    )
+
+
+def conv1x1(in_planes, out_planes, stride=1):
+    """1x1 convolution"""
+    return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False)
+
+
+class BasicBlock(nn.Module):
+    expansion = 1
+    __constants__ = ["downsample"]
+
+    def __init__(
+        self,
+        inplanes,
+        planes,
+        stride=1,
+        downsample=None,
+        groups=1,
+        base_width=64,
+        dilation=1,
+        norm_layer=None,
+    ):
+        super(BasicBlock, self).__init__()
+        if norm_layer is None:
+            norm_layer = nn.BatchNorm2d
+        if groups != 1 or base_width != 64:
+            raise ValueError("BasicBlock only supports groups=1 and base_width=64")
+        if dilation > 1:
+            raise NotImplementedError("Dilation > 1 not supported in BasicBlock")
+        # Both self.conv1 and self.downsample layers downsample the input when stride != 1
+        self.conv1 = conv3x3(inplanes, planes, stride)
+        self.bn1 = norm_layer(planes)
+        self.relu = nn.ReLU(inplace=True)
+        self.conv2 = conv3x3(planes, planes)
+        self.bn2 = norm_layer(planes)
+        self.downsample = downsample
+        self.stride = stride
+
+    def forward(self, x):
+        identity = x
+
+        out = self.conv1(x)
+        out = self.bn1(out)
+        out = self.relu(out)
+
+        out = self.conv2(out)
+        out = self.bn2(out)
+
+        if self.downsample is not None:
+            identity = self.downsample(x)
+
+        out += identity
+        out = self.relu(out)
+
+        return out
+
+
+class Bottleneck(nn.Module):
+    expansion = 4
+    __constants__ = ["downsample"]
+
+    def __init__(
+        self,
+        inplanes,
+        planes,
+        stride=1,
+        downsample=None,
+        groups=1,
+        base_width=64,
+        dilation=1,
+        norm_layer=None,
+    ):
+        super(Bottleneck, self).__init__()
+        if norm_layer is None:
+            norm_layer = nn.BatchNorm2d
+        width = int(planes * (base_width / 64.0)) * groups
+        # Both self.conv2 and self.downsample layers downsample the input when stride != 1
+        self.conv1 = conv1x1(inplanes, width)
+        self.bn1 = norm_layer(width)
+        self.conv2 = conv3x3(width, width, stride, groups, dilation)
+        self.bn2 = norm_layer(width)
+        self.conv3 = conv1x1(width, planes * self.expansion)
+        self.bn3 = norm_layer(planes * self.expansion)
+        self.relu = nn.ReLU(inplace=True)
+        self.downsample = downsample
+        self.stride = stride
+
+    def forward(self, x):
+        identity = x
+
+        out = self.conv1(x)
+        out = self.bn1(out)
+        out = self.relu(out)
+
+        out = self.conv2(out)
+        out = self.bn2(out)
+        out = self.relu(out)
+
+        out = self.conv3(out)
+        out = self.bn3(out)
+
+        if self.downsample is not None:
+            identity = self.downsample(x)
+
+        out += identity
+        out = self.relu(out)
+
+        return out
+
+
+class ResNet(nn.Module):
+    def __init__(
+        self,
+        block,
+        layers,
+        num_classes=1000,
+        zero_init_residual=False,
+        groups=1,
+        width_per_group=64,
+        replace_stride_with_dilation=None,
+        norm_layer=None,
+    ):
+        super(ResNet, self).__init__()
+        if norm_layer is None:
+            norm_layer = nn.BatchNorm2d
+        self._norm_layer = norm_layer
+
+        self.inplanes = 64
+        self.dilation = 1
+        if replace_stride_with_dilation is None:
+            # each element in the tuple indicates if we should replace
+            # the 2x2 stride with a dilated convolution instead
+            replace_stride_with_dilation = [False, False, False]
+        if len(replace_stride_with_dilation) != 3:
+            raise ValueError(
+                "replace_stride_with_dilation should be None "
+                "or a 3-element tuple, got {}".format(replace_stride_with_dilation)
+            )
+        self.groups = groups
+        self.base_width = width_per_group
+        self.conv1 = nn.Conv2d(
+            3, self.inplanes, kernel_size=7, stride=2, padding=3, bias=False
+        )
+        self.bn1 = norm_layer(self.inplanes)
+        self.relu = nn.ReLU(inplace=True)
+        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+        self.layer1 = self._make_layer(block, 64, layers[0])
+        self.layer2 = self._make_layer(
+            block, 128, layers[1], stride=2, dilate=replace_stride_with_dilation[0]
+        )
+        self.layer3 = self._make_layer(
+            block, 256, layers[2], stride=2, dilate=replace_stride_with_dilation[1]
+        )
+        self.layer4 = self._make_layer(
+            block, 512, layers[3], stride=2, dilate=replace_stride_with_dilation[2]
+        )
+        #         self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
+        #         self.fc = nn.Linear(512 * block.expansion, num_classes)
+        self.feature_dim = 512 * block.expansion
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                nn.init.kaiming_normal_(m.weight, mode="fan_out", nonlinearity="relu")
+            elif isinstance(m, (nn.BatchNorm2d, nn.GroupNorm)):
+                nn.init.constant_(m.weight, 1)
+                nn.init.constant_(m.bias, 0)
+
+        # Zero-initialize the last BN in each residual branch,
+        # so that the residual branch starts with zeros, and each residual block behaves like an identity.
