diff --git a/parse/train/2vubO341F_E/2vubO341F_E.md b/parse/train/2vubO341F_E/2vubO341F_E.md new file mode 100644 index 0000000000000000000000000000000000000000..f326b41e0b92f4034bf0451bc29059ad0e8ba8f4 --- /dev/null +++ b/parse/train/2vubO341F_E/2vubO341F_E.md @@ -0,0 +1,255 @@ +# All Tokens Matter: Token Labeling for Training Better Vision Transformers + +Zihang Jiang1∗ Qibin Hou2,1† Li Yuan3 Daquan Zhou1 Yujun Shi1 + +Xiaojie Jin4 Anran Wang4 Jiashi Feng4 + +1National University of Singapore 2Nankai University + +# 3 Peking University 4ByteDance + +{jzh0103,andrewhoux,ylustcnus,zhoudaquan21,shiyujun1016}@gmail.com xjjin0731@gmail.com, anran.wang@bytedance.com, jshfeng@gmail.com + +# Abstract + +In this paper, we present token labeling—a new training objective for training high-performance vision transformers (ViTs). Different from the standard training objective of ViTs that computes the classification loss on an additional trainable class token, our proposed one takes advantage of all the image patch tokens to compute the training loss in a dense manner. Specifically, token labeling reformulates the image classification problem into multiple token-level recognition problems and assigns each patch token with an individual location-specific supervision generated by a machine annotator. Experiments show that token labeling can clearly and consistently improve the performance of various ViT models across a wide spectrum. For a vision transformer with 26M learnable parameters serving as an example, with token labeling, the model can achieve $8 4 . 4 \%$ Top-1 accuracy on ImageNet. The result can be further increased to $8 6 . 4 \%$ by slightly scaling the model size up to 150M, delivering the minimal-sized model among previous models $( 2 5 0 \mathbf { M } + )$ reaching $86 \%$ . We also show that token labeling can clearly improve the generalization capability of the pretrained models on downstream tasks with dense prediction, such as semantic segmentation. Our code and model are publicly available at https://github.com/zihangJiang/TokenLabeling. + +# 1 Introduction + +Transformers [39] have achieved great performance for almost all the natural language processing (NLP) tasks over the past years [4, 14, 24]. Motivated by such success, recently, many researchers attempt to build transformer models for vision tasks, and their encouraging results have shown the great potential of transformer based models for image classification [6, 15, 25, 36, 40, 46], especially the strong benefits of the self-attention mechanism in building long-range dependencies between pairs of input tokens. + +Despite the importance of gathering long-range dependencies, recent work on local data augmentation [57] has demonstrated that well modeling and leveraging local information for image classification would avoid biasing the model towards skewed and non-generalizable patterns and substantially improve the model performance. However, recent vision transformers normally utilize class tokens that aggregate global information to predict the output class while neglecting the role of other patch tokens that encode rich information on their respective local image patches. + +![](images/9c7bb663ccabca3baefdad385c72c84db533d97dbd497ac18638700c9e1d548b.jpg) +Figure 1: Comparison between the proposed LV-ViT and other recent works based on vision transformers, including T2T-ViT [46], ConViT [12], BoTNet [31], DeepViT [59], DeiT [36], ViT [15], Swin Transformer [25], LambdaNet [1], CvT [43], CrossViT [6], PVT [40], CaiT [37]. Note that we only show models whose model sizes are under 100M. As can be seen, our LV-ViT achieves the best results using the least amount of learnable parameters. The default test resolution is $2 2 4 \times 2 2 4$ unless specified after $@$ . + +In this paper, we present a new training objective for vision transformers, termed token labeling, that takes advantage of both the patch tokens and the class tokens. Our method takes a $K$ -dimensional score map generated by a machine annotator as supervision to supervise all the tokens in a dense manner, where $K$ is the number of categories for the target dataset. In this way, each patch token is explicitly associated with an individual location-specific supervision indicating the existence of the target objects inside the corresponding image patch, so as to improve the object grounding and recognition capabilities of vision transformers with negligible computation overhead. To the best of our knowledge, this is the first work demonstrating that dense supervision is beneficial to vision transformers in image classification. + +According to our experiments, utilizing the proposed token labeling objective can clearly boost the performance of vision transformers. As shown in Figure 1, our model, named LV-ViT, with 56M parameters, yields $8 5 . 4 \%$ top-1 accuracy on ImageNet [13], behaving better than all the other transformer-based models having no more than 100M parameters. When the model size is scaled up to 150M, the result can be further improved to $8 6 . 4 \%$ . In addition, we have empirically found that the pretrained models with token labeling are also beneficial to downstream tasks with dense prediction, such as semantic segmentation. + +# 2 Related Work + +Transformers [39] refer to the models that entirely rely on the self-attention mechanism to build global dependencies, which are originally designed for natural language processing tasks. Due to their strong capability of capturing spatial information, transformers have also been successfully applied to a variety of vision problems, including low-level vision tasks like image enhancement [7, 45], as well as more challenging tasks such as image classification [9, 15], object detection [5, 11, 55, 61], segmentation [7, 33, 41] and image generation [28]. Some works also extend transformers for video and 3D point cloud processing [50, 53, 60]. + +Vision Transformer (ViT) is one of the earlier attempts that achieved state-of-the-art performance on ImageNet classification, using pure transformers as basic building blocks. However, ViTs need pretraining on very large datasets, such as ImageNet-22k and JFT-300M, and huge computation resources to achieve comparable performance to ResNet [18] with a similar model size trained on ImageNet. Later, DeiT [36] manages to tackle the data-inefficiency problem by simply adjusting the network architecture and adding an additional token along with the class token for Knowledge Distillation [21, 47] to improve model performance. + +![](images/77890e011e8cc08c2d9f50310be754cb08f7894ee6cb766ac590a7580f42e616.jpg) +Figure 2: Pipeline of training vision transformers with token labeling. Other than utilizing the class token (pink rectangle), we also take advantage of all the output patch tokens (orange rounded rectangle) by assigning each patch token an individual location-specific prediction generated by a machine annotator [3] as supervision (see the part in the red dash rectangle). Our proposed token labeling method can be treated as an auxiliary objective to provide each patch token the local details that aid vision transformers to more accurately locate and recognize the target objects. Note that the traditional vision transformer training does not include the red dash rectangle part. + +Some recent works [6, 16, 43, 46] also attempt to introduce the local dependency into vision transformers by modifying the patch embedding block or the transformer block or both, leading to significant performance gains. Moreover, there are also some works [20, 25, 40] adopting a pyramid structure to reduce the overall computation while maintaining the model’s ability to capture low-level features. + +Unlike most aforementioned works that design new transformer blocks or transformer architectures, we attempt to improve vision transformers by studying the role of patch tokens that embed rich local information inside image patches. We show that by slightly tuning the structure of vision transformers and employing the proposed token labeling objective, we can achieve strong baselines for transformer models at different model size levels. + +# 3 Token Labeling Method + +In this section, we first briefly review the structure of the vision transformer [15] and then describe the proposed training objective—token labeling. + +# 3.1 Revisiting Vision Transformer + +A typical vision transformer [15] first decomposes a fixed-size input image into a sequence of small patches. Each small patch is mapped to a feature vector, or called a token, by projection with a linear layer. Then, all the tokens combined with an additional learnable class token for classification score prediction are sent into a stack of transformer blocks for feature encoding. + +In loss computing, the class token from the output tokens of the last transformer block is usually selected and sent into a linear layer for the classification score prediction. Mathematically, given an image $I$ , denote the output of the last transformer block as $[ \bar { X } ^ { c l s } , X ^ { 1 } , . . . , X ^ { N } ]$ , where $N$ is the total number of patch tokens, and $X ^ { c l s }$ and $X ^ { 1 } , . . . , X ^ { N }$ correspond to the class token and the patch tokens, respectively. The classification loss for image $I$ can be written as + +$$ +L _ { c l s } = H ( { X } ^ { c l s } , { y } ^ { c l s } ) , +$$ + +where $H ( \cdot , \cdot )$ is the softmax cross-entropy loss and $y ^ { c l s }$ is the class label. + +![](images/c741b481df54f67e411d3cc1b30dc7deb904c80d9d11e57036dbf5b27cdbfb8f.jpg) +Figure 3: Comparison between CutMix [48] (Left) and our proposed MixToken (Right). CutMix is operated on the input images. This results in patches containing mixed regions from the two images (see the patches enclosed by red bounding boxes). Differently, MixToken targets at mixing tokens after patch embedding. This enables each token after patch embedding to have clean content as shown in the right part of this figure. The detailed advantage of MixToken can be found in Sec. 4.2. + +# 3.2 Token Labeling + +The above classification problem only adopts an image-level label as supervision whereas it neglects the rich information embedded in each image patch. In this subsection, we present a new training objective—token labeling—that takes advantage of the complementary information between the patch tokens and the class tokens. + +Token Labeling: Different from the classification loss as formulated in Eqn. (1) that measures the distance between the single class token (representing the whole input image) and the corresponding image-level label, token labeling emphasizes the importance of all output tokens and advocates that each output token should be associated with an individual location-specific label. Therefore, in our method, the ground truth for an input image involves not only a single $K$ -dimensional vector $y ^ { c l s }$ but also a $K \times N$ matrix or called a $K$ -dimensional score map as represented by $[ y ^ { 1 } , . . . , y ^ { N } ]$ , where $N$ is the number of the output patch tokens. + +Specifically, we leverage a dense score map for each training image and use the cross-entropy loss between each output patch token and the corresponding aligned label in the dense score map as an auxiliary loss at the training phase. Figure 2 provides an intuitive interpretation. Given the output patch tokens $X ^ { 1 } , . . . , X ^ { N }$ and the corresponding labels $[ y ^ { 1 } , . . . , y ^ { N } ]$ , the token labeling objective can be defined as + +$$ +L _ { t l } = \frac { 1 } { N } \sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) . +$$ + +Recall that $H$ is the cross-entropy loss. Therefore, the total loss function can be written as + +$$ +\begin{array} { l } { { { \cal L } _ { t o t a l } = H ( X ^ { c l s } , y ^ { c l s } ) + \beta \cdot L _ { t l } , } } \\ { { { } } } \\ { { = H ( X ^ { c l s } , y ^ { c l s } ) + \beta \cdot \displaystyle \frac { 1 } { N } \sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) , } } \end{array} +$$ + +where $\beta$ is a hyper-parameter to balance the two terms. In our experiment, we empirically set it to 0.5. + +Advantages: Our token labeling offers the following advantages. First of all, unlike knowledge distillation methods that require a teacher model to generate supervision labels online, token labeling is a cheap operation. The dense score map can be generated by a pretrained model in advance (e.g., EfficientNet [34] or NFNet [3]). During training, we only need to crop the score map and perform interpolation to make it aligned with the cropped image in the spatial coordinate. Thus, the additional computations are negligible. Second, rather than utilizing a single label vector as supervision as done in most classification models and the ReLabel strategy [49], we also harness score maps to supervise the models in a dense manner and thereby the label for each patch token provides location-specific information, which can aid the training models to easily discover the target objects and improve the recognition accuracy. Last but not the least, as dense supervision is adopted in training, we found that the pretrained models with token labeling benefit downstream tasks with dense prediction, like semantic segmentation. + +# 3.3 Token Labeling with MixToken + +While training vision transformer, previous studies [36, 46] have shown that augmentation methods, like MixUp [52] and CutMix [48], can effectively boost the performance and robustness of the models. However, vision transformers rely on patch-based tokenization to map each input image to a sequence of tokens and our token labeling strategy also operates on patch-based token labels. If we apply CutMix directly on the raw image, some of the resulting patches may contain content from two images, leading to mixed regions within a small patch as shown in Figure 3. When performing token labeling, it is difficult to assign each output token a clean and correct label. Taking this situation into account, we rethink the CutMix augmentation method and present MixToken, which can be viewed as a modified version of CutMix operating on the tokens after patch embedding as illustrated in the right part of Figure 3. + +To be specific, for two images denoted as $I _ { 1 } , I _ { 2 }$ and their corresponding token labels $Y _ { 1 } = [ y _ { 1 } ^ { 1 } , . . . , y _ { 1 } ^ { N } ]$ as well as $Y _ { 2 } = [ y _ { 2 } ^ { 1 } , . . . , y _ { 2 } ^ { N } ]$ , we first feed the two images into the patch embedding module to tokenize each as a sequence of tokens, resulting in $T _ { 1 } = [ t _ { 1 } ^ { 1 } , . . . , t _ { 1 } ^ { N } ]$ and $\bar { T _ { 2 } } = [ t _ { 2 } ^ { 1 } , . . . , t _ { 2 } ^ { N } ]$ . Then, we produce a new sequence of tokens by applying MixToken using a binary mask $M$ as follows: + +$$ +\hat { T } = T _ { 1 } \odot M + T _ { 2 } \odot ( 1 - M ) , +$$ + +where $\odot$ is element-wise multiplication. We use the same way to generate the mask $M$ as in [48]. For the corresponding token labels, we also mix them using the same mask $M$ : + +$$ +\hat { Y } = Y _ { 1 } \odot M + Y _ { 2 } \odot ( 1 - M ) . +$$ + +The label for the class token can be written as + +$$ +y ^ { \hat { c } l s } = \bar { M } y _ { 1 } ^ { c l s } + ( 1 - \bar { M } ) y _ { 2 } ^ { c l s } , +$$ + +where $\bar { M }$ is the average of all element values of $M$ . + +# 4 Experiments + +# 4.1 Experiment Setup + +We evaluate our method on the ImageNet [13] dataset. All experiments are built and conducted upon PyTorch [29] and the timm [42] library. We follow the standard training schedule and train our models on the ImageNet dataset for 300 epochs. Besides normal augmentations like CutOut [57] and RandAug [10], we also explore the effect of applying MixUp [52] and CutMix [48] together with our proposed token labeling. Empirically, we have found that using MixUp together with token labeling brings no benefit to the performance, and thus we do not apply it in our experiments. + +For optimization, by default, we use the AdamW optimizer [27] with a linear learning rate scaling strategy $\begin{array} { r } { l r = 1 0 ^ { - 3 } \times \frac { b a t c h \_ s i z e } { 6 4 0 } } \end{array}$ and $5 \times 1 0 ^ { - 2 }$ weight decay rate. For Dropout regularization, we observe that for small models, using Dropout hurts the performance. This has also been observed in a few other works related to training vision transformers [36, 37, 46]. As a result, we do not apply Dropout [32] and use Stochastic Depth [23] instead. More details on hyper-parameters and finetuning can be found in our supplementary materials. + +We use the NFNet-F6 [3] trained on ImageNet with an $8 6 . 3 \%$ Top-1 accuracy as the machine annotator to generate dense score maps for the ImageNet dataset, yielding a 1000-dimensional score map for each image for training. The score map generation procedure is similar to [49], but we limit our experiment setting by training all models from scratch on ImageNet without extra data support, such as JFT-300M and ImageNet-22K. This is different from the original ReLabel paper [49], in which the EfficientNet-L2 model pretrained on JFT-300M is used. The input resolution for NFNet-F6 is $5 7 6 \times 5 7 6$ , and the dimension of the corresponding output score map for each image is $L \in \mathbb { R } ^ { 1 8 \times 1 8 \times 1 0 0 0 }$ . During training, the target labels for the tokens are generated by applying RoIAlign [17] on the corresponding score map. In practice, we only store the top-5 score maps for each position in half-precision to save space as storing the entire score maps for all the images results in 2TB storage. In our experiment, we only need 10GB of storage to store all the score maps. + +Table 1: Performance of the proposed LV-ViT with different model sizes. Here, ‘depth’ denotes the number of transformer blocks used in different models. By default, the test resolution is set to $2 2 4 \times 2 2 4$ except the last one which is $2 8 8 \times 2 8 8$ . + +
NameDepthEmbed dim.MLP Ratio#Heads#ParamsThroughput (im/s)Test sizeTop-1 Acc. (%)
LV-ViT-T122403.048.5M2032.622479.1
LV-ViT-S163843.0626M1018.222483.3
LV-ViT-M205123.0856M668.922484.1
LV-ViT-L247683.012150M204.828885.3
+ +# 4.2 Ablation Analysis + +Model Settings: The default settings of the proposed LV-ViT are given in Table 1, where both token labeling and MixToken are used. A slight architecture modification to ViT [15] is that we replace the patch embedding module with a 4-layer convolution to better tokenize the input image and integrate local information. Detailed ablation about patch embedding can be found in our supplementary materials. As can be seen, our LV-ViT-T with only $8 . 5 { \bf M }$ parameters can already achieve a top-1 accuracy of $7 9 . 1 \%$ on ImageNet. Increasing the embedding dimension and network depth can further boost the performance. More experiments compared to other methods can be found in Sec. 4.3. In the following ablation experiments, we will set our LV-ViT-S as baseline and show the advantages of the proposed token labeling and MixToken methods. + +MixToken: We use MixToken as a substitution for CutMix while applying token labeling. Our experiments show that MixToken performs better than CutMix for token-based transformer models. As shown in Table 2, when training with the original ImageNet labels, using MixToken is $0 . 1 \%$ higher than using CutMix. When using the ReLabel supervision, we can also see an advantage of $0 . { \bar { 2 } } \%$ over the CutMix baseline. Combining with our token labeling, the performance can be further raised to $8 3 . 3 \%$ . + +Table 2: Ablation on the proposed MixToken and token labeling augmentations. We also show results with either the ImageNet hard label and the ReLabel [49] as supervision. + +
Aug. MethodSupervisionTop-1 Acc.
MixTokenToken labeling83.3
MixTokenReLabel83.0
CutMixReLabel82.8
MixtokenImageNet Label82.5
CutMixImageNet Label82.4
+ +Table 3: Ablation on different widely-used data augmentations. We have empirically found our proposed MixToken performs even better than the combination of MixUp and CutMix in vision transformers. + +
MixTokenMixUpCutOutRandAugTop-1 Acc.
83.3
81.3
>x>xx>>>>>>>83.1
X√83.0
广XXX82.8
+ +Data Augmentation: Here, we study the compatibility of MixToken with other augmentation techniques, such as MixUp [52], CutOut [57] and RandAug [10]. The ablation results are shown in Table 3. We can see when all the four augmentation methods are used, a top-1 accuracy of $8 3 . 1 \%$ is achieved. Interestingly, when the MixUp augmentation is removed, the performance can be improved to $8 3 . 3 \%$ . This may be explained as, using MixToken and MixUp at the same time would bring too much noise in the label, and consequently cause confusion of the model. Moreover, the CutOut augmentation, which randomly erases some parts of the image, is also effective and removing it brings a performance drop of $\dot { 0 } . 3 \%$ . Similarly, the RandAug augmentation also contributes to the performance and using it brings an improvement of $0 . 5 \%$ . + +All Tokens Matter: To show the importance of involving all tokens in our token labeling method, we attempt to randomly drop some tokens and use the remaining ones for computing the token labeling loss. The percentage of the remaining tokens is denoted as Token Participation Rate. As shown in Figure 4 (Left), we conduct experiments on two models: LV-ViT-S and LV-ViT-M. As can be seen, using only $2 0 \%$ of the tokens to compute the token labeling loss decreases the performance $( - 0 . 5 \%$ for LV-ViT-S and $- 0 . 4 \%$ for LV-ViT-M). Involving more tokens for loss computation consistently leads to better performance. Since involving all tokens brings negligible computation cost and gives the best performance, we always set the token participation rate as $\mathrm { \bar { 1 0 0 \% } }$ in the following experiments. + +![](images/810578c2e36580b46aa98e4ec01030b20f6e00994c33dafcaa4cde435beb4d46.jpg) +Figure 4: Left: LV-ViT ImageNet Top-1 Accuracy w.r.t. the token participation rate while applying token labeling. Token participation rate indicates the percentage of patch tokens involved in computing the token labeling loss. This experiment reflects that all tokens matter for vision transformers. Right: LV-ViT-S ImageNet Top-1 Accuracy w.r.t. different annotator models. The point size indicates the parameter number of the annotator model. Clearly, our token labeling objective is robust to different annotator models. + +Table 4: Comparison of token labeling (TL), knowledge distillation (KD) based method and ReLabel method based on utilized tokens, DeiT-S/LV-ViT-S Top-1 accuracy on ImageNet validation set and training time on a single V100 GPU node. + +
MethodOnline KDOnline TLTLReLabelVanilla
Tokens Utilized2AllAll11
DeiT-S Acc. (%)81.281.881.080.479.9
LV-ViT-S Acc. (%)83.083.583.382.882.4
Training Time (8× V100)63 hrs63 hrs45 hrs45 hrs41 hrs
+ +Online Token Labeling: Unlike the online knowledge distillation method which generates labels by a teacher model online, our token labeling approach utilizes the dense label map generated in advance and directly applies the corresponding augmentation methods, such as random crop, on the label map to obtain token-level labels. To directly compare with the online knowledge distillation based method and validate the effectiveness of token-level supervision, we further conduct experiments on the online version of our token labeling method, which generates token-level labels online during training. Following DeiT [36], we use RegNetY-16GF [30] as the online teacher model. Results in terms of DeiT-S/LV-ViT-S Top-1 accuracy and training time for our token labeling, online knowledge distillation, and ReLabel [49] are listed in Table 4, with number of utilized tokens also included for clear comparison. As can be seen, for both online and offline cases, using token-level supervision can improve the overall performance with only negligible additional training cost. Meanwhile, compared to the vanilla training baseline, our proposed offline token labeling brings almost no additional training cost, and boosts the overall performance of LV-ViT-S by $0 . 9 \%$ , which well demonstrates its efficiency and effectiveness. + +Robustness to Different Annotators: To evaluate the robustness of our token labeling method, we use different pretrained CNNs, including EfficientNet-B3,B4,B5,B6,B7,B8 [34], NFNet-F6 [3] and ResNest269E [51], as annotator models to provide dense supervision. Results are shown in the right part of Figure 4. We can see that, even if we use an annotator with relatively lower performance, such as EfficientNet-B3 whose Top-1 accuracy is $8 1 . 6 \%$ , it can still provide multi-label location-specific supervision and help improve the performance of our LV-ViT-S model. Meanwhile, annotator models with better performance can provide more accurate supervision, bringing even better performance, as stronger annotator models can generate better token-level labels. The largest annotator NFNet-F6 [3], which has the best performance of $8 6 . 3 \%$ , allows us to achieve the best result for LV-ViT-S, which is $8 3 . 3 \%$ . In addition, we also attempt to use a better model, EfficientNet-L2 pretrained on JFT-300M as described in [49] which has $8 8 . 2 \%$ Top-1 ImageNet accuracy, as our annotator. The performance of LV-ViT-S can be further improved to $8 \mathrm { { 3 . 5 \% } }$ . However, to fairly compare with the models without extra training data, we only report results based on dense supervision produced by NFNet-F6 [3] that uses only ImageNet training data. + +![](images/a3d559defddf629a8a69e1ea7f71228daa68e20f028756c9b3671f019cf5ad40.jpg) +Figure 5: Performance of the proposed token labeling objective on three different vision transformers: DeiT [36] (Left), T2T-ViT [46] (Middle), and LV-ViT (Right). Our method has a consistent improvement on all 7 different ViT models. + +Robustness to Different ViT Variants: To further evaluate the robustness of our token labeling, we train different transformer-based networks, including DeiT [36], T2T-ViT [3] and our model LV-ViT, with the proposed training objective. Results are shown in Figure 5. It can be found that, all the models trained with token labeling consistently outperform their vanilla counterparts, demonstrating the robustness of token labeling with respect to different variants of patch-based vision transformers. Meanwhile, for different scales of the models, the improvement is also consistent. Interestingly, we observe larger improvements for larger models. These indicate that our proposed token labeling method is widely applicable to a large range of patch-based vision transformer variants. + +Beyond Vision Transformers: We further explore the performance of token labeling on other CNN-based and MLP-based models. Results are shown in Table 5. Besides our re-implementation with more data augmentation and regularization techniques, we also provide the results from the original papers. It can be found that for both MLP-based and CNN-based models, our token labeling objective can also improve the performance over strong baselines by providing location-specific dense supervision. + +Table 5: Performance of the proposed token labeling objective on representative CNN-based (ResNeSt) and MLP-based (Mixer-MLP) models. Our method has a consistent improvement on all different models. Here † indicates results reported in original papers. + +
ModelMixer-S/16 [35]Mixer-B/16 [35]Mixer-L/16 [35]ResNeSt-50 [51]
Token LabelingXX√×X√××√××√
Parameters18M18M18M59M59M 59M207M207M207M27M27M 27M
Top-1 Acc. (%)73.8t75.676.176.478.379.571.6t77.780.181.1t80.981.5
+ +# 4.3 Comparison to Other Methods + +We compare our proposed model LV-ViT with other state-of-the-art methods in Table 6. For smallsized models, when the test resolution is set to $2 2 4 \times 2 2 4$ , we achieve an $8 3 . 3 \%$ accuracy on ImageNet with only 26M parameters, which is $3 . 4 \%$ higher than the strong baseline DeiT-S [36]. For mediumsized models, when the test resolution is set to $3 8 4 \times 3 8 4$ we achieve the performance of $8 5 . 4 \%$ , the same as CaiT-S36 [37], but with much less computational cost and parameters. Note that both DeiT and CaiT use knowledge distillation to improve their models, which introduce much more computations in training. However, we do not require any extra computations in training and only have to compute and store the dense score maps in advance. For large-sized models, our LV-ViT-L with a test resolution of $4 4 8 \times 4 4 8$ achieves an $8 6 . 2 \%$ top-1 accuracy, which is comparable to CaiT-M36 [37] but with far fewer FLOPs and parameters. + +Table 6: Top-1 accuracy comparison with other methods on ImageNet [13] and ImageNet Real [2]. All models are trained without external data. With the same computation and parameter constraint, our model consistently outperforms other CNN-based and transformer-based counterparts. The results of CNNs and ViT are referenced from [37]. + +
NetworkParamsFLOPsTrain sizeTest sizeTop-1(%)Real Top-1 (%)
EfficientNet-B5 [34] SNNO30M9.9B45645683.688.3
EfficientNet-B7 [34]66M37.0B60060084.3
Fix-EfficientNet-B8 [34,38]87M89.5B67280085.790.0
NFNet-F3 [3]255M114.8B32041685.789.4
NFNet-F4 [3]316M215.3B38451285.989.4
NFNet-F5[3]377M289.8B41654486.089.2
ViT-B/16 [15]86M55.4B22438477.983.6
ViT-L/16[15]307M190.7B22438476.582.2
T2T-ViT-14 [46]22M5.2B22422481.5
T2T-ViT-14↑384 [46]22M17.1B22438483.31
Cross ViT [6]45M56.6B22448084.1一
Swin-B[25]88M47.0B22438484.2
TNT-B[16]66M14.1B22422482.8
iriirrrrrs DeepViT-S [59]27M6.2B22422482.3
DeepViT-L [59]55M12.5B22422483.1
DeiT-S[36]22M4.6B22422479.985.7
Distilled DeiT-S [36]22M4.6B22422481.286.8
DeiT-B [36]86M17.5B22422481.886.7
DeiT-B↑384 [36]86M55.4B22438483.187.7
Distilled DeiT-B [36]87M17.5B22422483.488.3
BoTNet-S1-128 [31]79.1M19.3B25625684.2
BoTNet-S1-128↑384 [31]79.1M45.8B25638484.7-
CaiT-S36↑384 [37]68M48.0B22438485.4- 89.8
CaiT-M36[37]271M53.7B22422485.189.3
CaiT-M36↑448 [37]271M247.8B22444886.390.2
LV-ViT-S26M6.6B22422483.3
江 LV-ViT-S↑38426M22.2B22438484.488.1 88.9
LV-ViT-M56M16.0B22422484.188.4
W LV-ViT-M↑38456M42.2B22438485.489.5
LV-ViT-L150M59.0B28828885.389.3
0 LV-ViT-L↑448150M157.2B28844885.989.7
LV-ViT-L↑448150M157.2B44844886.289.9
LV-ViT-L↑512151M214.8B44851286.490.1
+ +# 4.4 Semantic Segmentation on ADE20K + +It has been shown in [19] that different training techniques for pretrained models have different impacts on downstream tasks with dense prediction, like semantic segmentation. To demonstrate the advantage of the proposed token labeling objective on tasks with dense prediction, we apply our pretrained LV-ViT with token labeling to the semantic segmentation task. + +Similar to previous work [25], we run experiments on the widely-used ADE20K [58] dataset. ADE20K contains 25K images in total, including 20K images for training, 2K images for validation and 3K images for test, and covering 150 different foreground categories. We take both FCN [26] and UperNet [44] as our segmentation frameworks and use the mmseg toolbox to implement. During training, following [25], we use the AdamW optimizer with an initial learning rate of 6e-5 and a weight decay of 0.01. We also use a linear learning schedule with a minimum learning rate of 5e-6. All models are trained on 8 GPUs and with a batch size of 16 (i.e., 2 images on each GPU). The input resolution is set to $5 1 2 \times 5 1 2$ . In inference, a multi-scale test with interpolation rates of [0.75, 1.0, 1.25, 1.5, 1.75] is used. As suggested by [58], we report results in terms of both mean intersection-over-union (mIoU) and the average pixel accuracy (Pixel Acc.). + +In Table 7, we test the performance of token labeling on both FCN and UperNet frameworks. The FCN framework has a light convolutional head and can directly reflect the performance of the pretrained models in terms of transferable capability. As can be seen, pretrained models with token labeling perform better than those without token labeling. This indicates token labeling is indeed beneficial to semantic segmentation. + +Table 7: Transfer performance of the proposed LV-ViT in semantic segmentation. We take two classic methods, FCN and UperNet, as segmentation architectures and show both single-scale (SS) and multi-scale (MS) results on the validation set. + +
MethodToken LabelingModel SizemIoU (SS)P. Acc. (SS)mIoU (MS)P. Acc. (MS)
LV-ViT-S + FCN×30M46.181.947.382.6
LV-ViT-S + FCN30M47.282.448.483.0
LV-ViT-S + UperNet44M46.582.147.682.7
LV-ViT-S + UperNetX44M47.982.648.683.1
+ +We also compare our segmentation results with previous state-of-the-art segmentation methods in Table 8. Without pretraining on large-scale datasets such as ImageNet-22K, our LV-ViT-M with the UperNet segmentation architecture achieves an mIoU score of 50.6 with only 77M parameters. This result is much better than the previous CNN-based and transformer-based models. Furthermore, using our LV-ViT-L as the pretrained model yields a better result of 51.8 in terms of mIoU. As far as we know, this is the best result reported on ADE20K with no pretraining on ImageNet-22K or other large-scale datasets. + +Table 8: Comparison with previous work on ADE20K validation set. As far as we know, our LVViT-L $^ +$ UperNet achieves the best result on ADE20K with only ImageNet-1K as training data in pretraining. †Pretrained on ImageNet-22K. + +
BackboneSegmentation ArchitectureModel SizemIoU (MS)Pixel Acc. (MS)
SNNOResNet-269PSPNet [54]44.981.7
ResNet-101UperNet [44]86M44.9-
ResNet-101Strip Pooling [22]45.682.1
ResNeSt200DeepLabV3+ [8]88M48.41
DeiT-SUperNet52M44.0-
TirriiirrssViT-LargetSETR [56]308M50.383.5
Swin-T[25]UperNet60M46.11
Swin-S [25]UperNet81M49.3=
Swin-B [25]UperNet121M49.7
Swin-B† [25]UperNet121M51.6-
LIA-ATLV-ViT-SFCN30M48.483.0
LV-ViT-SUperNet44M48.683.1
LV-ViT-MUperNet77M50.683.5
LV-ViT-LUperNet209M51.884.1
+ +# 5 Conclusions and Discussion + +In this paper, we introduce a new token labeling method to help improve the performance of vision transformers. We also analyze the effectiveness and robustness of our token labeling with respect to different annotators and different variants of patch-based vision transformers. By applying token labeling, our proposed LV-ViT achieves $8 4 . 4 \%$ Top-1 accuracy with only 26M parameters and $8 6 . 4 \%$ Top-1 accuracy with 150M parameters on ImageNet-1K benchmark. + +Despite the effectiveness, token labeling has a limitation of requiring a pretrained model as the machine annotator. Fortunately, the machine annotating procedure can be done in advance to avoid introducing extra computational cost in training. This makes our method quite different from knowledge distillation methods that rely on online teaching. For users with limited machine resources on hand, our token labeling provides a promising training technique to improve the performance of vision transformers. + +References +[1] Irwan Bello. Lambdanetworks: Modeling long-range interactions without attention. arXiv preprint arXiv:2102.08602, 2021. +[2] Lucas Beyer, Olivier J Hénaff, Alexander Kolesnikov, Xiaohua Zhai, and Aäron van den Oord. Are we done with imagenet? arXiv preprint arXiv:2006.07159, 2020. +[3] Andrew Brock, Soham De, Samuel L Smith, and Karen Simonyan. High-performance large-scale image recognition without normalization. arXiv preprint arXiv:2102.06171, 2021. +[4] Tom B Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. arXiv preprint arXiv:2005.14165, 2020. +[5] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-end object detection with transformers. arXiv preprint arXiv:2005.12872, 2020. +[6] Chun-Fu Chen, Quanfu Fan, and Rameswar Panda. Crossvit: Cross-attention multi-scale vision transformer for image classification. arXiv preprint arXiv:2103.14899, 2021. +[7] Hanting Chen, Yunhe Wang, Tianyu Guo, Chang Xu, Yiping Deng, Zhenhua Liu, Siwei Ma, Chunjing Xu, Chao Xu, and Wen Gao. Pre-trained image processing transformer. arXiv preprint arXiv:2012.00364, 2020. +[8] Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. Encoder-decoder with atrous separable convolution for semantic image segmentation. In Proceedings of the European conference on computer vision (ECCV), pages 801–818, 2018. +[9] Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, and Ilya Sutskever. Generative pretraining from pixels. In International Conference on Machine Learning, pages 1691–1703. PMLR, 2020. +[10] Ekin D Cubuk, Barret Zoph, Jonathon Shlens, and Quoc V Le. Randaugment: Practical automated data augmentation with a reduced search space. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pages 702–703, 2020. +[11] Zhigang Dai, Bolun Cai, Yugeng Lin, and Junying Chen. Up-detr: Unsupervised pre-training for object detection with transformers. arXiv preprint arXiv:2011.09094, 2020. +[12] Stéphane d’Ascoli, Hugo Touvron, Matthew Leavitt, Ari Morcos, Giulio Biroli, and Levent Sagun. Convit: Improving vision transformers with soft convolutional inductive biases. arXiv preprint arXiv:2103.10697, 2021. +[13] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. +[14] Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018. +[15] Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. +[16] Kai Han, An Xiao, Enhua Wu, Jianyuan Guo, Chunjing Xu, and Yunhe Wang. Transformer in transformer. arXiv preprint arXiv:2103.00112, 2021. +[17] Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. Mask r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 2961–2969, 2017. +[18] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. +[19] Tong He, Zhi Zhang, Hang Zhang, Zhongyue Zhang, Junyuan Xie, and Mu Li. Bag of tricks for image classification with convolutional neural networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 558–567, 2019. +[20] Byeongho Heo, Sangdoo Yun, Dongyoon Han, Sanghyuk Chun, Junsuk Choe, and Seong Joon Oh. Rethinking spatial dimensions of vision transformers. arXiv preprint arXiv:2103.16302, 2021. +[21] Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531, 2015. +[22] Qibin Hou, Li Zhang, Ming-Ming Cheng, and Jiashi Feng. Strip pooling: Rethinking spatial pooling for scene parsing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4003–4012, 2020. +[23] Gao Huang, Yu Sun, Zhuang Liu, Daniel Sedra, and Kilian Q Weinberger. Deep networks with stochastic depth. In European conference on computer vision, pages 646–661. Springer, 2016. +[24] Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692, 2019. +[25] Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. arXiv preprint arXiv:2103.14030, 2021. +[26] Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3431–3440, 2015. +[27] Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. +[28] Niki Parmar, Ashish Vaswani, Jakob Uszkoreit, Łukasz Kaiser, Noam Shazeer, Alexander Ku, and Dustin Tran. Image transformer. arXiv preprint arXiv:1802.05751, 2018. +[29] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. In Advances in neural information processing systems, pages 8026–8037, 2019. +[30] Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, and Piotr Dollár. Designing network design spaces. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10428–10436, 2020. +[31] Aravind Srinivas, Tsung-Yi Lin, Niki Parmar, Jonathon Shlens, Pieter Abbeel, and Ashish Vaswani. Bottleneck transformers for visual recognition. arXiv preprint arXiv:2101.11605, 2021. +[32] Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. The journal of machine learning research, 15(1):1929–1958, 2014. +[33] Zhiqing Sun, Shengcao Cao, Yiming Yang, and Kris Kitani. Rethinking transformer-based set prediction for object detection. arXiv preprint arXiv:2011.10881, 2020. +[34] Mingxing Tan and Quoc V Le. Efficientnet: Rethinking model scaling for convolutional neural networks. arXiv preprint arXiv:1905.11946, 2019. +[35] Ilya Tolstikhin, Neil Houlsby, Alexander Kolesnikov, Lucas Beyer, Xiaohua Zhai, Thomas Unterthiner, Jessica Yung, Daniel Keysers, Jakob Uszkoreit, Mario Lucic, et al. Mlp-mixer: An all-mlp architecture for vision. arXiv preprint arXiv:2105.01601, 2021. +[36] Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, and Hervé Jégou. Training data-efficient image transformers & distillation through attention. arXiv preprint arXiv:2012.12877, 2020. +[37] Hugo Touvron, Matthieu Cord, Alexandre Sablayrolles, Gabriel Synnaeve, and Hervé Jégou. Going deeper with image transformers. arXiv preprint arXiv:2103.17239, 2021. +[38] Hugo Touvron, Andrea Vedaldi, Matthijs Douze, and Hervé Jégou. Fixing the train-test resolution discrepancy. arXiv preprint arXiv:1906.06423, 2019. +[39] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30:5998–6008, 2017. +[40] Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and Ling Shao. Pyramid vision transformer: A versatile backbone for dense prediction without convolutions. arXiv preprint arXiv:2102.12122, 2021. +[41] Yuqing Wang, Zhaoliang Xu, Xinlong Wang, Chunhua Shen, Baoshan Cheng, Hao Shen, and Huaxia Xia. End-to-end video instance segmentation with transformers. arXiv preprint arXiv:2011.14503, 2020. +[42] Ross Wightman. Pytorch image models. https://github.com/rwightman/ pytorch-image-models, 2019. +[43] Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, and Lei Zhang. Cvt: Introducing convolutions to vision transformers. arXiv preprint arXiv:2103.15808, 2021. +[44] Tete Xiao, Yingcheng Liu, Bolei Zhou, Yuning Jiang, and Jian Sun. Unified perceptual parsing for scene understanding. In Proceedings of the European Conference on Computer Vision (ECCV), pages 418–434, 2018. +[45] Fuzhi Yang, Huan Yang, Jianlong Fu, Hongtao Lu, and Baining Guo. Learning texture transformer network for image super-resolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5791–5800, 2020. +[46] Li Yuan, Yunpeng Chen, Tao Wang, Weihao Yu, Yujun Shi, Francis EH Tay, Jiashi Feng, and Shuicheng Yan. Tokens-to-token vit: Training vision transformers from scratch on imagenet. arXiv preprint arXiv:2101.11986, 2021. +[47] Li Yuan, Francis EH Tay, Guilin Li, Tao Wang, and Jiashi Feng. Revisiting knowledge distillation via label smoothing regularization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3903–3911, 2020. +[48] Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, and Youngjoon Yoo. Cutmix: Regularization strategy to train strong classifiers with localizable features. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 6023–6032, 2019. +[49] Sangdoo Yun, Seong Joon Oh, Byeongho Heo, Dongyoon Han, Junsuk Choe, and Sanghyuk Chun. Re-labeling imagenet: from single to multi-labels, from global to localized labels. arXiv preprint arXiv:2101.05022, 2021. +[50] Yanhong Zeng, Jianlong Fu, and Hongyang Chao. Learning joint spatial-temporal transformations for video inpainting. In European Conference on Computer Vision, pages 528–543. Springer, 2020. +[51] Hang Zhang, Chongruo Wu, Zhongyue Zhang, Yi Zhu, Zhi Zhang, Haibin Lin, Yue Sun, Tong He, Jonas Muller, R. Manmatha, Mu Li, and Alexander Smola. Resnest: Split-attention networks. arXiv preprint arXiv:2004.08955, 2020. +[52] Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and David Lopez-Paz. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412, 2017. +[53] Hengshuang Zhao, Li Jiang, Jiaya Jia, Philip Torr, and Vladlen Koltun. Point transformer. arXiv preprint arXiv:2012.09164, 2020. +[54] Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, and Jiaya Jia. Pyramid scene parsing network. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2881–2890, 2017. +[55] Minghang Zheng, Peng Gao, Xiaogang Wang, Hongsheng Li, and Hao Dong. End-to-end object detection with adaptive clustering transformer. arXiv preprint arXiv:2011.09315, 2020. +[56] Sixiao Zheng, Jiachen Lu, Hengshuang Zhao, Xiatian Zhu, Zekun Luo, Yabiao Wang, Yanwei Fu, Jianfeng Feng, Tao Xiang, Philip HS Torr, et al. Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers. arXiv preprint arXiv:2012.15840, 2020. +[57] Zhun Zhong, Liang Zheng, Guoliang Kang, Shaozi Li, and Yi Yang. Random erasing data augmentation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 13001–13008, 2020. +[58] Bolei Zhou, Hang Zhao, Xavier Puig, Tete Xiao, Sanja Fidler, Adela Barriuso, and Antonio Torralba. Semantic understanding of scenes through the ade20k dataset. International Journal of Computer Vision, 127(3):302–321, 2019. +[59] Daquan Zhou, Bingyi Kang, Xiaojie Jin, Linjie Yang, Xiaochen Lian, Qibin Hou, and Jiashi Feng. Deepvit: Towards deeper vision transformer. arXiv preprint arXiv:2103.11886, 2021. +[60] Luowei Zhou, Yingbo Zhou, Jason J Corso, Richard Socher, and Caiming Xiong. End-to-end dense video captioning with masked transformer. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 8739–8748, 2018. +[61] Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, and Jifeng Dai. Deformable detr: Deformable transformers for end-to-end object detection. arXiv preprint arXiv:2010.04159, 2020. \ No newline at end of file diff --git a/parse/train/2vubO341F_E/2vubO341F_E_content_list.json b/parse/train/2vubO341F_E/2vubO341F_E_content_list.json new file mode 100644 index 0000000000000000000000000000000000000000..a643f852198f407876d9c8963c55d9c148ea8d89 --- /dev/null +++ b/parse/train/2vubO341F_E/2vubO341F_E_content_list.json @@ -0,0 +1,1035 @@ +[ + { + "type": "text", + "text": "All Tokens Matter: Token Labeling for Training Better Vision Transformers ", + "text_level": 1, + "bbox": [ + 209, + 122, + 790, + 172 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Zihang Jiang1∗ Qibin Hou2,1† Li Yuan3 Daquan Zhou1 Yujun Shi1 ", + "bbox": [ + 199, + 219, + 794, + 237 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Xiaojie Jin4 Anran Wang4 Jiashi Feng4 ", + "bbox": [ + 264, + 256, + 733, + 272 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "1National University of Singapore 2Nankai University ", + "bbox": [ + 302, + 292, + 694, + 308 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "3 Peking University 4ByteDance ", + "text_level": 1, + "bbox": [ + 379, + 328, + 619, + 343 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "{jzh0103,andrewhoux,ylustcnus,zhoudaquan21,shiyujun1016}@gmail.com xjjin0731@gmail.com, anran.wang@bytedance.com, jshfeng@gmail.com ", + "bbox": [ + 209, + 344, + 789, + 371 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Abstract ", + "text_level": 1, + "bbox": [ + 462, + 406, + 535, + 422 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "In this paper, we present token labeling—a new training objective for training high-performance vision transformers (ViTs). Different from the standard training objective of ViTs that computes the classification loss on an additional trainable class token, our proposed one takes advantage of all the image patch tokens to compute the training loss in a dense manner. Specifically, token labeling reformulates the image classification problem into multiple token-level recognition problems and assigns each patch token with an individual location-specific supervision generated by a machine annotator. Experiments show that token labeling can clearly and consistently improve the performance of various ViT models across a wide spectrum. For a vision transformer with 26M learnable parameters serving as an example, with token labeling, the model can achieve $8 4 . 4 \\%$ Top-1 accuracy on ImageNet. The result can be further increased to $8 6 . 4 \\%$ by slightly scaling the model size up to 150M, delivering the minimal-sized model among previous models $( 2 5 0 \\mathbf { M } + )$ reaching $86 \\%$ . We also show that token labeling can clearly improve the generalization capability of the pretrained models on downstream tasks with dense prediction, such as semantic segmentation. Our code and model are publicly available at https://github.com/zihangJiang/TokenLabeling. ", + "bbox": [ + 232, + 438, + 766, + 674 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "1 Introduction ", + "text_level": 1, + "bbox": [ + 174, + 699, + 310, + 717 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Transformers [39] have achieved great performance for almost all the natural language processing (NLP) tasks over the past years [4, 14, 24]. Motivated by such success, recently, many researchers attempt to build transformer models for vision tasks, and their encouraging results have shown the great potential of transformer based models for image classification [6, 15, 25, 36, 40, 46], especially the strong benefits of the self-attention mechanism in building long-range dependencies between pairs of input tokens. ", + "bbox": [ + 174, + 731, + 825, + 814 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Despite the importance of gathering long-range dependencies, recent work on local data augmentation [57] has demonstrated that well modeling and leveraging local information for image classification would avoid biasing the model towards skewed and non-generalizable patterns and substantially improve the model performance. However, recent vision transformers normally utilize class tokens that aggregate global information to predict the output class while neglecting the role of other patch tokens that encode rich information on their respective local image patches. ", + "bbox": [ + 176, + 820, + 825, + 863 + ], + "page_idx": 0 + }, + { + "type": "image", + "img_path": "images/9c7bb663ccabca3baefdad385c72c84db533d97dbd497ac18638700c9e1d548b.jpg", + "image_caption": [ + "Figure 1: Comparison between the proposed LV-ViT and other recent works based on vision transformers, including T2T-ViT [46], ConViT [12], BoTNet [31], DeepViT [59], DeiT [36], ViT [15], Swin Transformer [25], LambdaNet [1], CvT [43], CrossViT [6], PVT [40], CaiT [37]. Note that we only show models whose model sizes are under 100M. As can be seen, our LV-ViT achieves the best results using the least amount of learnable parameters. The default test resolution is $2 2 4 \\times 2 2 4$ unless specified after $@$ . " + ], + "image_footnote": [], + "bbox": [ + 174, + 87, + 825, + 265 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "", + "bbox": [ + 174, + 382, + 825, + 425 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "In this paper, we present a new training objective for vision transformers, termed token labeling, that takes advantage of both the patch tokens and the class tokens. Our method takes a $K$ -dimensional score map generated by a machine annotator as supervision to supervise all the tokens in a dense manner, where $K$ is the number of categories for the target dataset. In this way, each patch token is explicitly associated with an individual location-specific supervision indicating the existence of the target objects inside the corresponding image patch, so as to improve the object grounding and recognition capabilities of vision transformers with negligible computation overhead. To the best of our knowledge, this is the first work demonstrating that dense supervision is beneficial to vision transformers in image classification. ", + "bbox": [ + 174, + 431, + 825, + 556 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "According to our experiments, utilizing the proposed token labeling objective can clearly boost the performance of vision transformers. As shown in Figure 1, our model, named LV-ViT, with 56M parameters, yields $8 5 . 4 \\%$ top-1 accuracy on ImageNet [13], behaving better than all the other transformer-based models having no more than 100M parameters. When the model size is scaled up to 150M, the result can be further improved to $8 6 . 4 \\%$ . In addition, we have empirically found that the pretrained models with token labeling are also beneficial to downstream tasks with dense prediction, such as semantic segmentation. ", + "bbox": [ + 174, + 563, + 825, + 660 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "2 Related Work ", + "text_level": 1, + "bbox": [ + 174, + 679, + 321, + 696 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "Transformers [39] refer to the models that entirely rely on the self-attention mechanism to build global dependencies, which are originally designed for natural language processing tasks. Due to their strong capability of capturing spatial information, transformers have also been successfully applied to a variety of vision problems, including low-level vision tasks like image enhancement [7, 45], as well as more challenging tasks such as image classification [9, 15], object detection [5, 11, 55, 61], segmentation [7, 33, 41] and image generation [28]. Some works also extend transformers for video and 3D point cloud processing [50, 53, 60]. ", + "bbox": [ + 174, + 710, + 825, + 809 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "Vision Transformer (ViT) is one of the earlier attempts that achieved state-of-the-art performance on ImageNet classification, using pure transformers as basic building blocks. However, ViTs need pretraining on very large datasets, such as ImageNet-22k and JFT-300M, and huge computation resources to achieve comparable performance to ResNet [18] with a similar model size trained on ImageNet. Later, DeiT [36] manages to tackle the data-inefficiency problem by simply adjusting the network architecture and adding an additional token along with the class token for Knowledge Distillation [21, 47] to improve model performance. ", + "bbox": [ + 174, + 814, + 825, + 911 + ], + "page_idx": 1 + }, + { + "type": "image", + "img_path": "images/77890e011e8cc08c2d9f50310be754cb08f7894ee6cb766ac590a7580f42e616.jpg", + "image_caption": [ + "Figure 2: Pipeline of training vision transformers with token labeling. Other than utilizing the class token (pink rectangle), we also take advantage of all the output patch tokens (orange rounded rectangle) by assigning each patch token an individual location-specific prediction generated by a machine annotator [3] as supervision (see the part in the red dash rectangle). Our proposed token labeling method can be treated as an auxiliary objective to provide each patch token the local details that aid vision transformers to more accurately locate and recognize the target objects. Note that the traditional vision transformer training does not include the red dash rectangle part. " + ], + "image_footnote": [], + "bbox": [ + 209, + 89, + 794, + 318 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "Some recent works [6, 16, 43, 46] also attempt to introduce the local dependency into vision transformers by modifying the patch embedding block or the transformer block or both, leading to significant performance gains. Moreover, there are also some works [20, 25, 40] adopting a pyramid structure to reduce the overall computation while maintaining the model’s ability to capture low-level features. ", + "bbox": [ + 174, + 452, + 825, + 522 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "Unlike most aforementioned works that design new transformer blocks or transformer architectures, we attempt to improve vision transformers by studying the role of patch tokens that embed rich local information inside image patches. We show that by slightly tuning the structure of vision transformers and employing the proposed token labeling objective, we can achieve strong baselines for transformer models at different model size levels. ", + "bbox": [ + 174, + 529, + 825, + 598 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "3 Token Labeling Method ", + "text_level": 1, + "bbox": [ + 176, + 619, + 406, + 637 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "In this section, we first briefly review the structure of the vision transformer [15] and then describe the proposed training objective—token labeling. ", + "bbox": [ + 174, + 652, + 823, + 681 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "3.1 Revisiting Vision Transformer ", + "text_level": 1, + "bbox": [ + 176, + 699, + 421, + 714 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "A typical vision transformer [15] first decomposes a fixed-size input image into a sequence of small patches. Each small patch is mapped to a feature vector, or called a token, by projection with a linear layer. Then, all the tokens combined with an additional learnable class token for classification score prediction are sent into a stack of transformer blocks for feature encoding. ", + "bbox": [ + 173, + 726, + 825, + 781 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "In loss computing, the class token from the output tokens of the last transformer block is usually selected and sent into a linear layer for the classification score prediction. Mathematically, given an image $I$ , denote the output of the last transformer block as $[ \\bar { X } ^ { c l s } , X ^ { 1 } , . . . , X ^ { N } ]$ , where $N$ is the total number of patch tokens, and $X ^ { c l s }$ and $X ^ { 1 } , . . . , X ^ { N }$ correspond to the class token and the patch tokens, respectively. The classification loss for image $I$ can be written as ", + "bbox": [ + 174, + 787, + 823, + 859 + ], + "page_idx": 2 + }, + { + "type": "equation", + "img_path": "images/592dd559a9882895a0bc985fe4c80cb9b1f3f3afab9946d0fd9a87b3424793a9.jpg", + "text": "$$\nL _ { c l s } = H ( { X } ^ { c l s } , { y } ^ { c l s } ) ,\n$$", + "text_format": "latex", + "bbox": [ + 424, + 868, + 573, + 887 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "where $H ( \\cdot , \\cdot )$ is the softmax cross-entropy loss and $y ^ { c l s }$ is the class label. ", + "bbox": [ + 173, + 896, + 651, + 912 + ], + "page_idx": 2 + }, + { + "type": "image", + "img_path": "images/c741b481df54f67e411d3cc1b30dc7deb904c80d9d11e57036dbf5b27cdbfb8f.jpg", + "image_caption": [ + "Figure 3: Comparison between CutMix [48] (Left) and our proposed MixToken (Right). CutMix is operated on the input images. This results in patches containing mixed regions from the two images (see the patches enclosed by red bounding boxes). Differently, MixToken targets at mixing tokens after patch embedding. This enables each token after patch embedding to have clean content as shown in the right part of this figure. The detailed advantage of MixToken can be found in Sec. 4.2. " + ], + "image_footnote": [], + "bbox": [ + 197, + 90, + 490, + 242 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "3.2 Token Labeling ", + "text_level": 1, + "bbox": [ + 174, + 275, + 321, + 291 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "The above classification problem only adopts an image-level label as supervision whereas it neglects the rich information embedded in each image patch. In this subsection, we present a new training objective—token labeling—that takes advantage of the complementary information between the patch tokens and the class tokens. ", + "bbox": [ + 174, + 304, + 825, + 359 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "Token Labeling: Different from the classification loss as formulated in Eqn. (1) that measures the distance between the single class token (representing the whole input image) and the corresponding image-level label, token labeling emphasizes the importance of all output tokens and advocates that each output token should be associated with an individual location-specific label. Therefore, in our method, the ground truth for an input image involves not only a single $K$ -dimensional vector $y ^ { c l s }$ but also a $K \\times N$ matrix or called a $K$ -dimensional score map as represented by $[ y ^ { 1 } , . . . , y ^ { N } ]$ , where $N$ is the number of the output patch tokens. ", + "bbox": [ + 173, + 366, + 825, + 464 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "Specifically, we leverage a dense score map for each training image and use the cross-entropy loss between each output patch token and the corresponding aligned label in the dense score map as an auxiliary loss at the training phase. Figure 2 provides an intuitive interpretation. Given the output patch tokens $X ^ { 1 } , . . . , X ^ { N }$ and the corresponding labels $[ y ^ { 1 } , . . . , y ^ { N } ]$ , the token labeling objective can be defined as ", + "bbox": [ + 174, + 469, + 825, + 540 + ], + "page_idx": 3 + }, + { + "type": "equation", + "img_path": "images/a9d480948e81c19661ff7ef1b5daada45648302552f0f35ccc9e1a4855e07862.jpg", + "text": "$$\nL _ { t l } = \\frac { 1 } { N } \\sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) .\n$$", + "text_format": "latex", + "bbox": [ + 415, + 546, + 583, + 590 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "Recall that $H$ is the cross-entropy loss. Therefore, the total loss function can be written as ", + "bbox": [ + 171, + 603, + 764, + 619 + ], + "page_idx": 3 + }, + { + "type": "equation", + "img_path": "images/c5da8361fd2faf0126b83503cc2cc4f9798363630c4f410a80eef81deafa2915.jpg", + "text": "$$\n\\begin{array} { l } { { { \\cal L } _ { t o t a l } = H ( X ^ { c l s } , y ^ { c l s } ) + \\beta \\cdot L _ { t l } , } } \\\\ { { { } } } \\\\ { { = H ( X ^ { c l s } , y ^ { c l s } ) + \\beta \\cdot \\displaystyle \\frac { 1 } { N } \\sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) , } } \\end{array}\n$$", + "text_format": "latex", + "bbox": [ + 338, + 632, + 658, + 698 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "where $\\beta$ is a hyper-parameter to balance the two terms. In our experiment, we empirically set it to 0.5. ", + "bbox": [ + 174, + 710, + 825, + 738 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "Advantages: Our token labeling offers the following advantages. First of all, unlike knowledge distillation methods that require a teacher model to generate supervision labels online, token labeling is a cheap operation. The dense score map can be generated by a pretrained model in advance (e.g., EfficientNet [34] or NFNet [3]). During training, we only need to crop the score map and perform interpolation to make it aligned with the cropped image in the spatial coordinate. Thus, the additional computations are negligible. Second, rather than utilizing a single label vector as supervision as done in most classification models and the ReLabel strategy [49], we also harness score maps to supervise the models in a dense manner and thereby the label for each patch token provides location-specific information, which can aid the training models to easily discover the target objects and improve the recognition accuracy. Last but not the least, as dense supervision is adopted in training, we found that the pretrained models with token labeling benefit downstream tasks with dense prediction, like semantic segmentation. ", + "bbox": [ + 173, + 744, + 825, + 911 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "3.3 Token Labeling with MixToken ", + "text_level": 1, + "bbox": [ + 176, + 90, + 431, + 106 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "While training vision transformer, previous studies [36, 46] have shown that augmentation methods, like MixUp [52] and CutMix [48], can effectively boost the performance and robustness of the models. However, vision transformers rely on patch-based tokenization to map each input image to a sequence of tokens and our token labeling strategy also operates on patch-based token labels. If we apply CutMix directly on the raw image, some of the resulting patches may contain content from two images, leading to mixed regions within a small patch as shown in Figure 3. When performing token labeling, it is difficult to assign each output token a clean and correct label. Taking this situation into account, we rethink the CutMix augmentation method and present MixToken, which can be viewed as a modified version of CutMix operating on the tokens after patch embedding as illustrated in the right part of Figure 3. ", + "bbox": [ + 173, + 117, + 825, + 257 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "To be specific, for two images denoted as $I _ { 1 } , I _ { 2 }$ and their corresponding token labels $Y _ { 1 } = [ y _ { 1 } ^ { 1 } , . . . , y _ { 1 } ^ { N } ]$ as well as $Y _ { 2 } = [ y _ { 2 } ^ { 1 } , . . . , y _ { 2 } ^ { N } ]$ , we first feed the two images into the patch embedding module to tokenize each as a sequence of tokens, resulting in $T _ { 1 } = [ t _ { 1 } ^ { 1 } , . . . , t _ { 1 } ^ { N } ]$ and $\\bar { T _ { 2 } } = [ t _ { 2 } ^ { 1 } , . . . , t _ { 2 } ^ { N } ]$ . Then, we produce a new sequence of tokens by applying MixToken using a binary mask $M$ as follows: ", + "bbox": [ + 174, + 262, + 825, + 319 + ], + "page_idx": 4 + }, + { + "type": "equation", + "img_path": "images/cdf5e3aab4339f76c0c2c44221d0f16054703d72f085272d761449ddf6e388dc.jpg", + "text": "$$\n\\hat { T } = T _ { 1 } \\odot M + T _ { 2 } \\odot ( 1 - M ) ,\n$$", + "text_format": "latex", + "bbox": [ + 392, + 329, + 604, + 349 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "where $\\odot$ is element-wise multiplication. We use the same way to generate the mask $M$ as in [48]. For the corresponding token labels, we also mix them using the same mask $M$ : ", + "bbox": [ + 173, + 358, + 823, + 387 + ], + "page_idx": 4 + }, + { + "type": "equation", + "img_path": "images/fa81133b791d6c30dfc1c80f85c7a073a5241a9fada2c61ef2a3b397042bb25e.jpg", + "text": "$$\n\\hat { Y } = Y _ { 1 } \\odot M + Y _ { 2 } \\odot ( 1 - M ) .\n$$", + "text_format": "latex", + "bbox": [ + 392, + 397, + 606, + 417 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "The label for the class token can be written as ", + "bbox": [ + 176, + 428, + 473, + 441 + ], + "page_idx": 4 + }, + { + "type": "equation", + "img_path": "images/1bb8081440df26328f22eeb83ce87e3030fb4e4f726532138a58c694ac3a6d6a.jpg", + "text": "$$\ny ^ { \\hat { c } l s } = \\bar { M } y _ { 1 } ^ { c l s } + ( 1 - \\bar { M } ) y _ { 2 } ^ { c l s } ,\n$$", + "text_format": "latex", + "bbox": [ + 397, + 449, + 599, + 469 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "where $\\bar { M }$ is the average of all element values of $M$ . ", + "bbox": [ + 173, + 479, + 513, + 494 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "4 Experiments ", + "text_level": 1, + "bbox": [ + 174, + 517, + 312, + 535 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "4.1 Experiment Setup ", + "text_level": 1, + "bbox": [ + 174, + 551, + 339, + 566 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "We evaluate our method on the ImageNet [13] dataset. All experiments are built and conducted upon PyTorch [29] and the timm [42] library. We follow the standard training schedule and train our models on the ImageNet dataset for 300 epochs. Besides normal augmentations like CutOut [57] and RandAug [10], we also explore the effect of applying MixUp [52] and CutMix [48] together with our proposed token labeling. Empirically, we have found that using MixUp together with token labeling brings no benefit to the performance, and thus we do not apply it in our experiments. ", + "bbox": [ + 173, + 578, + 825, + 662 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "For optimization, by default, we use the AdamW optimizer [27] with a linear learning rate scaling strategy $\\begin{array} { r } { l r = 1 0 ^ { - 3 } \\times \\frac { b a t c h \\_ s i z e } { 6 4 0 } } \\end{array}$ and $5 \\times 1 0 ^ { - 2 }$ weight decay rate. For Dropout regularization, we observe that for small models, using Dropout hurts the performance. This has also been observed in a few other works related to training vision transformers [36, 37, 46]. As a result, we do not apply Dropout [32] and use Stochastic Depth [23] instead. More details on hyper-parameters and finetuning can be found in our supplementary materials. ", + "bbox": [ + 174, + 667, + 825, + 752 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "We use the NFNet-F6 [3] trained on ImageNet with an $8 6 . 3 \\%$ Top-1 accuracy as the machine annotator to generate dense score maps for the ImageNet dataset, yielding a 1000-dimensional score map for each image for training. The score map generation procedure is similar to [49], but we limit our experiment setting by training all models from scratch on ImageNet without extra data support, such as JFT-300M and ImageNet-22K. This is different from the original ReLabel paper [49], in which the EfficientNet-L2 model pretrained on JFT-300M is used. The input resolution for NFNet-F6 is $5 7 6 \\times 5 7 6$ , and the dimension of the corresponding output score map for each image is $L \\in \\mathbb { R } ^ { 1 8 \\times 1 8 \\times 1 0 0 0 }$ . During training, the target labels for the tokens are generated by applying RoIAlign [17] on the corresponding score map. In practice, we only store the top-5 score maps for each position in half-precision to save space as storing the entire score maps for all the images results in 2TB storage. In our experiment, we only need 10GB of storage to store all the score maps. ", + "bbox": [ + 173, + 758, + 825, + 911 + ], + "page_idx": 4 + }, + { + "type": "table", + "img_path": "images/8ab634b915f17e8f746292568270ed165dd5bacd94d1fed7927449e1328c4232.jpg", + "table_caption": [ + "Table 1: Performance of the proposed LV-ViT with different model sizes. Here, ‘depth’ denotes the number of transformer blocks used in different models. By default, the test resolution is set to $2 2 4 \\times 2 2 4$ except the last one which is $2 8 8 \\times 2 8 8$ . " + ], + "table_footnote": [], + "table_body": "
NameDepthEmbed dim.MLP Ratio#Heads#ParamsThroughput (im/s)Test sizeTop-1 Acc. (%)
LV-ViT-T122403.048.5M2032.622479.1
LV-ViT-S163843.0626M1018.222483.3
LV-ViT-M205123.0856M668.922484.1
LV-ViT-L247683.012150M204.828885.3
", + "bbox": [ + 173, + 138, + 825, + 220 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "4.2 Ablation Analysis ", + "text_level": 1, + "bbox": [ + 174, + 244, + 336, + 260 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "Model Settings: The default settings of the proposed LV-ViT are given in Table 1, where both token labeling and MixToken are used. A slight architecture modification to ViT [15] is that we replace the patch embedding module with a 4-layer convolution to better tokenize the input image and integrate local information. Detailed ablation about patch embedding can be found in our supplementary materials. As can be seen, our LV-ViT-T with only $8 . 5 { \\bf M }$ parameters can already achieve a top-1 accuracy of $7 9 . 1 \\%$ on ImageNet. Increasing the embedding dimension and network depth can further boost the performance. More experiments compared to other methods can be found in Sec. 4.3. In the following ablation experiments, we will set our LV-ViT-S as baseline and show the advantages of the proposed token labeling and MixToken methods. ", + "bbox": [ + 173, + 270, + 825, + 395 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "MixToken: We use MixToken as a substitution for CutMix while applying token labeling. Our experiments show that MixToken performs better than CutMix for token-based transformer models. As shown in Table 2, when training with the original ImageNet labels, using MixToken is $0 . 1 \\%$ higher than using CutMix. When using the ReLabel supervision, we can also see an advantage of $0 . { \\bar { 2 } } \\%$ over the CutMix baseline. Combining with our token labeling, the performance can be further raised to $8 3 . 3 \\%$ . ", + "bbox": [ + 173, + 401, + 825, + 484 + ], + "page_idx": 5 + }, + { + "type": "table", + "img_path": "images/ecc151313b3142fbf655b4065fd6e4b0c147ea2803b47139c3e90b472d0b2137.jpg", + "table_caption": [ + "Table 2: Ablation on the proposed MixToken and token labeling augmentations. We also show results with either the ImageNet hard label and the ReLabel [49] as supervision. " + ], + "table_footnote": [], + "table_body": "
Aug. MethodSupervisionTop-1 Acc.