+        # This improves the model by 0.2~0.3% according to https://arxiv.org/abs/1706.02677
+        if zero_init_residual:
+            for m in self.modules():
+                if isinstance(m, Bottleneck):
+                    nn.init.constant_(m.bn3.weight, 0)
+                elif isinstance(m, BasicBlock):
+                    nn.init.constant_(m.bn2.weight, 0)
+
+    def _make_layer(self, block, planes, blocks, stride=1, dilate=False):
+        norm_layer = self._norm_layer
+        downsample = None
+        previous_dilation = self.dilation
+        if dilate:
+            self.dilation *= stride
+            stride = 1
+        if stride != 1 or self.inplanes != planes * block.expansion:
+            downsample = nn.Sequential(
+                conv1x1(self.inplanes, planes * block.expansion, stride),
+                norm_layer(planes * block.expansion),
+            )
+
+        layers = []
+        layers.append(
+            block(
+                self.inplanes,
+                planes,
+                stride,
+                downsample,
+                self.groups,
+                self.base_width,
+                previous_dilation,
+                norm_layer,
+            )
+        )
+        self.inplanes = planes * block.expansion
+        for _ in range(1, blocks):
+            layers.append(
+                block(
+                    self.inplanes,
+                    planes,
+                    groups=self.groups,
+                    base_width=self.base_width,
+                    dilation=self.dilation,
+                    norm_layer=norm_layer,
+                )
+            )
+
+        return nn.Sequential(*layers)
+
+    def forward(self, x):
+        x = self.conv1(x)
+        x = self.bn1(x)
+        x = self.relu(x)
+        x = self.maxpool(x)
+
+        x = self.layer1(x)
+        x = self.layer2(x)
+        x = self.layer3(x)
+        x = self.layer4(x)
+
+        # We remove the original classifier, to attach task-specific decoders.
+        # x = self.avgpool(x)
+        # x = torch.flatten(x, 1)
+        # x = self.fc(x)
+        return x
+
+
+def _resnet(arch, block, layers, pretrained, progress, **kwargs):
+    model = ResNet(block, layers, **kwargs)
+    if pretrained:
+        state_dict = load_state_dict_from_url(model_urls[arch], progress=progress)
+        model.load_state_dict(state_dict, strict=False)
+    return model
+
+
+def resnet18(pretrained=False, progress=True, **kwargs):
+    r"""ResNet-18 model from
+    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    return _resnet("resnet18", BasicBlock, [2, 2, 2, 2], pretrained, progress, **kwargs)
+
+
+def resnet34(pretrained=False, progress=True, **kwargs):
+    r"""ResNet-34 model from
+    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    return _resnet("resnet34", BasicBlock, [3, 4, 6, 3], pretrained, progress, **kwargs)
+
+
+def resnet50(pretrained=False, progress=True, **kwargs):
+    r"""ResNet-50 model from
+    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    return _resnet("resnet50", Bottleneck, [3, 4, 6, 3], pretrained, progress, **kwargs)
+
+
+def resnet101(pretrained=False, progress=True, **kwargs):
+    r"""ResNet-101 model from
+    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    return _resnet(
+        "resnet101", Bottleneck, [3, 4, 23, 3], pretrained, progress, **kwargs
+    )
+
+
+def resnet152(pretrained=False, progress=True, **kwargs):
+    r"""ResNet-152 model from
+    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    return _resnet(
+        "resnet152", Bottleneck, [3, 8, 36, 3], pretrained, progress, **kwargs
+    )
+
+
+def resnext50_32x4d(pretrained=False, progress=True, **kwargs):
+    r"""ResNeXt-50 32x4d model from
+    `"Aggregated Residual Transformation for Deep Neural Networks" <https://arxiv.org/pdf/1611.05431.pdf>`_
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    kwargs["groups"] = 32
+    kwargs["width_per_group"] = 4
+    return _resnet(
+        "resnext50_32x4d", Bottleneck, [3, 4, 6, 3], pretrained, progress, **kwargs
+    )
+
+
+def resnext101_32x8d(pretrained=False, progress=True, **kwargs):
+    r"""ResNeXt-101 32x8d model from
+    `"Aggregated Residual Transformation for Deep Neural Networks" <https://arxiv.org/pdf/1611.05431.pdf>`_
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    kwargs["groups"] = 32
+    kwargs["width_per_group"] = 8
+    return _resnet(
+        "resnext101_32x8d", Bottleneck, [3, 4, 23, 3], pretrained, progress, **kwargs
+    )
+
+
+def wide_resnet50_2(pretrained=False, progress=True, **kwargs):
+    r"""Wide ResNet-50-2 model from
+    `"Wide Residual Networks" <https://arxiv.org/pdf/1605.07146.pdf>`_
+
+    The model is the same as ResNet except for the number of bottleneck channels
+    which is twice larger in every block. The number of channels in outer 1x1
+    convolutions is the same, e.g., the last block in ResNet-50 has 2048-512-2048
+    channels, while in wide ResNet-50-2 there are 2048-1024-2048.
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    kwargs["width_per_group"] = 64 * 2
+    return _resnet(
+        "wide_resnet50_2", Bottleneck, [3, 4, 6, 3], pretrained, progress, **kwargs
+    )
+
+
+def wide_resnet101_2(pretrained=False, progress=True, **kwargs):
+    r"""Wide ResNet-101-2 model from
+    `"Wide Residual Networks" <https://arxiv.org/pdf/1605.07146.pdf>`_
+
+    The model is the same as ResNet except for the number of bottleneck channels
+    which is twice larger in every block. The number of channels in outer 1x1
+    convolutions is the same, e.g., the last block in ResNet-101 has 2048-512-2048
+    channels, while in wide ResNet-101-2 there are 2048-1024-2048.
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on the ImageNet dataset.
+        progress (bool): If True, displays a progress bar of the download to stderr.
+    """
+    kwargs["width_per_group"] = 64 * 2
+    return _resnet(
+        "wide_resnet101_2", Bottleneck, [3, 4, 23, 3], pretrained, progress, **kwargs
+    )