MixTokenToken labeling83.3
MixTokenReLabel83.0
CutMixReLabel82.8
MixtokenImageNet Label82.5
CutMixImageNet Label82.4
", + "bbox": [ + 173, + 560, + 472, + 654 + ], + "page_idx": 5 + }, + { + "type": "table", + "img_path": "images/51895f2ce4fa7182e488e57176934e6ab370cc3160e81c0730e961fdd370158d.jpg", + "table_caption": [ + "Table 3: Ablation on different widely-used data augmentations. We have empirically found our proposed MixToken performs even better than the combination of MixUp and CutMix in vision transformers. " + ], + "table_footnote": [], + "table_body": "
MixTokenMixUpCutOutRandAugTop-1 Acc.
83.3
81.3
>x>xx>>>>>>>83.1
X√83.0
广XXX82.8
", + "bbox": [ + 483, + 559, + 825, + 654 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "Data Augmentation: Here, we study the compatibility of MixToken with other augmentation techniques, such as MixUp [52], CutOut [57] and RandAug [10]. The ablation results are shown in Table 3. We can see when all the four augmentation methods are used, a top-1 accuracy of $8 3 . 1 \\%$ is achieved. Interestingly, when the MixUp augmentation is removed, the performance can be improved to $8 3 . 3 \\%$ . This may be explained as, using MixToken and MixUp at the same time would bring too much noise in the label, and consequently cause confusion of the model. Moreover, the CutOut augmentation, which randomly erases some parts of the image, is also effective and removing it brings a performance drop of $\\dot { 0 } . 3 \\%$ . Similarly, the RandAug augmentation also contributes to the performance and using it brings an improvement of $0 . 5 \\%$ . ", + "bbox": [ + 173, + 669, + 825, + 795 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "All Tokens Matter: To show the importance of involving all tokens in our token labeling method, we attempt to randomly drop some tokens and use the remaining ones for computing the token labeling loss. The percentage of the remaining tokens is denoted as Token Participation Rate. As shown in Figure 4 (Left), we conduct experiments on two models: LV-ViT-S and LV-ViT-M. As can be seen, using only $2 0 \\%$ of the tokens to compute the token labeling loss decreases the performance $( - 0 . 5 \\%$ for LV-ViT-S and $- 0 . 4 \\%$ for LV-ViT-M). Involving more tokens for loss computation consistently leads to better performance. Since involving all tokens brings negligible computation cost and gives the best performance, we always set the token participation rate as $\\mathrm { \\bar { 1 0 0 \\% } }$ in the following experiments. ", + "bbox": [ + 173, + 800, + 825, + 911 + ], + "page_idx": 5 + }, + { + "type": "image", + "img_path": "images/810578c2e36580b46aa98e4ec01030b20f6e00994c33dafcaa4cde435beb4d46.jpg", + "image_caption": [ + "Figure 4: Left: LV-ViT ImageNet Top-1 Accuracy w.r.t. the token participation rate while applying token labeling. Token participation rate indicates the percentage of patch tokens involved in computing the token labeling loss. This experiment reflects that all tokens matter for vision transformers. Right: LV-ViT-S ImageNet Top-1 Accuracy w.r.t. different annotator models. The point size indicates the parameter number of the annotator model. Clearly, our token labeling objective is robust to different annotator models. " + ], + "image_footnote": [], + "bbox": [ + 178, + 103, + 790, + 275 + ], + "page_idx": 6 + }, + { + "type": "table", + "img_path": "images/ae9fdf37e5d602133531450f1b69751afa30b0104c5bff5b6433f7a76be69066.jpg", + "table_caption": [ + "Table 4: Comparison of token labeling (TL), knowledge distillation (KD) based method and ReLabel method based on utilized tokens, DeiT-S/LV-ViT-S Top-1 accuracy on ImageNet validation set and training time on a single V100 GPU node. " + ], + "table_footnote": [], + "table_body": "
MethodOnline KDOnline TLTLReLabelVanilla
Tokens Utilized2AllAll11
DeiT-S Acc. (%)81.281.881.080.479.9
LV-ViT-S Acc. (%)83.083.583.382.882.4
Training Time (8× V100)63 hrs63 hrs45 hrs45 hrs41 hrs
", + "bbox": [ + 209, + 431, + 789, + 518 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Online Token Labeling: Unlike the online knowledge distillation method which generates labels by a teacher model online, our token labeling approach utilizes the dense label map generated in advance and directly applies the corresponding augmentation methods, such as random crop, on the label map to obtain token-level labels. To directly compare with the online knowledge distillation based method and validate the effectiveness of token-level supervision, we further conduct experiments on the online version of our token labeling method, which generates token-level labels online during training. Following DeiT [36], we use RegNetY-16GF [30] as the online teacher model. Results in terms of DeiT-S/LV-ViT-S Top-1 accuracy and training time for our token labeling, online knowledge distillation, and ReLabel [49] are listed in Table 4, with number of utilized tokens also included for clear comparison. As can be seen, for both online and offline cases, using token-level supervision can improve the overall performance with only negligible additional training cost. Meanwhile, compared to the vanilla training baseline, our proposed offline token labeling brings almost no additional training cost, and boosts the overall performance of LV-ViT-S by $0 . 9 \\%$ , which well demonstrates its efficiency and effectiveness. ", + "bbox": [ + 173, + 545, + 825, + 738 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Robustness to Different Annotators: To evaluate the robustness of our token labeling method, we use different pretrained CNNs, including EfficientNet-B3,B4,B5,B6,B7,B8 [34], NFNet-F6 [3] and ResNest269E [51], as annotator models to provide dense supervision. Results are shown in the right part of Figure 4. We can see that, even if we use an annotator with relatively lower performance, such as EfficientNet-B3 whose Top-1 accuracy is $8 1 . 6 \\%$ , it can still provide multi-label location-specific supervision and help improve the performance of our LV-ViT-S model. Meanwhile, annotator models with better performance can provide more accurate supervision, bringing even better performance, as stronger annotator models can generate better token-level labels. The largest annotator NFNet-F6 [3], which has the best performance of $8 6 . 3 \\%$ , allows us to achieve the best result for LV-ViT-S, which is $8 3 . 3 \\%$ . In addition, we also attempt to use a better model, EfficientNet-L2 pretrained on JFT-300M as described in [49] which has $8 8 . 2 \\%$ Top-1 ImageNet accuracy, as our annotator. The performance of LV-ViT-S can be further improved to $8 \\mathrm { { 3 . 5 \\% } }$ . However, to fairly compare with the models without extra training data, we only report results based on dense supervision produced by NFNet-F6 [3] that uses only ImageNet training data. ", + "bbox": [ + 174, + 746, + 825, + 911 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "", + "bbox": [ + 174, + 92, + 825, + 119 + ], + "page_idx": 7 + }, + { + "type": "image", + "img_path": "images/a3d559defddf629a8a69e1ea7f71228daa68e20f028756c9b3671f019cf5ad40.jpg", + "image_caption": [ + "Figure 5: Performance of the proposed token labeling objective on three different vision transformers: DeiT [36] (Left), T2T-ViT [46] (Middle), and LV-ViT (Right). Our method has a consistent improvement on all 7 different ViT models. " + ], + "image_footnote": [], + "bbox": [ + 179, + 143, + 825, + 260 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "Robustness to Different ViT Variants: To further evaluate the robustness of our token labeling, we train different transformer-based networks, including DeiT [36], T2T-ViT [3] and our model LV-ViT, with the proposed training objective. Results are shown in Figure 5. It can be found that, all the models trained with token labeling consistently outperform their vanilla counterparts, demonstrating the robustness of token labeling with respect to different variants of patch-based vision transformers. Meanwhile, for different scales of the models, the improvement is also consistent. Interestingly, we observe larger improvements for larger models. These indicate that our proposed token labeling method is widely applicable to a large range of patch-based vision transformer variants. ", + "bbox": [ + 173, + 337, + 825, + 449 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "Beyond Vision Transformers: We further explore the performance of token labeling on other CNN-based and MLP-based models. Results are shown in Table 5. Besides our re-implementation with more data augmentation and regularization techniques, we also provide the results from the original papers. It can be found that for both MLP-based and CNN-based models, our token labeling objective can also improve the performance over strong baselines by providing location-specific dense supervision. ", + "bbox": [ + 173, + 454, + 825, + 537 + ], + "page_idx": 7 + }, + { + "type": "table", + "img_path": "images/163ebee6ecc8c4b165c6f3e5a941947668dd1d76a42075b0fcd4b5d7c14a71c5.jpg", + "table_caption": [ + "Table 5: Performance of the proposed token labeling objective on representative CNN-based (ResNeSt) and MLP-based (Mixer-MLP) models. Our method has a consistent improvement on all different models. Here † indicates results reported in original papers. " + ], + "table_footnote": [], + "table_body": "
ModelMixer-S/16 [35]Mixer-B/16 [35]Mixer-L/16 [35]ResNeSt-50 [51]
Token LabelingXX√×X√××√××√
Parameters18M18M18M59M59M 59M207M207M207M27M27M 27M
Top-1 Acc. (%)73.8t75.676.176.478.379.571.6t77.780.181.1t80.981.5
", + "bbox": [ + 181, + 613, + 815, + 688 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "4.3 Comparison to Other Methods ", + "text_level": 1, + "bbox": [ + 174, + 742, + 426, + 757 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "We compare our proposed model LV-ViT with other state-of-the-art methods in Table 6. For smallsized models, when the test resolution is set to $2 2 4 \\times 2 2 4$ , we achieve an $8 3 . 3 \\%$ accuracy on ImageNet with only 26M parameters, which is $3 . 4 \\%$ higher than the strong baseline DeiT-S [36]. For mediumsized models, when the test resolution is set to $3 8 4 \\times 3 8 4$ we achieve the performance of $8 5 . 4 \\%$ , the same as CaiT-S36 [37], but with much less computational cost and parameters. Note that both DeiT and CaiT use knowledge distillation to improve their models, which introduce much more computations in training. However, we do not require any extra computations in training and only have to compute and store the dense score maps in advance. For large-sized models, our LV-ViT-L with a test resolution of $4 4 8 \\times 4 4 8$ achieves an $8 6 . 2 \\%$ top-1 accuracy, which is comparable to CaiT-M36 [37] but with far fewer FLOPs and parameters. ", + "bbox": [ + 173, + 772, + 826, + 911 + ], + "page_idx": 7 + }, + { + "type": "table", + "img_path": "images/1ffe2c2d40127fa5107472db1c5e025901208024a3c26b0c26ca4a1ac6751875.jpg", + "table_caption": [ + "Table 6: Top-1 accuracy comparison with other methods on ImageNet [13] and ImageNet Real [2]. All models are trained without external data. With the same computation and parameter constraint, our model consistently outperforms other CNN-based and transformer-based counterparts. The results of CNNs and ViT are referenced from [37]. " + ], + "table_footnote": [], + "table_body": "
NetworkParamsFLOPsTrain sizeTest sizeTop-1(%)Real Top-1 (%)
EfficientNet-B5 [34] SNNO30M9.9B45645683.688.3
EfficientNet-B7 [34]66M37.0B60060084.3
Fix-EfficientNet-B8 [34,38]87M89.5B67280085.790.0
NFNet-F3 [3]255M114.8B32041685.789.4
NFNet-F4 [3]316M215.3B38451285.989.4
NFNet-F5[3]377M289.8B41654486.089.2
ViT-B/16 [15]86M55.4B22438477.983.6
ViT-L/16[15]307M190.7B22438476.582.2
T2T-ViT-14 [46]22M5.2B22422481.5
T2T-ViT-14↑384 [46]22M17.1B22438483.31
Cross ViT [6]45M56.6B22448084.1一
Swin-B[25]88M47.0B22438484.2
TNT-B[16]66M14.1B22422482.8
iriirrrrrs DeepViT-S [59]27M6.2B22422482.3
DeepViT-L [59]55M12.5B22422483.1
DeiT-S[36]22M4.6B22422479.985.7
Distilled DeiT-S [36]22M4.6B22422481.286.8
DeiT-B [36]86M17.5B22422481.886.7
DeiT-B↑384 [36]86M55.4B22438483.187.7
Distilled DeiT-B [36]87M17.5B22422483.488.3
BoTNet-S1-128 [31]79.1M19.3B25625684.2
BoTNet-S1-128↑384 [31]79.1M45.8B25638484.7-
CaiT-S36↑384 [37]68M48.0B22438485.4- 89.8
CaiT-M36[37]271M53.7B22422485.189.3
CaiT-M36↑448 [37]271M247.8B22444886.390.2
LV-ViT-S26M6.6B22422483.3
江 LV-ViT-S↑38426M22.2B22438484.488.1 88.9
LV-ViT-M56M16.0B22422484.188.4
W LV-ViT-M↑38456M42.2B22438485.489.5
LV-ViT-L150M59.0B28828885.389.3
0 LV-ViT-L↑448150M157.2B28844885.989.7
LV-ViT-L↑448150M157.2B44844886.289.9
LV-ViT-L↑512151M214.8B44851286.490.1
", + "bbox": [ + 174, + 150, + 825, + 604 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "4.4 Semantic Segmentation on ADE20K ", + "text_level": 1, + "bbox": [ + 174, + 635, + 464, + 650 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "It has been shown in [19] that different training techniques for pretrained models have different impacts on downstream tasks with dense prediction, like semantic segmentation. To demonstrate the advantage of the proposed token labeling objective on tasks with dense prediction, we apply our pretrained LV-ViT with token labeling to the semantic segmentation task. ", + "bbox": [ + 174, + 662, + 825, + 718 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "Similar to previous work [25], we run experiments on the widely-used ADE20K [58] dataset. ADE20K contains 25K images in total, including 20K images for training, 2K images for validation and 3K images for test, and covering 150 different foreground categories. We take both FCN [26] and UperNet [44] as our segmentation frameworks and use the mmseg toolbox to implement. During training, following [25], we use the AdamW optimizer with an initial learning rate of 6e-5 and a weight decay of 0.01. We also use a linear learning schedule with a minimum learning rate of 5e-6. All models are trained on 8 GPUs and with a batch size of 16 (i.e., 2 images on each GPU). The input resolution is set to $5 1 2 \\times 5 1 2$ . In inference, a multi-scale test with interpolation rates of [0.75, 1.0, 1.25, 1.5, 1.75] is used. As suggested by [58], we report results in terms of both mean intersection-over-union (mIoU) and the average pixel accuracy (Pixel Acc.). ", + "bbox": [ + 174, + 724, + 825, + 863 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "In Table 7, we test the performance of token labeling on both FCN and UperNet frameworks. The FCN framework has a light convolutional head and can directly reflect the performance of the pretrained models in terms of transferable capability. As can be seen, pretrained models with token labeling perform better than those without token labeling. This indicates token labeling is indeed beneficial to semantic segmentation. ", + "bbox": [ + 176, + 869, + 825, + 911 + ], + "page_idx": 8 + }, + { + "type": "table", + "img_path": "images/70c1d4e41ceddea09bd2be3af5115392c9cbb1586a4c7cf96581f2f713f425f0.jpg", + "table_caption": [ + "Table 7: Transfer performance of the proposed LV-ViT in semantic segmentation. We take two classic methods, FCN and UperNet, as segmentation architectures and show both single-scale (SS) and multi-scale (MS) results on the validation set. " + ], + "table_footnote": [], + "table_body": "
MethodToken LabelingModel SizemIoU (SS)P. Acc. (SS)mIoU (MS)P. Acc. (MS)
LV-ViT-S + FCN×30M46.181.947.382.6
LV-ViT-S + FCN30M47.282.448.483.0
LV-ViT-S + UperNet44M46.582.147.682.7
LV-ViT-S + UperNetX44M47.982.648.683.1
", + "bbox": [ + 178, + 138, + 820, + 228 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "", + "bbox": [ + 174, + 256, + 821, + 284 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "We also compare our segmentation results with previous state-of-the-art segmentation methods in Table 8. Without pretraining on large-scale datasets such as ImageNet-22K, our LV-ViT-M with the UperNet segmentation architecture achieves an mIoU score of 50.6 with only 77M parameters. This result is much better than the previous CNN-based and transformer-based models. Furthermore, using our LV-ViT-L as the pretrained model yields a better result of 51.8 in terms of mIoU. As far as we know, this is the best result reported on ADE20K with no pretraining on ImageNet-22K or other large-scale datasets. ", + "bbox": [ + 173, + 290, + 825, + 387 + ], + "page_idx": 9 + }, + { + "type": "table", + "img_path": "images/7675f5b12a93d9e094cea27ffe903eaa1865bea2f706e0842c5d4c04af07ac46.jpg", + "table_caption": [ + "Table 8: Comparison with previous work on ADE20K validation set. As far as we know, our LVViT-L $^ +$ UperNet achieves the best result on ADE20K with only ImageNet-1K as training data in pretraining. †Pretrained on ImageNet-22K. " + ], + "table_footnote": [], + "table_body": "
BackboneSegmentation ArchitectureModel SizemIoU (MS)Pixel Acc. (MS)
SNNOResNet-269PSPNet [54]44.981.7
ResNet-101UperNet [44]86M44.9-
ResNet-101Strip Pooling [22]45.682.1
ResNeSt200DeepLabV3+ [8]88M48.41
DeiT-SUperNet52M44.0-
TirriiirrssViT-LargetSETR [56]308M50.383.5
Swin-T[25]UperNet60M46.11
Swin-S [25]UperNet81M49.3=
Swin-B [25]UperNet121M49.7
Swin-B† [25]UperNet121M51.6-
LIA-ATLV-ViT-SFCN30M48.483.0
LV-ViT-SUperNet44M48.683.1
LV-ViT-MUperNet77M50.683.5
LV-ViT-LUperNet209M51.884.1
", + "bbox": [ + 173, + 454, + 825, + 676 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "5 Conclusions and Discussion ", + "text_level": 1, + "bbox": [ + 174, + 708, + 436, + 727 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "In this paper, we introduce a new token labeling method to help improve the performance of vision transformers. We also analyze the effectiveness and robustness of our token labeling with respect to different annotators and different variants of patch-based vision transformers. By applying token labeling, our proposed LV-ViT achieves $8 4 . 4 \\%$ Top-1 accuracy with only 26M parameters and $8 6 . 4 \\%$ Top-1 accuracy with 150M parameters on ImageNet-1K benchmark. ", + "bbox": [ + 174, + 741, + 825, + 811 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Despite the effectiveness, token labeling has a limitation of requiring a pretrained model as the machine annotator. Fortunately, the machine annotating procedure can be done in advance to avoid introducing extra computational cost in training. This makes our method quite different from knowledge distillation methods that rely on online teaching. For users with limited machine resources on hand, our token labeling provides a promising training technique to improve the performance of vision transformers. ", + "bbox": [ + 174, + 818, + 825, + 901 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "References \n[1] Irwan Bello. Lambdanetworks: Modeling long-range interactions without attention. arXiv preprint arXiv:2102.08602, 2021. \n[2] Lucas Beyer, Olivier J Hénaff, Alexander Kolesnikov, Xiaohua Zhai, and Aäron van den Oord. Are we done with imagenet? arXiv preprint arXiv:2006.07159, 2020. \n[3] Andrew Brock, Soham De, Samuel L Smith, and Karen Simonyan. High-performance large-scale image recognition without normalization. arXiv preprint arXiv:2102.06171, 2021. \n[4] Tom B Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. arXiv preprint arXiv:2005.14165, 2020. \n[5] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-end object detection with transformers. arXiv preprint arXiv:2005.12872, 2020. \n[6] Chun-Fu Chen, Quanfu Fan, and Rameswar Panda. Crossvit: Cross-attention multi-scale vision transformer for image classification. arXiv preprint arXiv:2103.14899, 2021. \n[7] Hanting Chen, Yunhe Wang, Tianyu Guo, Chang Xu, Yiping Deng, Zhenhua Liu, Siwei Ma, Chunjing Xu, Chao Xu, and Wen Gao. Pre-trained image processing transformer. arXiv preprint arXiv:2012.00364, 2020. \n[8] Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. Encoder-decoder with atrous separable convolution for semantic image segmentation. In Proceedings of the European conference on computer vision (ECCV), pages 801–818, 2018. \n[9] Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, and Ilya Sutskever. Generative pretraining from pixels. In International Conference on Machine Learning, pages 1691–1703. PMLR, 2020. \n[10] Ekin D Cubuk, Barret Zoph, Jonathon Shlens, and Quoc V Le. Randaugment: Practical automated data augmentation with a reduced search space. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pages 702–703, 2020. \n[11] Zhigang Dai, Bolun Cai, Yugeng Lin, and Junying Chen. Up-detr: Unsupervised pre-training for object detection with transformers. arXiv preprint arXiv:2011.09094, 2020. \n[12] Stéphane d’Ascoli, Hugo Touvron, Matthew Leavitt, Ari Morcos, Giulio Biroli, and Levent Sagun. Convit: Improving vision transformers with soft convolutional inductive biases. arXiv preprint arXiv:2103.10697, 2021. \n[13] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. \n[14] Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018. \n[15] Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. \n[16] Kai Han, An Xiao, Enhua Wu, Jianyuan Guo, Chunjing Xu, and Yunhe Wang. Transformer in transformer. arXiv preprint arXiv:2103.00112, 2021. \n[17] Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. Mask r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 2961–2969, 2017. \n[18] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. \n[19] Tong He, Zhi Zhang, Hang Zhang, Zhongyue Zhang, Junyuan Xie, and Mu Li. Bag of tricks for image classification with convolutional neural networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 558–567, 2019. \n[20] Byeongho Heo, Sangdoo Yun, Dongyoon Han, Sanghyuk Chun, Junsuk Choe, and Seong Joon Oh. Rethinking spatial dimensions of vision transformers. arXiv preprint arXiv:2103.16302, 2021. \n[21] Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531, 2015. \n[22] Qibin Hou, Li Zhang, Ming-Ming Cheng, and Jiashi Feng. Strip pooling: Rethinking spatial pooling for scene parsing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4003–4012, 2020. \n[23] Gao Huang, Yu Sun, Zhuang Liu, Daniel Sedra, and Kilian Q Weinberger. Deep networks with stochastic depth. In European conference on computer vision, pages 646–661. Springer, 2016. \n[24] Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692, 2019. \n[25] Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. arXiv preprint arXiv:2103.14030, 2021. \n[26] Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3431–3440, 2015. \n[27] Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. \n[28] Niki Parmar, Ashish Vaswani, Jakob Uszkoreit, Łukasz Kaiser, Noam Shazeer, Alexander Ku, and Dustin Tran. Image transformer. arXiv preprint arXiv:1802.05751, 2018. \n[29] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. In Advances in neural information processing systems, pages 8026–8037, 2019. \n[30] Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, and Piotr Dollár. Designing network design spaces. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10428–10436, 2020. \n[31] Aravind Srinivas, Tsung-Yi Lin, Niki Parmar, Jonathon Shlens, Pieter Abbeel, and Ashish Vaswani. Bottleneck transformers for visual recognition. arXiv preprint arXiv:2101.11605, 2021. \n[32] Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. The journal of machine learning research, 15(1):1929–1958, 2014. \n[33] Zhiqing Sun, Shengcao Cao, Yiming Yang, and Kris Kitani. Rethinking transformer-based set prediction for object detection. arXiv preprint arXiv:2011.10881, 2020. \n[34] Mingxing Tan and Quoc V Le. Efficientnet: Rethinking model scaling for convolutional neural networks. arXiv preprint arXiv:1905.11946, 2019. \n[35] Ilya Tolstikhin, Neil Houlsby, Alexander Kolesnikov, Lucas Beyer, Xiaohua Zhai, Thomas Unterthiner, Jessica Yung, Daniel Keysers, Jakob Uszkoreit, Mario Lucic, et al. Mlp-mixer: An all-mlp architecture for vision. arXiv preprint arXiv:2105.01601, 2021. \n[36] Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, and Hervé Jégou. Training data-efficient image transformers & distillation through attention. arXiv preprint arXiv:2012.12877, 2020. \n[37] Hugo Touvron, Matthieu Cord, Alexandre Sablayrolles, Gabriel Synnaeve, and Hervé Jégou. Going deeper with image transformers. arXiv preprint arXiv:2103.17239, 2021. \n[38] Hugo Touvron, Andrea Vedaldi, Matthijs Douze, and Hervé Jégou. Fixing the train-test resolution discrepancy. arXiv preprint arXiv:1906.06423, 2019. \n[39] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30:5998–6008, 2017. \n[40] Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and Ling Shao. Pyramid vision transformer: A versatile backbone for dense prediction without convolutions. arXiv preprint arXiv:2102.12122, 2021. \n[41] Yuqing Wang, Zhaoliang Xu, Xinlong Wang, Chunhua Shen, Baoshan Cheng, Hao Shen, and Huaxia Xia. End-to-end video instance segmentation with transformers. arXiv preprint arXiv:2011.14503, 2020. \n[42] Ross Wightman. Pytorch image models. https://github.com/rwightman/ pytorch-image-models, 2019. \n[43] Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, and Lei Zhang. Cvt: Introducing convolutions to vision transformers. arXiv preprint arXiv:2103.15808, 2021. \n[44] Tete Xiao, Yingcheng Liu, Bolei Zhou, Yuning Jiang, and Jian Sun. Unified perceptual parsing for scene understanding. In Proceedings of the European Conference on Computer Vision (ECCV), pages 418–434, 2018. \n[45] Fuzhi Yang, Huan Yang, Jianlong Fu, Hongtao Lu, and Baining Guo. Learning texture transformer network for image super-resolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5791–5800, 2020. \n[46] Li Yuan, Yunpeng Chen, Tao Wang, Weihao Yu, Yujun Shi, Francis EH Tay, Jiashi Feng, and Shuicheng Yan. Tokens-to-token vit: Training vision transformers from scratch on imagenet. arXiv preprint arXiv:2101.11986, 2021. \n[47] Li Yuan, Francis EH Tay, Guilin Li, Tao Wang, and Jiashi Feng. Revisiting knowledge distillation via label smoothing regularization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3903–3911, 2020. \n[48] Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, and Youngjoon Yoo. Cutmix: Regularization strategy to train strong classifiers with localizable features. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 6023–6032, 2019. \n[49] Sangdoo Yun, Seong Joon Oh, Byeongho Heo, Dongyoon Han, Junsuk Choe, and Sanghyuk Chun. Re-labeling imagenet: from single to multi-labels, from global to localized labels. arXiv preprint arXiv:2101.05022, 2021. \n[50] Yanhong Zeng, Jianlong Fu, and Hongyang Chao. Learning joint spatial-temporal transformations for video inpainting. In European Conference on Computer Vision, pages 528–543. Springer, 2020. \n[51] Hang Zhang, Chongruo Wu, Zhongyue Zhang, Yi Zhu, Zhi Zhang, Haibin Lin, Yue Sun, Tong He, Jonas Muller, R. Manmatha, Mu Li, and Alexander Smola. Resnest: Split-attention networks. arXiv preprint arXiv:2004.08955, 2020. \n[52] Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and David Lopez-Paz. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412, 2017. \n[53] Hengshuang Zhao, Li Jiang, Jiaya Jia, Philip Torr, and Vladlen Koltun. Point transformer. arXiv preprint arXiv:2012.09164, 2020. \n[54] Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, and Jiaya Jia. Pyramid scene parsing network. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2881–2890, 2017. \n[55] Minghang Zheng, Peng Gao, Xiaogang Wang, Hongsheng Li, and Hao Dong. End-to-end object detection with adaptive clustering transformer. arXiv preprint arXiv:2011.09315, 2020. \n[56] Sixiao Zheng, Jiachen Lu, Hengshuang Zhao, Xiatian Zhu, Zekun Luo, Yabiao Wang, Yanwei Fu, Jianfeng Feng, Tao Xiang, Philip HS Torr, et al. Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers. arXiv preprint arXiv:2012.15840, 2020. \n[57] Zhun Zhong, Liang Zheng, Guoliang Kang, Shaozi Li, and Yi Yang. Random erasing data augmentation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 13001–13008, 2020. \n[58] Bolei Zhou, Hang Zhao, Xavier Puig, Tete Xiao, Sanja Fidler, Adela Barriuso, and Antonio Torralba. Semantic understanding of scenes through the ade20k dataset. International Journal of Computer Vision, 127(3):302–321, 2019. \n[59] Daquan Zhou, Bingyi Kang, Xiaojie Jin, Linjie Yang, Xiaochen Lian, Qibin Hou, and Jiashi Feng. Deepvit: Towards deeper vision transformer. arXiv preprint arXiv:2103.11886, 2021. \n[60] Luowei Zhou, Yingbo Zhou, Jason J Corso, Richard Socher, and Caiming Xiong. End-to-end dense video captioning with masked transformer. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 8739–8748, 2018. \n[61] Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, and Jifeng Dai. Deformable detr: Deformable transformers for end-to-end object detection. arXiv preprint arXiv:2010.04159, 2020. ", + "bbox": [ + 171, + 87, + 828, + 919 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "", + "bbox": [ + 171, + 73, + 828, + 919 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "", + "bbox": [ + 171, + 65, + 828, + 921 + ], + "page_idx": 12 + } +] \ No newline at end of file diff --git a/parse/train/2vubO341F_E/2vubO341F_E_middle.json b/parse/train/2vubO341F_E/2vubO341F_E_middle.json new file mode 100644 index 0000000000000000000000000000000000000000..d556ef9b879564e8f83653d2210add9de74012c2 --- /dev/null +++ b/parse/train/2vubO341F_E/2vubO341F_E_middle.json @@ -0,0 +1,31737 @@ +{ + "pdf_info": [ + { + "preproc_blocks": [ + { + "type": "title", + "bbox": [ + 128, + 97, + 484, + 137 + ], + "lines": [ + { + "bbox": [ + 125, + 95, + 485, + 119 + ], + "spans": [ + { + "bbox": [ + 125, + 95, + 485, + 119 + ], + "score": 1.0, + "content": "All Tokens Matter: Token Labeling for Training", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 203, + 117, + 409, + 138 + ], + "spans": [ + { + "bbox": [ + 203, + 117, + 409, + 138 + ], + "score": 1.0, + "content": "Better Vision Transformers", + "type": "text" + } + ], + "index": 1 + } + ], + "index": 0.5 + }, + { + "type": "text", + "bbox": [ + 122, + 174, + 486, + 188 + ], + "lines": [ + { + "bbox": [ + 122, + 173, + 488, + 191 + ], + "spans": [ + { + "bbox": [ + 122, + 173, + 194, + 191 + ], + "score": 1.0, + "content": "Zihang Jiang1∗", + "type": "text" + }, + { + "bbox": [ + 206, + 174, + 275, + 189 + ], + "score": 1.0, + "content": "Qibin Hou2,1†", + "type": "text" + }, + { + "bbox": [ + 285, + 174, + 331, + 189 + ], + "score": 1.0, + "content": "Li Yuan3", + "type": "text" + }, + { + "bbox": [ + 347, + 173, + 417, + 191 + ], + "score": 1.0, + "content": "Daquan Zhou1", + "type": "text" + }, + { + "bbox": [ + 435, + 173, + 488, + 190 + ], + "score": 1.0, + "content": "Yujun Shi1", + "type": "text" + } + ], + "index": 2 + } + ], + "index": 2 + }, + { + "type": "text", + "bbox": [ + 162, + 203, + 449, + 216 + ], + "lines": [ + { + "bbox": [ + 160, + 201, + 449, + 218 + ], + "spans": [ + { + "bbox": [ + 160, + 201, + 215, + 218 + ], + "score": 1.0, + "content": "Xiaojie Jin4", + "type": "text" + }, + { + "bbox": [ + 273, + 202, + 335, + 218 + ], + "score": 1.0, + "content": "Anran Wang4", + "type": "text" + }, + { + "bbox": [ + 393, + 202, + 449, + 218 + ], + "score": 1.0, + "content": "Jiashi Feng4", + "type": "text" + } + ], + "index": 3 + } + ], + "index": 3 + }, + { + "type": "text", + "bbox": [ + 185, + 232, + 425, + 244 + ], + "lines": [ + { + "bbox": [ + 184, + 230, + 427, + 249 + ], + "spans": [ + { + "bbox": [ + 184, + 230, + 427, + 249 + ], + "score": 1.0, + "content": "1National University of Singapore 2Nankai University", + "type": "text" + } + ], + "index": 4 + } + ], + "index": 4 + }, + { + "type": "title", + "bbox": [ + 232, + 260, + 379, + 272 + ], + "lines": [ + { + "bbox": [ + 230, + 257, + 380, + 276 + ], + "spans": [ + { + "bbox": [ + 230, + 257, + 380, + 276 + ], + "score": 1.0, + "content": "3 Peking University 4ByteDance", + "type": "text" + } + ], + "index": 5 + } + ], + "index": 5 + }, + { + "type": "text", + "bbox": [ + 128, + 273, + 483, + 294 + ], + "lines": [ + { + "bbox": [ + 128, + 272, + 485, + 284 + ], + "spans": [ + { + "bbox": [ + 128, + 272, + 485, + 284 + ], + "score": 1.0, + "content": "{jzh0103,andrewhoux,ylustcnus,zhoudaquan21,shiyujun1016}@gmail.com", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 131, + 283, + 480, + 295 + ], + "spans": [ + { + "bbox": [ + 131, + 283, + 480, + 295 + ], + "score": 1.0, + "content": "xjjin0731@gmail.com, anran.wang@bytedance.com, jshfeng@gmail.com", + "type": "text" + } + ], + "index": 7 + } + ], + "index": 6.5 + }, + { + "type": "title", + "bbox": [ + 283, + 322, + 328, + 335 + ], + "lines": [ + { + "bbox": [ + 281, + 322, + 330, + 336 + ], + "spans": [ + { + "bbox": [ + 281, + 322, + 330, + 336 + ], + "score": 1.0, + "content": "Abstract", + "type": "text" + } + ], + "index": 8 + } + ], + "index": 8 + }, + { + "type": "text", + "bbox": [ + 142, + 347, + 469, + 534 + ], + "lines": [ + { + "bbox": [ + 141, + 347, + 470, + 361 + ], + "spans": [ + { + "bbox": [ + 141, + 347, + 470, + 361 + ], + "score": 1.0, + "content": "In this paper, we present token labeling—a new training objective for training", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 141, + 357, + 470, + 372 + ], + "spans": [ + { + "bbox": [ + 141, + 357, + 470, + 372 + ], + "score": 1.0, + "content": "high-performance vision transformers (ViTs). Different from the standard training", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 142, + 370, + 469, + 382 + ], + "spans": [ + { + "bbox": [ + 142, + 370, + 469, + 382 + ], + "score": 1.0, + "content": "objective of ViTs that computes the classification loss on an additional trainable", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 141, + 381, + 470, + 393 + ], + "spans": [ + { + "bbox": [ + 141, + 381, + 470, + 393 + ], + "score": 1.0, + "content": "class token, our proposed one takes advantage of all the image patch tokens to com-", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 141, + 392, + 470, + 404 + ], + "spans": [ + { + "bbox": [ + 141, + 392, + 470, + 404 + ], + "score": 1.0, + "content": "pute the training loss in a dense manner. Specifically, token labeling reformulates", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 141, + 402, + 469, + 415 + ], + "spans": [ + { + "bbox": [ + 141, + 402, + 469, + 415 + ], + "score": 1.0, + "content": "the image classification problem into multiple token-level recognition problems", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 141, + 413, + 470, + 426 + ], + "spans": [ + { + "bbox": [ + 141, + 413, + 470, + 426 + ], + "score": 1.0, + "content": "and assigns each patch token with an individual location-specific supervision gen-", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 141, + 424, + 470, + 437 + ], + "spans": [ + { + "bbox": [ + 141, + 424, + 470, + 437 + ], + "score": 1.0, + "content": "erated by a machine annotator. Experiments show that token labeling can clearly", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 141, + 435, + 469, + 447 + ], + "spans": [ + { + "bbox": [ + 141, + 435, + 469, + 447 + ], + "score": 1.0, + "content": "and consistently improve the performance of various ViT models across a wide", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 141, + 445, + 470, + 460 + ], + "spans": [ + { + "bbox": [ + 141, + 445, + 470, + 460 + ], + "score": 1.0, + "content": "spectrum. For a vision transformer with 26M learnable parameters serving as an", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 141, + 456, + 469, + 469 + ], + "spans": [ + { + "bbox": [ + 141, + 456, + 361, + 469 + ], + "score": 1.0, + "content": "example, with token labeling, the model can achieve", + "type": "text" + }, + { + "bbox": [ + 361, + 457, + 389, + 467 + ], + "score": 0.86, + "content": "8 4 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 389, + 456, + 469, + 469 + ], + "score": 1.0, + "content": "Top-1 accuracy on", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 141, + 467, + 470, + 480 + ], + "spans": [ + { + "bbox": [ + 141, + 467, + 346, + 480 + ], + "score": 1.0, + "content": "ImageNet. The result can be further increased to", + "type": "text" + }, + { + "bbox": [ + 347, + 468, + 374, + 478 + ], + "score": 0.88, + "content": "8 6 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 375, + 467, + 470, + 480 + ], + "score": 1.0, + "content": "by slightly scaling the", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 141, + 479, + 470, + 491 + ], + "spans": [ + { + "bbox": [ + 141, + 479, + 470, + 491 + ], + "score": 1.0, + "content": "model size up to 150M, delivering the minimal-sized model among previous mod-", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 141, + 489, + 469, + 502 + ], + "spans": [ + { + "bbox": [ + 141, + 489, + 156, + 502 + ], + "score": 1.0, + "content": "els", + "type": "text" + }, + { + "bbox": [ + 157, + 489, + 192, + 500 + ], + "score": 0.76, + "content": "( 2 5 0 \\mathbf { M } + )", + "type": "inline_equation" + }, + { + "bbox": [ + 192, + 489, + 231, + 502 + ], + "score": 1.0, + "content": "reaching", + "type": "text" + }, + { + "bbox": [ + 231, + 489, + 250, + 500 + ], + "score": 0.87, + "content": "86 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 250, + 489, + 469, + 502 + ], + "score": 1.0, + "content": ". We also show that token labeling can clearly improve", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 142, + 500, + 469, + 513 + ], + "spans": [ + { + "bbox": [ + 142, + 500, + 469, + 513 + ], + "score": 1.0, + "content": "the generalization capability of the pretrained models on downstream tasks with", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 141, + 510, + 470, + 525 + ], + "spans": [ + { + "bbox": [ + 141, + 510, + 470, + 525 + ], + "score": 1.0, + "content": "dense prediction, such as semantic segmentation. Our code and model are publicly", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 141, + 521, + 459, + 536 + ], + "spans": [ + { + "bbox": [ + 141, + 521, + 459, + 536 + ], + "score": 1.0, + "content": "available at https://github.com/zihangJiang/TokenLabeling.", + "type": "text" + } + ], + "index": 25 + } + ], + "index": 17 + }, + { + "type": "title", + "bbox": [ + 107, + 554, + 190, + 568 + ], + "lines": [ + { + "bbox": [ + 105, + 552, + 192, + 570 + ], + "spans": [ + { + "bbox": [ + 105, + 552, + 192, + 570 + ], + "score": 1.0, + "content": "1 Introduction", + "type": "text" + } + ], + "index": 26 + } + ], + "index": 26 + }, + { + "type": "text", + "bbox": [ + 107, + 579, + 505, + 645 + ], + "lines": [ + { + "bbox": [ + 105, + 577, + 506, + 594 + ], + "spans": [ + { + "bbox": [ + 105, + 577, + 506, + 594 + ], + "score": 1.0, + "content": "Transformers [39] have achieved great performance for almost all the natural language processing", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 105, + 590, + 505, + 603 + ], + "spans": [ + { + "bbox": [ + 105, + 590, + 505, + 603 + ], + "score": 1.0, + "content": "(NLP) tasks over the past years [4, 14, 24]. Motivated by such success, recently, many researchers", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 601, + 505, + 614 + ], + "spans": [ + { + "bbox": [ + 105, + 601, + 505, + 614 + ], + "score": 1.0, + "content": "attempt to build transformer models for vision tasks, and their encouraging results have shown the", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 612, + 505, + 626 + ], + "spans": [ + { + "bbox": [ + 105, + 612, + 505, + 626 + ], + "score": 1.0, + "content": "great potential of transformer based models for image classification [6, 15, 25, 36, 40, 46], especially", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 622, + 505, + 636 + ], + "spans": [ + { + "bbox": [ + 105, + 622, + 505, + 636 + ], + "score": 1.0, + "content": "the strong benefits of the self-attention mechanism in building long-range dependencies between", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 634, + 193, + 646 + ], + "spans": [ + { + "bbox": [ + 105, + 634, + 193, + 646 + ], + "score": 1.0, + "content": "pairs of input tokens.", + "type": "text" + } + ], + "index": 32 + } + ], + "index": 29.5 + }, + { + "type": "text", + "bbox": [ + 108, + 650, + 505, + 684 + ], + "lines": [ + { + "bbox": [ + 106, + 650, + 507, + 663 + ], + "spans": [ + { + "bbox": [ + 106, + 650, + 507, + 663 + ], + "score": 1.0, + "content": "Despite the importance of gathering long-range dependencies, recent work on local data augmenta-", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 106, + 661, + 506, + 673 + ], + "spans": [ + { + "bbox": [ + 106, + 661, + 506, + 673 + ], + "score": 1.0, + "content": "tion [57] has demonstrated that well modeling and leveraging local information for image classifica-", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 105, + 672, + 505, + 685 + ], + "spans": [ + { + "bbox": [ + 105, + 672, + 505, + 685 + ], + "score": 1.0, + "content": "tion would avoid biasing the model towards skewed and non-generalizable patterns and substantially", + "type": "text" + } + ], + "index": 35 + } + ], + "index": 34 + } + ], + "page_idx": 0, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 118, + 691, + 407, + 713 + ], + "lines": [ + { + "bbox": [ + 118, + 689, + 288, + 702 + ], + "spans": [ + { + "bbox": [ + 118, + 689, + 288, + 702 + ], + "score": 1.0, + "content": "∗Work done as an intern at ByteDance AI Lab.", + "type": "text" + } + ] + }, + { + "bbox": [ + 118, + 700, + 408, + 714 + ], + "spans": [ + { + "bbox": [ + 118, + 700, + 408, + 714 + ], + "score": 1.0, + "content": "†Corresponding author. Part of this work was done as a research fellow at NUS.", + "type": "text" + } + ] + } + ] + }, + { + "type": "discarded", + "bbox": [ + 106, + 731, + 385, + 742 + ], + "lines": [ + { + "bbox": [ + 105, + 730, + 386, + 743 + ], + "spans": [ + { + "bbox": [ + 105, + 730, + 386, + 743 + ], + "score": 1.0, + "content": "35th Conference on Neural Information Processing Systems (NeurIPS 2021).", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "title", + "bbox": [ + 128, + 97, + 484, + 137 + ], + "lines": [ + { + "bbox": [ + 125, + 95, + 485, + 119 + ], + "spans": [ + { + "bbox": [ + 125, + 95, + 485, + 119 + ], + "score": 1.0, + "content": "All Tokens Matter: Token Labeling for Training", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 203, + 117, + 409, + 138 + ], + "spans": [ + { + "bbox": [ + 203, + 117, + 409, + 138 + ], + "score": 1.0, + "content": "Better Vision Transformers", + "type": "text" + } + ], + "index": 1 + } + ], + "index": 0.5 + }, + { + "type": "text", + "bbox": [ + 122, + 174, + 486, + 188 + ], + "lines": [ + { + "bbox": [ + 122, + 173, + 488, + 191 + ], + "spans": [ + { + "bbox": [ + 122, + 173, + 194, + 191 + ], + "score": 1.0, + "content": "Zihang Jiang1∗", + "type": "text" + }, + { + "bbox": [ + 206, + 174, + 275, + 189 + ], + "score": 1.0, + "content": "Qibin Hou2,1†", + "type": "text" + }, + { + "bbox": [ + 285, + 174, + 331, + 189 + ], + "score": 1.0, + "content": "Li Yuan3", + "type": "text" + }, + { + "bbox": [ + 347, + 173, + 417, + 191 + ], + "score": 1.0, + "content": "Daquan Zhou1", + "type": "text" + }, + { + "bbox": [ + 435, + 173, + 488, + 190 + ], + "score": 1.0, + "content": "Yujun Shi1", + "type": "text" + } + ], + "index": 2 + } + ], + "index": 2, + "bbox_fs": [ + 122, + 173, + 488, + 191 + ] + }, + { + "type": "text", + "bbox": [ + 162, + 203, + 449, + 216 + ], + "lines": [ + { + "bbox": [ + 160, + 201, + 449, + 218 + ], + "spans": [ + { + "bbox": [ + 160, + 201, + 215, + 218 + ], + "score": 1.0, + "content": "Xiaojie Jin4", + "type": "text" + }, + { + "bbox": [ + 273, + 202, + 335, + 218 + ], + "score": 1.0, + "content": "Anran Wang4", + "type": "text" + }, + { + "bbox": [ + 393, + 202, + 449, + 218 + ], + "score": 1.0, + "content": "Jiashi Feng4", + "type": "text" + } + ], + "index": 3 + } + ], + "index": 3, + "bbox_fs": [ + 160, + 201, + 449, + 218 + ] + }, + { + "type": "text", + "bbox": [ + 185, + 232, + 425, + 244 + ], + "lines": [ + { + "bbox": [ + 184, + 230, + 427, + 249 + ], + "spans": [ + { + "bbox": [ + 184, + 230, + 427, + 249 + ], + "score": 1.0, + "content": "1National University of Singapore 2Nankai University", + "type": "text" + } + ], + "index": 4 + } + ], + "index": 4, + "bbox_fs": [ + 184, + 230, + 427, + 249 + ] + }, + { + "type": "title", + "bbox": [ + 232, + 260, + 379, + 272 + ], + "lines": [ + { + "bbox": [ + 230, + 257, + 380, + 276 + ], + "spans": [ + { + "bbox": [ + 230, + 257, + 380, + 276 + ], + "score": 1.0, + "content": "3 Peking University 4ByteDance", + "type": "text" + } + ], + "index": 5 + } + ], + "index": 5 + }, + { + "type": "text", + "bbox": [ + 128, + 273, + 483, + 294 + ], + "lines": [ + { + "bbox": [ + 128, + 272, + 485, + 284 + ], + "spans": [ + { + "bbox": [ + 128, + 272, + 485, + 284 + ], + "score": 1.0, + "content": "{jzh0103,andrewhoux,ylustcnus,zhoudaquan21,shiyujun1016}@gmail.com", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 131, + 283, + 480, + 295 + ], + "spans": [ + { + "bbox": [ + 131, + 283, + 480, + 295 + ], + "score": 1.0, + "content": "xjjin0731@gmail.com, anran.wang@bytedance.com, jshfeng@gmail.com", + "type": "text" + } + ], + "index": 7 + } + ], + "index": 6.5, + "bbox_fs": [ + 128, + 272, + 485, + 295 + ] + }, + { + "type": "title", + "bbox": [ + 283, + 322, + 328, + 335 + ], + "lines": [ + { + "bbox": [ + 281, + 322, + 330, + 336 + ], + "spans": [ + { + "bbox": [ + 281, + 322, + 330, + 336 + ], + "score": 1.0, + "content": "Abstract", + "type": "text" + } + ], + "index": 8 + } + ], + "index": 8 + }, + { + "type": "text", + "bbox": [ + 142, + 347, + 469, + 534 + ], + "lines": [ + { + "bbox": [ + 141, + 347, + 470, + 361 + ], + "spans": [ + { + "bbox": [ + 141, + 347, + 470, + 361 + ], + "score": 1.0, + "content": "In this paper, we present token labeling—a new training objective for training", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 141, + 357, + 470, + 372 + ], + "spans": [ + { + "bbox": [ + 141, + 357, + 470, + 372 + ], + "score": 1.0, + "content": "high-performance vision transformers (ViTs). Different from the standard training", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 142, + 370, + 469, + 382 + ], + "spans": [ + { + "bbox": [ + 142, + 370, + 469, + 382 + ], + "score": 1.0, + "content": "objective of ViTs that computes the classification loss on an additional trainable", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 141, + 381, + 470, + 393 + ], + "spans": [ + { + "bbox": [ + 141, + 381, + 470, + 393 + ], + "score": 1.0, + "content": "class token, our proposed one takes advantage of all the image patch tokens to com-", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 141, + 392, + 470, + 404 + ], + "spans": [ + { + "bbox": [ + 141, + 392, + 470, + 404 + ], + "score": 1.0, + "content": "pute the training loss in a dense manner. Specifically, token labeling reformulates", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 141, + 402, + 469, + 415 + ], + "spans": [ + { + "bbox": [ + 141, + 402, + 469, + 415 + ], + "score": 1.0, + "content": "the image classification problem into multiple token-level recognition problems", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 141, + 413, + 470, + 426 + ], + "spans": [ + { + "bbox": [ + 141, + 413, + 470, + 426 + ], + "score": 1.0, + "content": "and assigns each patch token with an individual location-specific supervision gen-", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 141, + 424, + 470, + 437 + ], + "spans": [ + { + "bbox": [ + 141, + 424, + 470, + 437 + ], + "score": 1.0, + "content": "erated by a machine annotator. Experiments show that token labeling can clearly", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 141, + 435, + 469, + 447 + ], + "spans": [ + { + "bbox": [ + 141, + 435, + 469, + 447 + ], + "score": 1.0, + "content": "and consistently improve the performance of various ViT models across a wide", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 141, + 445, + 470, + 460 + ], + "spans": [ + { + "bbox": [ + 141, + 445, + 470, + 460 + ], + "score": 1.0, + "content": "spectrum. For a vision transformer with 26M learnable parameters serving as an", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 141, + 456, + 469, + 469 + ], + "spans": [ + { + "bbox": [ + 141, + 456, + 361, + 469 + ], + "score": 1.0, + "content": "example, with token labeling, the model can achieve", + "type": "text" + }, + { + "bbox": [ + 361, + 457, + 389, + 467 + ], + "score": 0.86, + "content": "8 4 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 389, + 456, + 469, + 469 + ], + "score": 1.0, + "content": "Top-1 accuracy on", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 141, + 467, + 470, + 480 + ], + "spans": [ + { + "bbox": [ + 141, + 467, + 346, + 480 + ], + "score": 1.0, + "content": "ImageNet. The result can be further increased to", + "type": "text" + }, + { + "bbox": [ + 347, + 468, + 374, + 478 + ], + "score": 0.88, + "content": "8 6 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 375, + 467, + 470, + 480 + ], + "score": 1.0, + "content": "by slightly scaling the", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 141, + 479, + 470, + 491 + ], + "spans": [ + { + "bbox": [ + 141, + 479, + 470, + 491 + ], + "score": 1.0, + "content": "model size up to 150M, delivering the minimal-sized model among previous mod-", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 141, + 489, + 469, + 502 + ], + "spans": [ + { + "bbox": [ + 141, + 489, + 156, + 502 + ], + "score": 1.0, + "content": "els", + "type": "text" + }, + { + "bbox": [ + 157, + 489, + 192, + 500 + ], + "score": 0.76, + "content": "( 2 5 0 \\mathbf { M } + )", + "type": "inline_equation" + }, + { + "bbox": [ + 192, + 489, + 231, + 502 + ], + "score": 1.0, + "content": "reaching", + "type": "text" + }, + { + "bbox": [ + 231, + 489, + 250, + 500 + ], + "score": 0.87, + "content": "86 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 250, + 489, + 469, + 502 + ], + "score": 1.0, + "content": ". We also show that token labeling can clearly improve", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 142, + 500, + 469, + 513 + ], + "spans": [ + { + "bbox": [ + 142, + 500, + 469, + 513 + ], + "score": 1.0, + "content": "the generalization capability of the pretrained models on downstream tasks with", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 141, + 510, + 470, + 525 + ], + "spans": [ + { + "bbox": [ + 141, + 510, + 470, + 525 + ], + "score": 1.0, + "content": "dense prediction, such as semantic segmentation. Our code and model are publicly", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 141, + 521, + 459, + 536 + ], + "spans": [ + { + "bbox": [ + 141, + 521, + 459, + 536 + ], + "score": 1.0, + "content": "available at https://github.com/zihangJiang/TokenLabeling.", + "type": "text" + } + ], + "index": 25 + } + ], + "index": 17, + "bbox_fs": [ + 141, + 347, + 470, + 536 + ] + }, + { + "type": "title", + "bbox": [ + 107, + 554, + 190, + 568 + ], + "lines": [ + { + "bbox": [ + 105, + 552, + 192, + 570 + ], + "spans": [ + { + "bbox": [ + 105, + 552, + 192, + 570 + ], + "score": 1.0, + "content": "1 Introduction", + "type": "text" + } + ], + "index": 26 + } + ], + "index": 26 + }, + { + "type": "text", + "bbox": [ + 107, + 579, + 505, + 645 + ], + "lines": [ + { + "bbox": [ + 105, + 577, + 506, + 594 + ], + "spans": [ + { + "bbox": [ + 105, + 577, + 506, + 594 + ], + "score": 1.0, + "content": "Transformers [39] have achieved great performance for almost all the natural language processing", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 105, + 590, + 505, + 603 + ], + "spans": [ + { + "bbox": [ + 105, + 590, + 505, + 603 + ], + "score": 1.0, + "content": "(NLP) tasks over the past years [4, 14, 24]. Motivated by such success, recently, many researchers", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 601, + 505, + 614 + ], + "spans": [ + { + "bbox": [ + 105, + 601, + 505, + 614 + ], + "score": 1.0, + "content": "attempt to build transformer models for vision tasks, and their encouraging results have shown the", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 612, + 505, + 626 + ], + "spans": [ + { + "bbox": [ + 105, + 612, + 505, + 626 + ], + "score": 1.0, + "content": "great potential of transformer based models for image classification [6, 15, 25, 36, 40, 46], especially", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 622, + 505, + 636 + ], + "spans": [ + { + "bbox": [ + 105, + 622, + 505, + 636 + ], + "score": 1.0, + "content": "the strong benefits of the self-attention mechanism in building long-range dependencies between", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 634, + 193, + 646 + ], + "spans": [ + { + "bbox": [ + 105, + 634, + 193, + 646 + ], + "score": 1.0, + "content": "pairs of input tokens.", + "type": "text" + } + ], + "index": 32 + } + ], + "index": 29.5, + "bbox_fs": [ + 105, + 577, + 506, + 646 + ] + }, + { + "type": "text", + "bbox": [ + 108, + 650, + 505, + 684 + ], + "lines": [ + { + "bbox": [ + 106, + 650, + 507, + 663 + ], + "spans": [ + { + "bbox": [ + 106, + 650, + 507, + 663 + ], + "score": 1.0, + "content": "Despite the importance of gathering long-range dependencies, recent work on local data augmenta-", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 106, + 661, + 506, + 673 + ], + "spans": [ + { + "bbox": [ + 106, + 661, + 506, + 673 + ], + "score": 1.0, + "content": "tion [57] has demonstrated that well modeling and leveraging local information for image classifica-", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 105, + 672, + 505, + 685 + ], + "spans": [ + { + "bbox": [ + 105, + 672, + 505, + 685 + ], + "score": 1.0, + "content": "tion would avoid biasing the model towards skewed and non-generalizable patterns and substantially", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 106, + 304, + 505, + 316 + ], + "spans": [ + { + "bbox": [ + 106, + 304, + 505, + 316 + ], + "score": 1.0, + "content": "improve the model performance. However, recent vision transformers normally utilize class tokens", + "type": "text", + "cross_page": true + } + ], + "index": 9 + }, + { + "bbox": [ + 106, + 315, + 505, + 327 + ], + "spans": [ + { + "bbox": [ + 106, + 315, + 505, + 327 + ], + "score": 1.0, + "content": "that aggregate global information to predict the output class while neglecting the role of other patch", + "type": "text", + "cross_page": true + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 325, + 410, + 339 + ], + "spans": [ + { + "bbox": [ + 105, + 325, + 410, + 339 + ], + "score": 1.0, + "content": "tokens that encode rich information on their respective local image patches.", + "type": "text", + "cross_page": true + } + ], + "index": 11 + } + ], + "index": 34, + "bbox_fs": [ + 105, + 650, + 507, + 685 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "image", + "bbox": [ + 107, + 69, + 505, + 210 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 107, + 69, + 505, + 210 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 107, + 69, + 505, + 210 + ], + "spans": [ + { + "bbox": [ + 107, + 69, + 505, + 210 + ], + "score": 0.971, + "type": "image", + "image_path": "9c7bb663ccabca3baefdad385c72c84db533d97dbd497ac18638700c9e1d548b.jpg" + } + ] + } + ], + "index": 1, + "virtual_lines": [ + { + "bbox": [ + 107, + 69, + 505, + 116.0 + ], + "spans": [], + "index": 0 + }, + { + "bbox": [ + 107, + 116.0, + 505, + 163.0 + ], + "spans": [], + "index": 1 + }, + { + "bbox": [ + 107, + 163.0, + 505, + 210.0 + ], + "spans": [], + "index": 2 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 106, + 216, + 506, + 281 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 215, + 507, + 229 + ], + "spans": [ + { + "bbox": [ + 105, + 215, + 507, + 229 + ], + "score": 1.0, + "content": "Figure 1: Comparison between the proposed LV-ViT and other recent works based on vision trans-", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 105, + 227, + 507, + 240 + ], + "spans": [ + { + "bbox": [ + 105, + 227, + 507, + 240 + ], + "score": 1.0, + "content": "formers, including T2T-ViT [46], ConViT [12], BoTNet [31], DeepViT [59], DeiT [36], ViT [15],", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 237, + 506, + 250 + ], + "spans": [ + { + "bbox": [ + 105, + 237, + 506, + 250 + ], + "score": 1.0, + "content": "Swin Transformer [25], LambdaNet [1], CvT [43], CrossViT [6], PVT [40], CaiT [37]. Note that", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 105, + 249, + 505, + 261 + ], + "spans": [ + { + "bbox": [ + 105, + 249, + 505, + 261 + ], + "score": 1.0, + "content": "we only show models whose model sizes are under 100M. As can be seen, our LV-ViT achieves the", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 105, + 259, + 505, + 272 + ], + "spans": [ + { + "bbox": [ + 105, + 259, + 461, + 272 + ], + "score": 1.0, + "content": "best results using the least amount of learnable parameters. The default test resolution is", + "type": "text" + }, + { + "bbox": [ + 461, + 260, + 505, + 270 + ], + "score": 0.9, + "content": "2 2 4 \\times 2 2 4", + "type": "inline_equation" + } + ], + "index": 7 + }, + { + "bbox": [ + 106, + 270, + 207, + 283 + ], + "spans": [ + { + "bbox": [ + 106, + 270, + 193, + 283 + ], + "score": 1.0, + "content": "unless specified after", + "type": "text" + }, + { + "bbox": [ + 193, + 271, + 203, + 280 + ], + "score": 0.8, + "content": "@", + "type": "inline_equation" + }, + { + "bbox": [ + 203, + 270, + 207, + 283 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 8 + } + ], + "index": 5.5 + } + ], + "index": 3.25 + }, + { + "type": "text", + "bbox": [ + 107, + 303, + 505, + 337 + ], + "lines": [ + { + "bbox": [ + 106, + 304, + 505, + 316 + ], + "spans": [ + { + "bbox": [ + 106, + 304, + 505, + 316 + ], + "score": 1.0, + "content": "improve the model performance. However, recent vision transformers normally utilize class tokens", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 106, + 315, + 505, + 327 + ], + "spans": [ + { + "bbox": [ + 106, + 315, + 505, + 327 + ], + "score": 1.0, + "content": "that aggregate global information to predict the output class while neglecting the role of other patch", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 325, + 410, + 339 + ], + "spans": [ + { + "bbox": [ + 105, + 325, + 410, + 339 + ], + "score": 1.0, + "content": "tokens that encode rich information on their respective local image patches.", + "type": "text" + } + ], + "index": 11 + } + ], + "index": 10 + }, + { + "type": "text", + "bbox": [ + 107, + 342, + 505, + 441 + ], + "lines": [ + { + "bbox": [ + 106, + 343, + 505, + 354 + ], + "spans": [ + { + "bbox": [ + 106, + 343, + 505, + 354 + ], + "score": 1.0, + "content": "In this paper, we present a new training objective for vision transformers, termed token labeling, that", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 106, + 353, + 505, + 365 + ], + "spans": [ + { + "bbox": [ + 106, + 353, + 440, + 365 + ], + "score": 1.0, + "content": "takes advantage of both the patch tokens and the class tokens. Our method takes a", + "type": "text" + }, + { + "bbox": [ + 441, + 354, + 451, + 363 + ], + "score": 0.82, + "content": "K", + "type": "inline_equation" + }, + { + "bbox": [ + 451, + 353, + 505, + 365 + ], + "score": 1.0, + "content": "-dimensional", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 364, + 506, + 377 + ], + "spans": [ + { + "bbox": [ + 105, + 364, + 506, + 377 + ], + "score": 1.0, + "content": "score map generated by a machine annotator as supervision to supervise all the tokens in a dense", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 105, + 375, + 505, + 388 + ], + "spans": [ + { + "bbox": [ + 105, + 375, + 169, + 388 + ], + "score": 1.0, + "content": "manner, where", + "type": "text" + }, + { + "bbox": [ + 169, + 375, + 180, + 385 + ], + "score": 0.81, + "content": "K", + "type": "inline_equation" + }, + { + "bbox": [ + 180, + 375, + 505, + 388 + ], + "score": 1.0, + "content": "is the number of categories for the target dataset. In this way, each patch token", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 105, + 386, + 506, + 399 + ], + "spans": [ + { + "bbox": [ + 105, + 386, + 506, + 399 + ], + "score": 1.0, + "content": "is explicitly associated with an individual location-specific supervision indicating the existence of", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 397, + 506, + 410 + ], + "spans": [ + { + "bbox": [ + 105, + 397, + 506, + 410 + ], + "score": 1.0, + "content": "the target objects inside the corresponding image patch, so as to improve the object grounding and", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 408, + 506, + 420 + ], + "spans": [ + { + "bbox": [ + 105, + 408, + 506, + 420 + ], + "score": 1.0, + "content": "recognition capabilities of vision transformers with negligible computation overhead. To the best", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 105, + 417, + 506, + 431 + ], + "spans": [ + { + "bbox": [ + 105, + 417, + 506, + 431 + ], + "score": 1.0, + "content": "of our knowledge, this is the first work demonstrating that dense supervision is beneficial to vision", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 105, + 429, + 253, + 442 + ], + "spans": [ + { + "bbox": [ + 105, + 429, + 253, + 442 + ], + "score": 1.0, + "content": "transformers in image classification.", + "type": "text" + } + ], + "index": 20 + } + ], + "index": 16 + }, + { + "type": "text", + "bbox": [ + 107, + 446, + 505, + 523 + ], + "lines": [ + { + "bbox": [ + 105, + 446, + 506, + 458 + ], + "spans": [ + { + "bbox": [ + 105, + 446, + 506, + 458 + ], + "score": 1.0, + "content": "According to our experiments, utilizing the proposed token labeling objective can clearly boost", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 105, + 457, + 505, + 469 + ], + "spans": [ + { + "bbox": [ + 105, + 457, + 505, + 469 + ], + "score": 1.0, + "content": "the performance of vision transformers. As shown in Figure 1, our model, named LV-ViT, with", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 467, + 506, + 480 + ], + "spans": [ + { + "bbox": [ + 105, + 467, + 203, + 480 + ], + "score": 1.0, + "content": "56M parameters, yields", + "type": "text" + }, + { + "bbox": [ + 204, + 468, + 231, + 478 + ], + "score": 0.87, + "content": "8 5 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 231, + 467, + 506, + 480 + ], + "score": 1.0, + "content": "top-1 accuracy on ImageNet [13], behaving better than all the other", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 477, + 505, + 492 + ], + "spans": [ + { + "bbox": [ + 105, + 477, + 505, + 492 + ], + "score": 1.0, + "content": "transformer-based models having no more than 100M parameters. When the model size is scaled up", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 106, + 489, + 504, + 501 + ], + "spans": [ + { + "bbox": [ + 106, + 490, + 289, + 501 + ], + "score": 1.0, + "content": "to 150M, the result can be further improved to", + "type": "text" + }, + { + "bbox": [ + 289, + 489, + 316, + 500 + ], + "score": 0.87, + "content": "8 6 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 316, + 490, + 504, + 501 + ], + "score": 1.0, + "content": ". In addition, we have empirically found that the", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 105, + 500, + 507, + 513 + ], + "spans": [ + { + "bbox": [ + 105, + 500, + 507, + 513 + ], + "score": 1.0, + "content": "pretrained models with token labeling are also beneficial to downstream tasks with dense prediction,", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 105, + 511, + 234, + 524 + ], + "spans": [ + { + "bbox": [ + 105, + 511, + 234, + 524 + ], + "score": 1.0, + "content": "such as semantic segmentation.", + "type": "text" + } + ], + "index": 27 + } + ], + "index": 24 + }, + { + "type": "title", + "bbox": [ + 107, + 538, + 197, + 552 + ], + "lines": [ + { + "bbox": [ + 105, + 537, + 198, + 553 + ], + "spans": [ + { + "bbox": [ + 105, + 537, + 198, + 553 + ], + "score": 1.0, + "content": "2 Related Work", + "type": "text" + } + ], + "index": 28 + } + ], + "index": 28 + }, + { + "type": "text", + "bbox": [ + 107, + 563, + 505, + 641 + ], + "lines": [ + { + "bbox": [ + 105, + 563, + 505, + 576 + ], + "spans": [ + { + "bbox": [ + 105, + 563, + 505, + 576 + ], + "score": 1.0, + "content": "Transformers [39] refer to the models that entirely rely on the self-attention mechanism to build", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 574, + 506, + 587 + ], + "spans": [ + { + "bbox": [ + 105, + 574, + 506, + 587 + ], + "score": 1.0, + "content": "global dependencies, which are originally designed for natural language processing tasks. Due to their", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 585, + 505, + 597 + ], + "spans": [ + { + "bbox": [ + 105, + 585, + 505, + 597 + ], + "score": 1.0, + "content": "strong capability of capturing spatial information, transformers have also been successfully applied", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 596, + 505, + 609 + ], + "spans": [ + { + "bbox": [ + 105, + 596, + 505, + 609 + ], + "score": 1.0, + "content": "to a variety of vision problems, including low-level vision tasks like image enhancement [7, 45], as", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 105, + 607, + 506, + 619 + ], + "spans": [ + { + "bbox": [ + 105, + 607, + 506, + 619 + ], + "score": 1.0, + "content": "well as more challenging tasks such as image classification [9, 15], object detection [5, 11, 55, 61],", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 618, + 506, + 630 + ], + "spans": [ + { + "bbox": [ + 105, + 618, + 506, + 630 + ], + "score": 1.0, + "content": "segmentation [7, 33, 41] and image generation [28]. Some works also extend transformers for video", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 106, + 630, + 282, + 642 + ], + "spans": [ + { + "bbox": [ + 106, + 630, + 282, + 642 + ], + "score": 1.0, + "content": "and 3D point cloud processing [50, 53, 60].", + "type": "text" + } + ], + "index": 35 + } + ], + "index": 32 + }, + { + "type": "text", + "bbox": [ + 107, + 645, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 105, + 645, + 505, + 658 + ], + "spans": [ + { + "bbox": [ + 105, + 645, + 505, + 658 + ], + "score": 1.0, + "content": "Vision Transformer (ViT) is one of the earlier attempts that achieved state-of-the-art performance", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 106, + 656, + 505, + 669 + ], + "spans": [ + { + "bbox": [ + 106, + 656, + 505, + 669 + ], + "score": 1.0, + "content": "on ImageNet classification, using pure transformers as basic building blocks. However, ViTs need", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 105, + 667, + 506, + 680 + ], + "spans": [ + { + "bbox": [ + 105, + 667, + 506, + 680 + ], + "score": 1.0, + "content": "pretraining on very large datasets, such as ImageNet-22k and JFT-300M, and huge computation", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 105, + 678, + 505, + 690 + ], + "spans": [ + { + "bbox": [ + 105, + 678, + 505, + 690 + ], + "score": 1.0, + "content": "resources to achieve comparable performance to ResNet [18] with a similar model size trained on", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 105, + 688, + 505, + 702 + ], + "spans": [ + { + "bbox": [ + 105, + 688, + 505, + 702 + ], + "score": 1.0, + "content": "ImageNet. Later, DeiT [36] manages to tackle the data-inefficiency problem by simply adjusting", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 105, + 700, + 505, + 712 + ], + "spans": [ + { + "bbox": [ + 105, + 700, + 505, + 712 + ], + "score": 1.0, + "content": "the network architecture and adding an additional token along with the class token for Knowledge", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 106, + 710, + 316, + 723 + ], + "spans": [ + { + "bbox": [ + 106, + 710, + 316, + 723 + ], + "score": 1.0, + "content": "Distillation [21, 47] to improve model performance.", + "type": "text" + } + ], + "index": 42 + } + ], + "index": 39 + } + ], + "page_idx": 1, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 302, + 741, + 309, + 750 + ], + "lines": [ + { + "bbox": [ + 301, + 740, + 310, + 753 + ], + "spans": [ + { + "bbox": [ + 301, + 740, + 310, + 753 + ], + "score": 1.0, + "content": "2", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "image", + "bbox": [ + 107, + 69, + 505, + 210 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 107, + 69, + 505, + 210 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 107, + 69, + 505, + 210 + ], + "spans": [ + { + "bbox": [ + 107, + 69, + 505, + 210 + ], + "score": 0.971, + "type": "image", + "image_path": "9c7bb663ccabca3baefdad385c72c84db533d97dbd497ac18638700c9e1d548b.jpg" + } + ] + } + ], + "index": 1, + "virtual_lines": [ + { + "bbox": [ + 107, + 69, + 505, + 116.0 + ], + "spans": [], + "index": 0 + }, + { + "bbox": [ + 107, + 116.0, + 505, + 163.0 + ], + "spans": [], + "index": 1 + }, + { + "bbox": [ + 107, + 163.0, + 505, + 210.0 + ], + "spans": [], + "index": 2 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 106, + 216, + 506, + 281 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 215, + 507, + 229 + ], + "spans": [ + { + "bbox": [ + 105, + 215, + 507, + 229 + ], + "score": 1.0, + "content": "Figure 1: Comparison between the proposed LV-ViT and other recent works based on vision trans-", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 105, + 227, + 507, + 240 + ], + "spans": [ + { + "bbox": [ + 105, + 227, + 507, + 240 + ], + "score": 1.0, + "content": "formers, including T2T-ViT [46], ConViT [12], BoTNet [31], DeepViT [59], DeiT [36], ViT [15],", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 237, + 506, + 250 + ], + "spans": [ + { + "bbox": [ + 105, + 237, + 506, + 250 + ], + "score": 1.0, + "content": "Swin Transformer [25], LambdaNet [1], CvT [43], CrossViT [6], PVT [40], CaiT [37]. Note that", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 105, + 249, + 505, + 261 + ], + "spans": [ + { + "bbox": [ + 105, + 249, + 505, + 261 + ], + "score": 1.0, + "content": "we only show models whose model sizes are under 100M. As can be seen, our LV-ViT achieves the", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 105, + 259, + 505, + 272 + ], + "spans": [ + { + "bbox": [ + 105, + 259, + 461, + 272 + ], + "score": 1.0, + "content": "best results using the least amount of learnable parameters. The default test resolution is", + "type": "text" + }, + { + "bbox": [ + 461, + 260, + 505, + 270 + ], + "score": 0.9, + "content": "2 2 4 \\times 2 2 4", + "type": "inline_equation" + } + ], + "index": 7 + }, + { + "bbox": [ + 106, + 270, + 207, + 283 + ], + "spans": [ + { + "bbox": [ + 106, + 270, + 193, + 283 + ], + "score": 1.0, + "content": "unless specified after", + "type": "text" + }, + { + "bbox": [ + 193, + 271, + 203, + 280 + ], + "score": 0.8, + "content": "@", + "type": "inline_equation" + }, + { + "bbox": [ + 203, + 270, + 207, + 283 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 8 + } + ], + "index": 5.5 + } + ], + "index": 3.25 + }, + { + "type": "text", + "bbox": [ + 107, + 303, + 505, + 337 + ], + "lines": [], + "index": 10, + "bbox_fs": [ + 105, + 304, + 505, + 339 + ], + "lines_deleted": true + }, + { + "type": "text", + "bbox": [ + 107, + 342, + 505, + 441 + ], + "lines": [ + { + "bbox": [ + 106, + 343, + 505, + 354 + ], + "spans": [ + { + "bbox": [ + 106, + 343, + 505, + 354 + ], + "score": 1.0, + "content": "In this paper, we present a new training objective for vision transformers, termed token labeling, that", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 106, + 353, + 505, + 365 + ], + "spans": [ + { + "bbox": [ + 106, + 353, + 440, + 365 + ], + "score": 1.0, + "content": "takes advantage of both the patch tokens and the class tokens. Our method takes a", + "type": "text" + }, + { + "bbox": [ + 441, + 354, + 451, + 363 + ], + "score": 0.82, + "content": "K", + "type": "inline_equation" + }, + { + "bbox": [ + 451, + 353, + 505, + 365 + ], + "score": 1.0, + "content": "-dimensional", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 364, + 506, + 377 + ], + "spans": [ + { + "bbox": [ + 105, + 364, + 506, + 377 + ], + "score": 1.0, + "content": "score map generated by a machine annotator as supervision to supervise all the tokens in a dense", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 105, + 375, + 505, + 388 + ], + "spans": [ + { + "bbox": [ + 105, + 375, + 169, + 388 + ], + "score": 1.0, + "content": "manner, where", + "type": "text" + }, + { + "bbox": [ + 169, + 375, + 180, + 385 + ], + "score": 0.81, + "content": "K", + "type": "inline_equation" + }, + { + "bbox": [ + 180, + 375, + 505, + 388 + ], + "score": 1.0, + "content": "is the number of categories for the target dataset. In this way, each patch token", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 105, + 386, + 506, + 399 + ], + "spans": [ + { + "bbox": [ + 105, + 386, + 506, + 399 + ], + "score": 1.0, + "content": "is explicitly associated with an individual location-specific supervision indicating the existence of", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 397, + 506, + 410 + ], + "spans": [ + { + "bbox": [ + 105, + 397, + 506, + 410 + ], + "score": 1.0, + "content": "the target objects inside the corresponding image patch, so as to improve the object grounding and", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 408, + 506, + 420 + ], + "spans": [ + { + "bbox": [ + 105, + 408, + 506, + 420 + ], + "score": 1.0, + "content": "recognition capabilities of vision transformers with negligible computation overhead. To the best", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 105, + 417, + 506, + 431 + ], + "spans": [ + { + "bbox": [ + 105, + 417, + 506, + 431 + ], + "score": 1.0, + "content": "of our knowledge, this is the first work demonstrating that dense supervision is beneficial to vision", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 105, + 429, + 253, + 442 + ], + "spans": [ + { + "bbox": [ + 105, + 429, + 253, + 442 + ], + "score": 1.0, + "content": "transformers in image classification.", + "type": "text" + } + ], + "index": 20 + } + ], + "index": 16, + "bbox_fs": [ + 105, + 343, + 506, + 442 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 446, + 505, + 523 + ], + "lines": [ + { + "bbox": [ + 105, + 446, + 506, + 458 + ], + "spans": [ + { + "bbox": [ + 105, + 446, + 506, + 458 + ], + "score": 1.0, + "content": "According to our experiments, utilizing the proposed token labeling objective can clearly boost", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 105, + 457, + 505, + 469 + ], + "spans": [ + { + "bbox": [ + 105, + 457, + 505, + 469 + ], + "score": 1.0, + "content": "the performance of vision transformers. As shown in Figure 1, our model, named LV-ViT, with", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 467, + 506, + 480 + ], + "spans": [ + { + "bbox": [ + 105, + 467, + 203, + 480 + ], + "score": 1.0, + "content": "56M parameters, yields", + "type": "text" + }, + { + "bbox": [ + 204, + 468, + 231, + 478 + ], + "score": 0.87, + "content": "8 5 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 231, + 467, + 506, + 480 + ], + "score": 1.0, + "content": "top-1 accuracy on ImageNet [13], behaving better than all the other", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 477, + 505, + 492 + ], + "spans": [ + { + "bbox": [ + 105, + 477, + 505, + 492 + ], + "score": 1.0, + "content": "transformer-based models having no more than 100M parameters. When the model size is scaled up", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 106, + 489, + 504, + 501 + ], + "spans": [ + { + "bbox": [ + 106, + 490, + 289, + 501 + ], + "score": 1.0, + "content": "to 150M, the result can be further improved to", + "type": "text" + }, + { + "bbox": [ + 289, + 489, + 316, + 500 + ], + "score": 0.87, + "content": "8 6 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 316, + 490, + 504, + 501 + ], + "score": 1.0, + "content": ". In addition, we have empirically found that the", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 105, + 500, + 507, + 513 + ], + "spans": [ + { + "bbox": [ + 105, + 500, + 507, + 513 + ], + "score": 1.0, + "content": "pretrained models with token labeling are also beneficial to downstream tasks with dense prediction,", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 105, + 511, + 234, + 524 + ], + "spans": [ + { + "bbox": [ + 105, + 511, + 234, + 524 + ], + "score": 1.0, + "content": "such as semantic segmentation.", + "type": "text" + } + ], + "index": 27 + } + ], + "index": 24, + "bbox_fs": [ + 105, + 446, + 507, + 524 + ] + }, + { + "type": "title", + "bbox": [ + 107, + 538, + 197, + 552 + ], + "lines": [ + { + "bbox": [ + 105, + 537, + 198, + 553 + ], + "spans": [ + { + "bbox": [ + 105, + 537, + 198, + 553 + ], + "score": 1.0, + "content": "2 Related Work", + "type": "text" + } + ], + "index": 28 + } + ], + "index": 28 + }, + { + "type": "text", + "bbox": [ + 107, + 563, + 505, + 641 + ], + "lines": [ + { + "bbox": [ + 105, + 563, + 505, + 576 + ], + "spans": [ + { + "bbox": [ + 105, + 563, + 505, + 576 + ], + "score": 1.0, + "content": "Transformers [39] refer to the models that entirely rely on the self-attention mechanism to build", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 574, + 506, + 587 + ], + "spans": [ + { + "bbox": [ + 105, + 574, + 506, + 587 + ], + "score": 1.0, + "content": "global dependencies, which are originally designed for natural language processing tasks. Due to their", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 585, + 505, + 597 + ], + "spans": [ + { + "bbox": [ + 105, + 585, + 505, + 597 + ], + "score": 1.0, + "content": "strong capability of capturing spatial information, transformers have also been successfully applied", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 596, + 505, + 609 + ], + "spans": [ + { + "bbox": [ + 105, + 596, + 505, + 609 + ], + "score": 1.0, + "content": "to a variety of vision problems, including low-level vision tasks like image enhancement [7, 45], as", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 105, + 607, + 506, + 619 + ], + "spans": [ + { + "bbox": [ + 105, + 607, + 506, + 619 + ], + "score": 1.0, + "content": "well as more challenging tasks such as image classification [9, 15], object detection [5, 11, 55, 61],", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 618, + 506, + 630 + ], + "spans": [ + { + "bbox": [ + 105, + 618, + 506, + 630 + ], + "score": 1.0, + "content": "segmentation [7, 33, 41] and image generation [28]. Some works also extend transformers for video", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 106, + 630, + 282, + 642 + ], + "spans": [ + { + "bbox": [ + 106, + 630, + 282, + 642 + ], + "score": 1.0, + "content": "and 3D point cloud processing [50, 53, 60].", + "type": "text" + } + ], + "index": 35 + } + ], + "index": 32, + "bbox_fs": [ + 105, + 563, + 506, + 642 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 645, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 105, + 645, + 505, + 658 + ], + "spans": [ + { + "bbox": [ + 105, + 645, + 505, + 658 + ], + "score": 1.0, + "content": "Vision Transformer (ViT) is one of the earlier attempts that achieved state-of-the-art performance", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 106, + 656, + 505, + 669 + ], + "spans": [ + { + "bbox": [ + 106, + 656, + 505, + 669 + ], + "score": 1.0, + "content": "on ImageNet classification, using pure transformers as basic building blocks. However, ViTs need", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 105, + 667, + 506, + 680 + ], + "spans": [ + { + "bbox": [ + 105, + 667, + 506, + 680 + ], + "score": 1.0, + "content": "pretraining on very large datasets, such as ImageNet-22k and JFT-300M, and huge computation", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 105, + 678, + 505, + 690 + ], + "spans": [ + { + "bbox": [ + 105, + 678, + 505, + 690 + ], + "score": 1.0, + "content": "resources to achieve comparable performance to ResNet [18] with a similar model size trained on", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 105, + 688, + 505, + 702 + ], + "spans": [ + { + "bbox": [ + 105, + 688, + 505, + 702 + ], + "score": 1.0, + "content": "ImageNet. Later, DeiT [36] manages to tackle the data-inefficiency problem by simply adjusting", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 105, + 700, + 505, + 712 + ], + "spans": [ + { + "bbox": [ + 105, + 700, + 505, + 712 + ], + "score": 1.0, + "content": "the network architecture and adding an additional token along with the class token for Knowledge", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 106, + 710, + 316, + 723 + ], + "spans": [ + { + "bbox": [ + 106, + 710, + 316, + 723 + ], + "score": 1.0, + "content": "Distillation [21, 47] to improve model performance.", + "type": "text" + } + ], + "index": 42 + } + ], + "index": 39, + "bbox_fs": [ + 105, + 645, + 506, + 723 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "image", + "bbox": [ + 128, + 71, + 486, + 252 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 128, + 71, + 486, + 252 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 128, + 71, + 486, + 252 + ], + "spans": [ + { + "bbox": [ + 128, + 71, + 486, + 252 + ], + "score": 0.972, + "type": "image", + "image_path": "77890e011e8cc08c2d9f50310be754cb08f7894ee6cb766ac590a7580f42e616.jpg" + } + ] + } + ], + "index": 1, + "virtual_lines": [ + { + "bbox": [ + 128, + 71, + 486, + 131.33333333333334 + ], + "spans": [], + "index": 0 + }, + { + "bbox": [ + 128, + 131.33333333333334, + 486, + 191.66666666666669 + ], + "spans": [], + "index": 1 + }, + { + "bbox": [ + 128, + 191.66666666666669, + 486, + 252.00000000000003 + ], + "spans": [], + "index": 2 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 106, + 258, + 505, + 336 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 106, + 258, + 505, + 271 + ], + "spans": [ + { + "bbox": [ + 106, + 258, + 505, + 271 + ], + "score": 1.0, + "content": "Figure 2: Pipeline of training vision transformers with token labeling. Other than utilizing the", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 106, + 269, + 505, + 281 + ], + "spans": [ + { + "bbox": [ + 106, + 269, + 505, + 281 + ], + "score": 1.0, + "content": "class token (pink rectangle), we also take advantage of all the output patch tokens (orange rounded", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 280, + 506, + 294 + ], + "spans": [ + { + "bbox": [ + 105, + 280, + 506, + 294 + ], + "score": 1.0, + "content": "rectangle) by assigning each patch token an individual location-specific prediction generated by a", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 106, + 292, + 505, + 303 + ], + "spans": [ + { + "bbox": [ + 106, + 292, + 505, + 303 + ], + "score": 1.0, + "content": "machine annotator [3] as supervision (see the part in the red dash rectangle). Our proposed token", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 105, + 302, + 506, + 315 + ], + "spans": [ + { + "bbox": [ + 105, + 302, + 506, + 315 + ], + "score": 1.0, + "content": "labeling method can be treated as an auxiliary objective to provide each patch token the local details", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 106, + 313, + 505, + 325 + ], + "spans": [ + { + "bbox": [ + 106, + 313, + 505, + 325 + ], + "score": 1.0, + "content": "that aid vision transformers to more accurately locate and recognize the target objects. Note that the", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 105, + 323, + 437, + 336 + ], + "spans": [ + { + "bbox": [ + 105, + 323, + 437, + 336 + ], + "score": 1.0, + "content": "traditional vision transformer training does not include the red dash rectangle part.", + "type": "text" + } + ], + "index": 9 + } + ], + "index": 6 + } + ], + "index": 3.5 + }, + { + "type": "text", + "bbox": [ + 107, + 358, + 505, + 414 + ], + "lines": [ + { + "bbox": [ + 106, + 359, + 505, + 371 + ], + "spans": [ + { + "bbox": [ + 106, + 359, + 505, + 371 + ], + "score": 1.0, + "content": "Some recent works [6, 16, 43, 46] also attempt to introduce the local dependency into vision", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 106, + 370, + 505, + 382 + ], + "spans": [ + { + "bbox": [ + 106, + 370, + 505, + 382 + ], + "score": 1.0, + "content": "transformers by modifying the patch embedding block or the transformer block or both, leading to", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 380, + 505, + 394 + ], + "spans": [ + { + "bbox": [ + 105, + 380, + 505, + 394 + ], + "score": 1.0, + "content": "significant performance gains. Moreover, there are also some works [20, 25, 40] adopting a pyramid", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 392, + 505, + 404 + ], + "spans": [ + { + "bbox": [ + 105, + 392, + 505, + 404 + ], + "score": 1.0, + "content": "structure to reduce the overall computation while maintaining the model’s ability to capture low-level", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 402, + 143, + 414 + ], + "spans": [ + { + "bbox": [ + 105, + 402, + 143, + 414 + ], + "score": 1.0, + "content": "features.", + "type": "text" + } + ], + "index": 14 + } + ], + "index": 12 + }, + { + "type": "text", + "bbox": [ + 107, + 419, + 505, + 474 + ], + "lines": [ + { + "bbox": [ + 106, + 419, + 506, + 430 + ], + "spans": [ + { + "bbox": [ + 106, + 419, + 506, + 430 + ], + "score": 1.0, + "content": "Unlike most aforementioned works that design new transformer blocks or transformer architectures,", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 105, + 430, + 505, + 442 + ], + "spans": [ + { + "bbox": [ + 105, + 430, + 505, + 442 + ], + "score": 1.0, + "content": "we attempt to improve vision transformers by studying the role of patch tokens that embed rich local", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 106, + 441, + 505, + 453 + ], + "spans": [ + { + "bbox": [ + 106, + 441, + 505, + 453 + ], + "score": 1.0, + "content": "information inside image patches. We show that by slightly tuning the structure of vision transformers", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 106, + 452, + 505, + 464 + ], + "spans": [ + { + "bbox": [ + 106, + 452, + 505, + 464 + ], + "score": 1.0, + "content": "and employing the proposed token labeling objective, we can achieve strong baselines for transformer", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 106, + 464, + 256, + 474 + ], + "spans": [ + { + "bbox": [ + 106, + 464, + 256, + 474 + ], + "score": 1.0, + "content": "models at different model size levels.", + "type": "text" + } + ], + "index": 19 + } + ], + "index": 17 + }, + { + "type": "title", + "bbox": [ + 108, + 491, + 249, + 505 + ], + "lines": [ + { + "bbox": [ + 104, + 489, + 249, + 507 + ], + "spans": [ + { + "bbox": [ + 104, + 489, + 249, + 507 + ], + "score": 1.0, + "content": "3 Token Labeling Method", + "type": "text" + } + ], + "index": 20 + } + ], + "index": 20 + }, + { + "type": "text", + "bbox": [ + 107, + 517, + 504, + 540 + ], + "lines": [ + { + "bbox": [ + 105, + 517, + 505, + 530 + ], + "spans": [ + { + "bbox": [ + 105, + 517, + 505, + 530 + ], + "score": 1.0, + "content": "In this section, we first briefly review the structure of the vision transformer [15] and then describe", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 105, + 528, + 301, + 541 + ], + "spans": [ + { + "bbox": [ + 105, + 528, + 301, + 541 + ], + "score": 1.0, + "content": "the proposed training objective—token labeling.", + "type": "text" + } + ], + "index": 22 + } + ], + "index": 21.5 + }, + { + "type": "title", + "bbox": [ + 108, + 554, + 258, + 566 + ], + "lines": [ + { + "bbox": [ + 105, + 554, + 260, + 568 + ], + "spans": [ + { + "bbox": [ + 105, + 554, + 260, + 568 + ], + "score": 1.0, + "content": "3.1 Revisiting Vision Transformer", + "type": "text" + } + ], + "index": 23 + } + ], + "index": 23 + }, + { + "type": "text", + "bbox": [ + 106, + 575, + 505, + 619 + ], + "lines": [ + { + "bbox": [ + 105, + 576, + 505, + 587 + ], + "spans": [ + { + "bbox": [ + 105, + 576, + 505, + 587 + ], + "score": 1.0, + "content": "A typical vision transformer [15] first decomposes a fixed-size input image into a sequence of small", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 105, + 586, + 505, + 598 + ], + "spans": [ + { + "bbox": [ + 105, + 586, + 505, + 598 + ], + "score": 1.0, + "content": "patches. Each small patch is mapped to a feature vector, or called a token, by projection with a linear", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 106, + 597, + 505, + 609 + ], + "spans": [ + { + "bbox": [ + 106, + 597, + 505, + 609 + ], + "score": 1.0, + "content": "layer. Then, all the tokens combined with an additional learnable class token for classification score", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 105, + 608, + 404, + 621 + ], + "spans": [ + { + "bbox": [ + 105, + 608, + 404, + 621 + ], + "score": 1.0, + "content": "prediction are sent into a stack of transformer blocks for feature encoding.", + "type": "text" + } + ], + "index": 27 + } + ], + "index": 25.5 + }, + { + "type": "text", + "bbox": [ + 107, + 624, + 504, + 681 + ], + "lines": [ + { + "bbox": [ + 105, + 624, + 505, + 637 + ], + "spans": [ + { + "bbox": [ + 105, + 624, + 505, + 637 + ], + "score": 1.0, + "content": "In loss computing, the class token from the output tokens of the last transformer block is usually", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 633, + 506, + 649 + ], + "spans": [ + { + "bbox": [ + 105, + 633, + 506, + 649 + ], + "score": 1.0, + "content": "selected and sent into a linear layer for the classification score prediction. Mathematically, given", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 644, + 506, + 659 + ], + "spans": [ + { + "bbox": [ + 105, + 644, + 146, + 659 + ], + "score": 1.0, + "content": "an image", + "type": "text" + }, + { + "bbox": [ + 146, + 647, + 153, + 657 + ], + "score": 0.56, + "content": "I", + "type": "inline_equation" + }, + { + "bbox": [ + 153, + 644, + 360, + 659 + ], + "score": 1.0, + "content": ", denote the output of the last transformer block as", + "type": "text" + }, + { + "bbox": [ + 360, + 646, + 437, + 658 + ], + "score": 0.93, + "content": "[ \\bar { X } ^ { c l s } , X ^ { 1 } , . . . , X ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 438, + 644, + 469, + 659 + ], + "score": 1.0, + "content": ", where", + "type": "text" + }, + { + "bbox": [ + 469, + 647, + 479, + 656 + ], + "score": 0.82, + "content": "N", + "type": "inline_equation" + }, + { + "bbox": [ + 479, + 644, + 506, + 659 + ], + "score": 1.0, + "content": "is the", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 655, + 506, + 672 + ], + "spans": [ + { + "bbox": [ + 105, + 655, + 243, + 672 + ], + "score": 1.0, + "content": "total number of patch tokens, and", + "type": "text" + }, + { + "bbox": [ + 243, + 659, + 263, + 668 + ], + "score": 0.9, + "content": "X ^ { c l s }", + "type": "inline_equation" + }, + { + "bbox": [ + 263, + 655, + 281, + 672 + ], + "score": 1.0, + "content": "and", + "type": "text" + }, + { + "bbox": [ + 282, + 659, + 329, + 670 + ], + "score": 0.92, + "content": "X ^ { 1 } , . . . , X ^ { N }", + "type": "inline_equation" + }, + { + "bbox": [ + 330, + 655, + 506, + 672 + ], + "score": 1.0, + "content": "correspond to the class token and the patch", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 106, + 669, + 399, + 682 + ], + "spans": [ + { + "bbox": [ + 106, + 669, + 321, + 682 + ], + "score": 1.0, + "content": "tokens, respectively. The classification loss for image", + "type": "text" + }, + { + "bbox": [ + 322, + 671, + 328, + 679 + ], + "score": 0.85, + "content": "I", + "type": "inline_equation" + }, + { + "bbox": [ + 328, + 669, + 399, + 682 + ], + "score": 1.0, + "content": "can be written as", + "type": "text" + } + ], + "index": 32 + } + ], + "index": 30 + }, + { + "type": "interline_equation", + "bbox": [ + 260, + 688, + 351, + 703 + ], + "lines": [ + { + "bbox": [ + 260, + 688, + 351, + 703 + ], + "spans": [ + { + "bbox": [ + 260, + 688, + 351, + 703 + ], + "score": 0.92, + "content": "L _ { c l s } = H ( { X } ^ { c l s } , { y } ^ { c l s } ) ,", + "type": "interline_equation", + "image_path": "592dd559a9882895a0bc985fe4c80cb9b1f3f3afab9946d0fd9a87b3424793a9.jpg" + } + ] + } + ], + "index": 33, + "virtual_lines": [ + { + "bbox": [ + 260, + 688, + 351, + 703 + ], + "spans": [], + "index": 33 + } + ] + }, + { + "type": "text", + "bbox": [ + 106, + 710, + 399, + 723 + ], + "lines": [ + { + "bbox": [ + 105, + 709, + 401, + 724 + ], + "spans": [ + { + "bbox": [ + 105, + 709, + 133, + 724 + ], + "score": 1.0, + "content": "where", + "type": "text" + }, + { + "bbox": [ + 133, + 711, + 161, + 723 + ], + "score": 0.93, + "content": "H ( \\cdot , \\cdot )", + "type": "inline_equation" + }, + { + "bbox": [ + 161, + 709, + 312, + 724 + ], + "score": 1.0, + "content": "is the softmax cross-entropy loss and", + "type": "text" + }, + { + "bbox": [ + 313, + 711, + 328, + 722 + ], + "score": 0.92, + "content": "y ^ { c l s }", + "type": "inline_equation" + }, + { + "bbox": [ + 329, + 709, + 401, + 724 + ], + "score": 1.0, + "content": "is the class label.", + "type": "text" + } + ], + "index": 34 + } + ], + "index": 34 + } + ], + "page_idx": 2, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 302, + 741, + 309, + 750 + ], + "lines": [ + { + "bbox": [ + 301, + 740, + 310, + 752 + ], + "spans": [ + { + "bbox": [ + 301, + 740, + 310, + 752 + ], + "score": 1.0, + "content": "3", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "image", + "bbox": [ + 128, + 71, + 486, + 252 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 128, + 71, + 486, + 252 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 128, + 71, + 486, + 252 + ], + "spans": [ + { + "bbox": [ + 128, + 71, + 486, + 252 + ], + "score": 0.972, + "type": "image", + "image_path": "77890e011e8cc08c2d9f50310be754cb08f7894ee6cb766ac590a7580f42e616.jpg" + } + ] + } + ], + "index": 1, + "virtual_lines": [ + { + "bbox": [ + 128, + 71, + 486, + 131.33333333333334 + ], + "spans": [], + "index": 0 + }, + { + "bbox": [ + 128, + 131.33333333333334, + 486, + 191.66666666666669 + ], + "spans": [], + "index": 1 + }, + { + "bbox": [ + 128, + 191.66666666666669, + 486, + 252.00000000000003 + ], + "spans": [], + "index": 2 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 106, + 258, + 505, + 336 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 106, + 258, + 505, + 271 + ], + "spans": [ + { + "bbox": [ + 106, + 258, + 505, + 271 + ], + "score": 1.0, + "content": "Figure 2: Pipeline of training vision transformers with token labeling. Other than utilizing the", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 106, + 269, + 505, + 281 + ], + "spans": [ + { + "bbox": [ + 106, + 269, + 505, + 281 + ], + "score": 1.0, + "content": "class token (pink rectangle), we also take advantage of all the output patch tokens (orange rounded", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 280, + 506, + 294 + ], + "spans": [ + { + "bbox": [ + 105, + 280, + 506, + 294 + ], + "score": 1.0, + "content": "rectangle) by assigning each patch token an individual location-specific prediction generated by a", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 106, + 292, + 505, + 303 + ], + "spans": [ + { + "bbox": [ + 106, + 292, + 505, + 303 + ], + "score": 1.0, + "content": "machine annotator [3] as supervision (see the part in the red dash rectangle). Our proposed token", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 105, + 302, + 506, + 315 + ], + "spans": [ + { + "bbox": [ + 105, + 302, + 506, + 315 + ], + "score": 1.0, + "content": "labeling method can be treated as an auxiliary objective to provide each patch token the local details", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 106, + 313, + 505, + 325 + ], + "spans": [ + { + "bbox": [ + 106, + 313, + 505, + 325 + ], + "score": 1.0, + "content": "that aid vision transformers to more accurately locate and recognize the target objects. Note that the", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 105, + 323, + 437, + 336 + ], + "spans": [ + { + "bbox": [ + 105, + 323, + 437, + 336 + ], + "score": 1.0, + "content": "traditional vision transformer training does not include the red dash rectangle part.", + "type": "text" + } + ], + "index": 9 + } + ], + "index": 6 + } + ], + "index": 3.5 + }, + { + "type": "text", + "bbox": [ + 107, + 358, + 505, + 414 + ], + "lines": [ + { + "bbox": [ + 106, + 359, + 505, + 371 + ], + "spans": [ + { + "bbox": [ + 106, + 359, + 505, + 371 + ], + "score": 1.0, + "content": "Some recent works [6, 16, 43, 46] also attempt to introduce the local dependency into vision", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 106, + 370, + 505, + 382 + ], + "spans": [ + { + "bbox": [ + 106, + 370, + 505, + 382 + ], + "score": 1.0, + "content": "transformers by modifying the patch embedding block or the transformer block or both, leading to", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 380, + 505, + 394 + ], + "spans": [ + { + "bbox": [ + 105, + 380, + 505, + 394 + ], + "score": 1.0, + "content": "significant performance gains. Moreover, there are also some works [20, 25, 40] adopting a pyramid", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 392, + 505, + 404 + ], + "spans": [ + { + "bbox": [ + 105, + 392, + 505, + 404 + ], + "score": 1.0, + "content": "structure to reduce the overall computation while maintaining the model’s ability to capture low-level", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 402, + 143, + 414 + ], + "spans": [ + { + "bbox": [ + 105, + 402, + 143, + 414 + ], + "score": 1.0, + "content": "features.", + "type": "text" + } + ], + "index": 14 + } + ], + "index": 12, + "bbox_fs": [ + 105, + 359, + 505, + 414 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 419, + 505, + 474 + ], + "lines": [ + { + "bbox": [ + 106, + 419, + 506, + 430 + ], + "spans": [ + { + "bbox": [ + 106, + 419, + 506, + 430 + ], + "score": 1.0, + "content": "Unlike most aforementioned works that design new transformer blocks or transformer architectures,", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 105, + 430, + 505, + 442 + ], + "spans": [ + { + "bbox": [ + 105, + 430, + 505, + 442 + ], + "score": 1.0, + "content": "we attempt to improve vision transformers by studying the role of patch tokens that embed rich local", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 106, + 441, + 505, + 453 + ], + "spans": [ + { + "bbox": [ + 106, + 441, + 505, + 453 + ], + "score": 1.0, + "content": "information inside image patches. We show that by slightly tuning the structure of vision transformers", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 106, + 452, + 505, + 464 + ], + "spans": [ + { + "bbox": [ + 106, + 452, + 505, + 464 + ], + "score": 1.0, + "content": "and employing the proposed token labeling objective, we can achieve strong baselines for transformer", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 106, + 464, + 256, + 474 + ], + "spans": [ + { + "bbox": [ + 106, + 464, + 256, + 474 + ], + "score": 1.0, + "content": "models at different model size levels.", + "type": "text" + } + ], + "index": 19 + } + ], + "index": 17, + "bbox_fs": [ + 105, + 419, + 506, + 474 + ] + }, + { + "type": "title", + "bbox": [ + 108, + 491, + 249, + 505 + ], + "lines": [ + { + "bbox": [ + 104, + 489, + 249, + 507 + ], + "spans": [ + { + "bbox": [ + 104, + 489, + 249, + 507 + ], + "score": 1.0, + "content": "3 Token Labeling Method", + "type": "text" + } + ], + "index": 20 + } + ], + "index": 20 + }, + { + "type": "text", + "bbox": [ + 107, + 517, + 504, + 540 + ], + "lines": [ + { + "bbox": [ + 105, + 517, + 505, + 530 + ], + "spans": [ + { + "bbox": [ + 105, + 517, + 505, + 530 + ], + "score": 1.0, + "content": "In this section, we first briefly review the structure of the vision transformer [15] and then describe", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 105, + 528, + 301, + 541 + ], + "spans": [ + { + "bbox": [ + 105, + 528, + 301, + 541 + ], + "score": 1.0, + "content": "the proposed training objective—token labeling.", + "type": "text" + } + ], + "index": 22 + } + ], + "index": 21.5, + "bbox_fs": [ + 105, + 517, + 505, + 541 + ] + }, + { + "type": "title", + "bbox": [ + 108, + 554, + 258, + 566 + ], + "lines": [ + { + "bbox": [ + 105, + 554, + 260, + 568 + ], + "spans": [ + { + "bbox": [ + 105, + 554, + 260, + 568 + ], + "score": 1.0, + "content": "3.1 Revisiting Vision Transformer", + "type": "text" + } + ], + "index": 23 + } + ], + "index": 23 + }, + { + "type": "text", + "bbox": [ + 106, + 575, + 505, + 619 + ], + "lines": [ + { + "bbox": [ + 105, + 576, + 505, + 587 + ], + "spans": [ + { + "bbox": [ + 105, + 576, + 505, + 587 + ], + "score": 1.0, + "content": "A typical vision transformer [15] first decomposes a fixed-size input image into a sequence of small", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 105, + 586, + 505, + 598 + ], + "spans": [ + { + "bbox": [ + 105, + 586, + 505, + 598 + ], + "score": 1.0, + "content": "patches. Each small patch is mapped to a feature vector, or called a token, by projection with a linear", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 106, + 597, + 505, + 609 + ], + "spans": [ + { + "bbox": [ + 106, + 597, + 505, + 609 + ], + "score": 1.0, + "content": "layer. Then, all the tokens combined with an additional learnable class token for classification score", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 105, + 608, + 404, + 621 + ], + "spans": [ + { + "bbox": [ + 105, + 608, + 404, + 621 + ], + "score": 1.0, + "content": "prediction are sent into a stack of transformer blocks for feature encoding.", + "type": "text" + } + ], + "index": 27 + } + ], + "index": 25.5, + "bbox_fs": [ + 105, + 576, + 505, + 621 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 624, + 504, + 681 + ], + "lines": [ + { + "bbox": [ + 105, + 624, + 505, + 637 + ], + "spans": [ + { + "bbox": [ + 105, + 624, + 505, + 637 + ], + "score": 1.0, + "content": "In loss computing, the class token from the output tokens of the last transformer block is usually", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 633, + 506, + 649 + ], + "spans": [ + { + "bbox": [ + 105, + 633, + 506, + 649 + ], + "score": 1.0, + "content": "selected and sent into a linear layer for the classification score prediction. Mathematically, given", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 644, + 506, + 659 + ], + "spans": [ + { + "bbox": [ + 105, + 644, + 146, + 659 + ], + "score": 1.0, + "content": "an image", + "type": "text" + }, + { + "bbox": [ + 146, + 647, + 153, + 657 + ], + "score": 0.56, + "content": "I", + "type": "inline_equation" + }, + { + "bbox": [ + 153, + 644, + 360, + 659 + ], + "score": 1.0, + "content": ", denote the output of the last transformer block as", + "type": "text" + }, + { + "bbox": [ + 360, + 646, + 437, + 658 + ], + "score": 0.93, + "content": "[ \\bar { X } ^ { c l s } , X ^ { 1 } , . . . , X ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 438, + 644, + 469, + 659 + ], + "score": 1.0, + "content": ", where", + "type": "text" + }, + { + "bbox": [ + 469, + 647, + 479, + 656 + ], + "score": 0.82, + "content": "N", + "type": "inline_equation" + }, + { + "bbox": [ + 479, + 644, + 506, + 659 + ], + "score": 1.0, + "content": "is the", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 655, + 506, + 672 + ], + "spans": [ + { + "bbox": [ + 105, + 655, + 243, + 672 + ], + "score": 1.0, + "content": "total number of patch tokens, and", + "type": "text" + }, + { + "bbox": [ + 243, + 659, + 263, + 668 + ], + "score": 0.9, + "content": "X ^ { c l s }", + "type": "inline_equation" + }, + { + "bbox": [ + 263, + 655, + 281, + 672 + ], + "score": 1.0, + "content": "and", + "type": "text" + }, + { + "bbox": [ + 282, + 659, + 329, + 670 + ], + "score": 0.92, + "content": "X ^ { 1 } , . . . , X ^ { N }", + "type": "inline_equation" + }, + { + "bbox": [ + 330, + 655, + 506, + 672 + ], + "score": 1.0, + "content": "correspond to the class token and the patch", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 106, + 669, + 399, + 682 + ], + "spans": [ + { + "bbox": [ + 106, + 669, + 321, + 682 + ], + "score": 1.0, + "content": "tokens, respectively. The classification loss for image", + "type": "text" + }, + { + "bbox": [ + 322, + 671, + 328, + 679 + ], + "score": 0.85, + "content": "I", + "type": "inline_equation" + }, + { + "bbox": [ + 328, + 669, + 399, + 682 + ], + "score": 1.0, + "content": "can be written as", + "type": "text" + } + ], + "index": 32 + } + ], + "index": 30, + "bbox_fs": [ + 105, + 624, + 506, + 682 + ] + }, + { + "type": "interline_equation", + "bbox": [ + 260, + 688, + 351, + 703 + ], + "lines": [ + { + "bbox": [ + 260, + 688, + 351, + 703 + ], + "spans": [ + { + "bbox": [ + 260, + 688, + 351, + 703 + ], + "score": 0.92, + "content": "L _ { c l s } = H ( { X } ^ { c l s } , { y } ^ { c l s } ) ,", + "type": "interline_equation", + "image_path": "592dd559a9882895a0bc985fe4c80cb9b1f3f3afab9946d0fd9a87b3424793a9.jpg" + } + ] + } + ], + "index": 33, + "virtual_lines": [ + { + "bbox": [ + 260, + 688, + 351, + 703 + ], + "spans": [], + "index": 33 + } + ] + }, + { + "type": "text", + "bbox": [ + 106, + 710, + 399, + 723 + ], + "lines": [ + { + "bbox": [ + 105, + 709, + 401, + 724 + ], + "spans": [ + { + "bbox": [ + 105, + 709, + 133, + 724 + ], + "score": 1.0, + "content": "where", + "type": "text" + }, + { + "bbox": [ + 133, + 711, + 161, + 723 + ], + "score": 0.93, + "content": "H ( \\cdot , \\cdot )", + "type": "inline_equation" + }, + { + "bbox": [ + 161, + 709, + 312, + 724 + ], + "score": 1.0, + "content": "is the softmax cross-entropy loss and", + "type": "text" + }, + { + "bbox": [ + 313, + 711, + 328, + 722 + ], + "score": 0.92, + "content": "y ^ { c l s }", + "type": "inline_equation" + }, + { + "bbox": [ + 329, + 709, + 401, + 724 + ], + "score": 1.0, + "content": "is the class label.", + "type": "text" + } + ], + "index": 34 + } + ], + "index": 34, + "bbox_fs": [ + 105, + 709, + 401, + 724 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "image", + "bbox": [ + 121, + 72, + 300, + 192 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 121, + 72, + 300, + 192 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 121, + 72, + 300, + 192 + ], + "spans": [ + { + "bbox": [ + 121, + 72, + 300, + 192 + ], + "score": 0.949, + "type": "image", + "image_path": "c741b481df54f67e411d3cc1b30dc7deb904c80d9d11e57036dbf5b27cdbfb8f.jpg" + } + ] + } + ], + "index": 4, + "virtual_lines": [ + { + "bbox": [ + 121, + 72, + 300, + 85.33333333333333 + ], + "spans": [], + "index": 0 + }, + { + "bbox": [ + 121, + 85.33333333333333, + 300, + 98.66666666666666 + ], + "spans": [], + "index": 1 + }, + { + "bbox": [ + 121, + 98.66666666666666, + 300, + 111.99999999999999 + ], + "spans": [], + "index": 2 + }, + { + "bbox": [ + 121, + 111.99999999999999, + 300, + 125.33333333333331 + ], + "spans": [], + "index": 3 + }, + { + "bbox": [ + 121, + 125.33333333333331, + 300, + 138.66666666666666 + ], + "spans": [], + "index": 4 + }, + { + "bbox": [ + 121, + 138.66666666666666, + 300, + 152.0 + ], + "spans": [], + "index": 5 + }, + { + "bbox": [ + 121, + 152.0, + 300, + 165.33333333333334 + ], + "spans": [], + "index": 6 + }, + { + "bbox": [ + 121, + 165.33333333333334, + 300, + 178.66666666666669 + ], + "spans": [], + "index": 7 + }, + { + "bbox": [ + 121, + 178.66666666666669, + 300, + 192.00000000000003 + ], + "spans": [], + "index": 8 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 309, + 70, + 491, + 191 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 309, + 70, + 490, + 82 + ], + "spans": [ + { + "bbox": [ + 309, + 70, + 490, + 82 + ], + "score": 1.0, + "content": "Figure 3: Comparison between CutMix [48]", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 309, + 81, + 492, + 93 + ], + "spans": [ + { + "bbox": [ + 309, + 81, + 492, + 93 + ], + "score": 1.0, + "content": "(Left) and our proposed MixToken (Right).", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 309, + 92, + 491, + 104 + ], + "spans": [ + { + "bbox": [ + 309, + 92, + 491, + 104 + ], + "score": 1.0, + "content": "CutMix is operated on the input images. This", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 309, + 104, + 491, + 114 + ], + "spans": [ + { + "bbox": [ + 309, + 104, + 491, + 114 + ], + "score": 1.0, + "content": "results in patches containing mixed regions", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 309, + 114, + 491, + 126 + ], + "spans": [ + { + "bbox": [ + 309, + 114, + 491, + 126 + ], + "score": 1.0, + "content": "from the two images (see the patches enclosed", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 309, + 124, + 492, + 136 + ], + "spans": [ + { + "bbox": [ + 309, + 124, + 492, + 136 + ], + "score": 1.0, + "content": "by red bounding boxes). Differently, MixTo-", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 309, + 136, + 491, + 147 + ], + "spans": [ + { + "bbox": [ + 309, + 136, + 491, + 147 + ], + "score": 1.0, + "content": "ken targets at mixing tokens after patch em-", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 309, + 146, + 491, + 158 + ], + "spans": [ + { + "bbox": [ + 309, + 146, + 491, + 158 + ], + "score": 1.0, + "content": "bedding. This enables each token after patch", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 309, + 158, + 491, + 169 + ], + "spans": [ + { + "bbox": [ + 309, + 158, + 491, + 169 + ], + "score": 1.0, + "content": "embedding to have clean content as shown in", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 309, + 168, + 492, + 181 + ], + "spans": [ + { + "bbox": [ + 309, + 168, + 492, + 181 + ], + "score": 1.0, + "content": "the right part of this figure. The detailed ad-", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 309, + 179, + 491, + 191 + ], + "spans": [ + { + "bbox": [ + 309, + 179, + 491, + 191 + ], + "score": 1.0, + "content": "vantage of MixToken can be found in Sec. 4.2.", + "type": "text" + } + ], + "index": 19 + } + ], + "index": 14 + } + ], + "index": 9.0 + }, + { + "type": "title", + "bbox": [ + 107, + 218, + 197, + 231 + ], + "lines": [ + { + "bbox": [ + 105, + 216, + 198, + 233 + ], + "spans": [ + { + "bbox": [ + 105, + 216, + 198, + 233 + ], + "score": 1.0, + "content": "3.2 Token Labeling", + "type": "text" + } + ], + "index": 20 + } + ], + "index": 20 + }, + { + "type": "text", + "bbox": [ + 107, + 241, + 505, + 285 + ], + "lines": [ + { + "bbox": [ + 105, + 241, + 505, + 254 + ], + "spans": [ + { + "bbox": [ + 105, + 241, + 505, + 254 + ], + "score": 1.0, + "content": "The above classification problem only adopts an image-level label as supervision whereas it neglects", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 105, + 251, + 505, + 266 + ], + "spans": [ + { + "bbox": [ + 105, + 251, + 505, + 266 + ], + "score": 1.0, + "content": "the rich information embedded in each image patch. In this subsection, we present a new training", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 264, + 505, + 277 + ], + "spans": [ + { + "bbox": [ + 105, + 264, + 505, + 277 + ], + "score": 1.0, + "content": "objective—token labeling—that takes advantage of the complementary information between the", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 275, + 243, + 286 + ], + "spans": [ + { + "bbox": [ + 105, + 275, + 243, + 286 + ], + "score": 1.0, + "content": "patch tokens and the class tokens.", + "type": "text" + } + ], + "index": 24 + } + ], + "index": 22.5 + }, + { + "type": "text", + "bbox": [ + 106, + 290, + 505, + 368 + ], + "lines": [ + { + "bbox": [ + 106, + 290, + 505, + 303 + ], + "spans": [ + { + "bbox": [ + 106, + 290, + 505, + 303 + ], + "score": 1.0, + "content": "Token Labeling: Different from the classification loss as formulated in Eqn. (1) that measures the", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 105, + 300, + 505, + 315 + ], + "spans": [ + { + "bbox": [ + 105, + 300, + 505, + 315 + ], + "score": 1.0, + "content": "distance between the single class token (representing the whole input image) and the corresponding", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 106, + 312, + 506, + 325 + ], + "spans": [ + { + "bbox": [ + 106, + 312, + 506, + 325 + ], + "score": 1.0, + "content": "image-level label, token labeling emphasizes the importance of all output tokens and advocates that", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 106, + 324, + 505, + 335 + ], + "spans": [ + { + "bbox": [ + 106, + 324, + 505, + 335 + ], + "score": 1.0, + "content": "each output token should be associated with an individual location-specific label. Therefore, in our", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 334, + 505, + 346 + ], + "spans": [ + { + "bbox": [ + 105, + 334, + 383, + 346 + ], + "score": 1.0, + "content": "method, the ground truth for an input image involves not only a single", + "type": "text" + }, + { + "bbox": [ + 383, + 335, + 393, + 345 + ], + "score": 0.82, + "content": "K", + "type": "inline_equation" + }, + { + "bbox": [ + 393, + 334, + 472, + 346 + ], + "score": 1.0, + "content": "-dimensional vector", + "type": "text" + }, + { + "bbox": [ + 472, + 334, + 489, + 346 + ], + "score": 0.92, + "content": "y ^ { c l s }", + "type": "inline_equation" + }, + { + "bbox": [ + 489, + 334, + 505, + 346 + ], + "score": 1.0, + "content": "but", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 104, + 344, + 504, + 359 + ], + "spans": [ + { + "bbox": [ + 104, + 344, + 132, + 359 + ], + "score": 1.0, + "content": "also a", + "type": "text" + }, + { + "bbox": [ + 132, + 346, + 164, + 356 + ], + "score": 0.88, + "content": "K \\times N", + "type": "inline_equation" + }, + { + "bbox": [ + 164, + 344, + 237, + 359 + ], + "score": 1.0, + "content": "matrix or called a", + "type": "text" + }, + { + "bbox": [ + 238, + 346, + 248, + 355 + ], + "score": 0.84, + "content": "K", + "type": "inline_equation" + }, + { + "bbox": [ + 248, + 344, + 416, + 359 + ], + "score": 1.0, + "content": "-dimensional score map as represented by", + "type": "text" + }, + { + "bbox": [ + 417, + 345, + 462, + 357 + ], + "score": 0.94, + "content": "[ y ^ { 1 } , . . . , y ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 463, + 344, + 493, + 359 + ], + "score": 1.0, + "content": ", where", + "type": "text" + }, + { + "bbox": [ + 494, + 346, + 504, + 355 + ], + "score": 0.82, + "content": "N", + "type": "inline_equation" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 356, + 272, + 369 + ], + "spans": [ + { + "bbox": [ + 105, + 356, + 272, + 369 + ], + "score": 1.0, + "content": "is the number of the output patch tokens.", + "type": "text" + } + ], + "index": 31 + } + ], + "index": 28 + }, + { + "type": "text", + "bbox": [ + 107, + 372, + 505, + 428 + ], + "lines": [ + { + "bbox": [ + 106, + 372, + 505, + 385 + ], + "spans": [ + { + "bbox": [ + 106, + 372, + 505, + 385 + ], + "score": 1.0, + "content": "Specifically, we leverage a dense score map for each training image and use the cross-entropy loss", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 106, + 384, + 505, + 396 + ], + "spans": [ + { + "bbox": [ + 106, + 384, + 505, + 396 + ], + "score": 1.0, + "content": "between each output patch token and the corresponding aligned label in the dense score map as an", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 106, + 394, + 505, + 407 + ], + "spans": [ + { + "bbox": [ + 106, + 394, + 505, + 407 + ], + "score": 1.0, + "content": "auxiliary loss at the training phase. Figure 2 provides an intuitive interpretation. Given the output", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 105, + 403, + 506, + 419 + ], + "spans": [ + { + "bbox": [ + 105, + 403, + 159, + 419 + ], + "score": 1.0, + "content": "patch tokens", + "type": "text" + }, + { + "bbox": [ + 159, + 405, + 207, + 417 + ], + "score": 0.93, + "content": "X ^ { 1 } , . . . , X ^ { N }", + "type": "inline_equation" + }, + { + "bbox": [ + 208, + 403, + 326, + 419 + ], + "score": 1.0, + "content": "and the corresponding labels", + "type": "text" + }, + { + "bbox": [ + 326, + 405, + 372, + 417 + ], + "score": 0.94, + "content": "[ y ^ { 1 } , . . . , y ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 372, + 403, + 506, + 419 + ], + "score": 1.0, + "content": ", the token labeling objective can", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 106, + 416, + 161, + 428 + ], + "spans": [ + { + "bbox": [ + 106, + 416, + 161, + 428 + ], + "score": 1.0, + "content": "be defined as", + "type": "text" + } + ], + "index": 36 + } + ], + "index": 34 + }, + { + "type": "interline_equation", + "bbox": [ + 254, + 433, + 357, + 468 + ], + "lines": [ + { + "bbox": [ + 254, + 433, + 357, + 468 + ], + "spans": [ + { + "bbox": [ + 254, + 433, + 357, + 468 + ], + "score": 0.95, + "content": "L _ { t l } = \\frac { 1 } { N } \\sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) .", + "type": "interline_equation", + "image_path": "a9d480948e81c19661ff7ef1b5daada45648302552f0f35ccc9e1a4855e07862.jpg" + } + ] + } + ], + "index": 37.5, + "virtual_lines": [ + { + "bbox": [ + 254, + 433, + 357, + 450.5 + ], + "spans": [], + "index": 37 + }, + { + "bbox": [ + 254, + 450.5, + 357, + 468.0 + ], + "spans": [], + "index": 38 + } + ] + }, + { + "type": "text", + "bbox": [ + 105, + 478, + 468, + 491 + ], + "lines": [ + { + "bbox": [ + 106, + 478, + 468, + 491 + ], + "spans": [ + { + "bbox": [ + 106, + 478, + 152, + 491 + ], + "score": 1.0, + "content": "Recall that", + "type": "text" + }, + { + "bbox": [ + 152, + 480, + 162, + 489 + ], + "score": 0.84, + "content": "H", + "type": "inline_equation" + }, + { + "bbox": [ + 163, + 478, + 468, + 491 + ], + "score": 1.0, + "content": "is the cross-entropy loss. Therefore, the total loss function can be written as", + "type": "text" + } + ], + "index": 39 + } + ], + "index": 39 + }, + { + "type": "interline_equation", + "bbox": [ + 207, + 501, + 403, + 553 + ], + "lines": [ + { + "bbox": [ + 207, + 501, + 403, + 553 + ], + "spans": [ + { + "bbox": [ + 207, + 501, + 403, + 553 + ], + "score": 0.94, + "content": "\\begin{array} { l } { { { \\cal L } _ { t o t a l } = H ( X ^ { c l s } , y ^ { c l s } ) + \\beta \\cdot L _ { t l } , } } \\\\ { { { } } } \\\\ { { = H ( X ^ { c l s } , y ^ { c l s } ) + \\beta \\cdot \\displaystyle \\frac { 1 } { N } \\sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) , } } \\end{array}", + "type": "interline_equation", + "image_path": "c5da8361fd2faf0126b83503cc2cc4f9798363630c4f410a80eef81deafa2915.jpg" + } + ] + } + ], + "index": 41.5, + "virtual_lines": [ + { + "bbox": [ + 207, + 501, + 403, + 514.0 + ], + "spans": [], + "index": 40 + }, + { + "bbox": [ + 207, + 514.0, + 403, + 527.0 + ], + "spans": [], + "index": 41 + }, + { + "bbox": [ + 207, + 527.0, + 403, + 540.0 + ], + "spans": [], + "index": 42 + }, + { + "bbox": [ + 207, + 540.0, + 403, + 553.0 + ], + "spans": [], + "index": 43 + } + ] + }, + { + "type": "text", + "bbox": [ + 107, + 563, + 505, + 585 + ], + "lines": [ + { + "bbox": [ + 105, + 561, + 505, + 577 + ], + "spans": [ + { + "bbox": [ + 105, + 561, + 133, + 577 + ], + "score": 1.0, + "content": "where", + "type": "text" + }, + { + "bbox": [ + 133, + 564, + 141, + 575 + ], + "score": 0.85, + "content": "\\beta", + "type": "inline_equation" + }, + { + "bbox": [ + 141, + 561, + 505, + 577 + ], + "score": 1.0, + "content": "is a hyper-parameter to balance the two terms. In our experiment, we empirically set it to", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 104, + 572, + 126, + 587 + ], + "spans": [ + { + "bbox": [ + 104, + 572, + 126, + 587 + ], + "score": 1.0, + "content": "0.5.", + "type": "text" + } + ], + "index": 45 + } + ], + "index": 44.5 + }, + { + "type": "text", + "bbox": [ + 106, + 590, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 104, + 590, + 505, + 604 + ], + "spans": [ + { + "bbox": [ + 104, + 590, + 505, + 604 + ], + "score": 1.0, + "content": "Advantages: Our token labeling offers the following advantages. First of all, unlike knowledge", + "type": "text" + } + ], + "index": 46 + }, + { + "bbox": [ + 105, + 601, + 506, + 615 + ], + "spans": [ + { + "bbox": [ + 105, + 601, + 506, + 615 + ], + "score": 1.0, + "content": "distillation methods that require a teacher model to generate supervision labels online, token labeling", + "type": "text" + } + ], + "index": 47 + }, + { + "bbox": [ + 104, + 612, + 506, + 626 + ], + "spans": [ + { + "bbox": [ + 104, + 612, + 506, + 626 + ], + "score": 1.0, + "content": "is a cheap operation. The dense score map can be generated by a pretrained model in advance (e.g.,", + "type": "text" + } + ], + "index": 48 + }, + { + "bbox": [ + 106, + 624, + 505, + 636 + ], + "spans": [ + { + "bbox": [ + 106, + 624, + 505, + 636 + ], + "score": 1.0, + "content": "EfficientNet [34] or NFNet [3]). During training, we only need to crop the score map and perform", + "type": "text" + } + ], + "index": 49 + }, + { + "bbox": [ + 105, + 634, + 506, + 647 + ], + "spans": [ + { + "bbox": [ + 105, + 634, + 506, + 647 + ], + "score": 1.0, + "content": "interpolation to make it aligned with the cropped image in the spatial coordinate. Thus, the additional", + "type": "text" + } + ], + "index": 50 + }, + { + "bbox": [ + 105, + 645, + 506, + 658 + ], + "spans": [ + { + "bbox": [ + 105, + 645, + 506, + 658 + ], + "score": 1.0, + "content": "computations are negligible. Second, rather than utilizing a single label vector as supervision as done", + "type": "text" + } + ], + "index": 51 + }, + { + "bbox": [ + 105, + 656, + 505, + 669 + ], + "spans": [ + { + "bbox": [ + 105, + 656, + 505, + 669 + ], + "score": 1.0, + "content": "in most classification models and the ReLabel strategy [49], we also harness score maps to supervise", + "type": "text" + } + ], + "index": 52 + }, + { + "bbox": [ + 106, + 667, + 505, + 679 + ], + "spans": [ + { + "bbox": [ + 106, + 667, + 505, + 679 + ], + "score": 1.0, + "content": "the models in a dense manner and thereby the label for each patch token provides location-specific", + "type": "text" + } + ], + "index": 53 + }, + { + "bbox": [ + 106, + 678, + 505, + 690 + ], + "spans": [ + { + "bbox": [ + 106, + 678, + 505, + 690 + ], + "score": 1.0, + "content": "information, which can aid the training models to easily discover the target objects and improve the", + "type": "text" + } + ], + "index": 54 + }, + { + "bbox": [ + 105, + 689, + 505, + 701 + ], + "spans": [ + { + "bbox": [ + 105, + 689, + 505, + 701 + ], + "score": 1.0, + "content": "recognition accuracy. Last but not the least, as dense supervision is adopted in training, we found", + "type": "text" + } + ], + "index": 55 + }, + { + "bbox": [ + 106, + 700, + 505, + 713 + ], + "spans": [ + { + "bbox": [ + 106, + 700, + 505, + 713 + ], + "score": 1.0, + "content": "that the pretrained models with token labeling benefit downstream tasks with dense prediction, like", + "type": "text" + } + ], + "index": 56 + }, + { + "bbox": [ + 106, + 712, + 201, + 722 + ], + "spans": [ + { + "bbox": [ + 106, + 712, + 201, + 722 + ], + "score": 1.0, + "content": "semantic segmentation.", + "type": "text" + } + ], + "index": 57 + } + ], + "index": 51.5 + } + ], + "page_idx": 3, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 302, + 742, + 308, + 750 + ], + "lines": [ + { + "bbox": [ + 301, + 741, + 310, + 752 + ], + "spans": [ + { + "bbox": [ + 301, + 741, + 310, + 752 + ], + "score": 1.0, + "content": "", + "type": "text", + "height": 11, + "width": 9 + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "image", + "bbox": [ + 121, + 72, + 300, + 192 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 121, + 72, + 300, + 192 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 121, + 72, + 300, + 192 + ], + "spans": [ + { + "bbox": [ + 121, + 72, + 300, + 192 + ], + "score": 0.949, + "type": "image", + "image_path": "c741b481df54f67e411d3cc1b30dc7deb904c80d9d11e57036dbf5b27cdbfb8f.jpg" + } + ] + } + ], + "index": 4, + "virtual_lines": [ + { + "bbox": [ + 121, + 72, + 300, + 85.33333333333333 + ], + "spans": [], + "index": 0 + }, + { + "bbox": [ + 121, + 85.33333333333333, + 300, + 98.66666666666666 + ], + "spans": [], + "index": 1 + }, + { + "bbox": [ + 121, + 98.66666666666666, + 300, + 111.99999999999999 + ], + "spans": [], + "index": 2 + }, + { + "bbox": [ + 121, + 111.99999999999999, + 300, + 125.33333333333331 + ], + "spans": [], + "index": 3 + }, + { + "bbox": [ + 121, + 125.33333333333331, + 300, + 138.66666666666666 + ], + "spans": [], + "index": 4 + }, + { + "bbox": [ + 121, + 138.66666666666666, + 300, + 152.0 + ], + "spans": [], + "index": 5 + }, + { + "bbox": [ + 121, + 152.0, + 300, + 165.33333333333334 + ], + "spans": [], + "index": 6 + }, + { + "bbox": [ + 121, + 165.33333333333334, + 300, + 178.66666666666669 + ], + "spans": [], + "index": 7 + }, + { + "bbox": [ + 121, + 178.66666666666669, + 300, + 192.00000000000003 + ], + "spans": [], + "index": 8 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 309, + 70, + 491, + 191 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 309, + 70, + 490, + 82 + ], + "spans": [ + { + "bbox": [ + 309, + 70, + 490, + 82 + ], + "score": 1.0, + "content": "Figure 3: Comparison between CutMix [48]", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 309, + 81, + 492, + 93 + ], + "spans": [ + { + "bbox": [ + 309, + 81, + 492, + 93 + ], + "score": 1.0, + "content": "(Left) and our proposed MixToken (Right).", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 309, + 92, + 491, + 104 + ], + "spans": [ + { + "bbox": [ + 309, + 92, + 491, + 104 + ], + "score": 1.0, + "content": "CutMix is operated on the input images. This", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 309, + 104, + 491, + 114 + ], + "spans": [ + { + "bbox": [ + 309, + 104, + 491, + 114 + ], + "score": 1.0, + "content": "results in patches containing mixed regions", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 309, + 114, + 491, + 126 + ], + "spans": [ + { + "bbox": [ + 309, + 114, + 491, + 126 + ], + "score": 1.0, + "content": "from the two images (see the patches enclosed", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 309, + 124, + 492, + 136 + ], + "spans": [ + { + "bbox": [ + 309, + 124, + 492, + 136 + ], + "score": 1.0, + "content": "by red bounding boxes). Differently, MixTo-", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 309, + 136, + 491, + 147 + ], + "spans": [ + { + "bbox": [ + 309, + 136, + 491, + 147 + ], + "score": 1.0, + "content": "ken targets at mixing tokens after patch em-", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 309, + 146, + 491, + 158 + ], + "spans": [ + { + "bbox": [ + 309, + 146, + 491, + 158 + ], + "score": 1.0, + "content": "bedding. This enables each token after patch", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 309, + 158, + 491, + 169 + ], + "spans": [ + { + "bbox": [ + 309, + 158, + 491, + 169 + ], + "score": 1.0, + "content": "embedding to have clean content as shown in", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 309, + 168, + 492, + 181 + ], + "spans": [ + { + "bbox": [ + 309, + 168, + 492, + 181 + ], + "score": 1.0, + "content": "the right part of this figure. The detailed ad-", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 309, + 179, + 491, + 191 + ], + "spans": [ + { + "bbox": [ + 309, + 179, + 491, + 191 + ], + "score": 1.0, + "content": "vantage of MixToken can be found in Sec. 4.2.", + "type": "text" + } + ], + "index": 19 + } + ], + "index": 14 + } + ], + "index": 9.0 + }, + { + "type": "title", + "bbox": [ + 107, + 218, + 197, + 231 + ], + "lines": [ + { + "bbox": [ + 105, + 216, + 198, + 233 + ], + "spans": [ + { + "bbox": [ + 105, + 216, + 198, + 233 + ], + "score": 1.0, + "content": "3.2 Token Labeling", + "type": "text" + } + ], + "index": 20 + } + ], + "index": 20 + }, + { + "type": "text", + "bbox": [ + 107, + 241, + 505, + 285 + ], + "lines": [ + { + "bbox": [ + 105, + 241, + 505, + 254 + ], + "spans": [ + { + "bbox": [ + 105, + 241, + 505, + 254 + ], + "score": 1.0, + "content": "The above classification problem only adopts an image-level label as supervision whereas it neglects", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 105, + 251, + 505, + 266 + ], + "spans": [ + { + "bbox": [ + 105, + 251, + 505, + 266 + ], + "score": 1.0, + "content": "the rich information embedded in each image patch. In this subsection, we present a new training", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 264, + 505, + 277 + ], + "spans": [ + { + "bbox": [ + 105, + 264, + 505, + 277 + ], + "score": 1.0, + "content": "objective—token labeling—that takes advantage of the complementary information between the", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 275, + 243, + 286 + ], + "spans": [ + { + "bbox": [ + 105, + 275, + 243, + 286 + ], + "score": 1.0, + "content": "patch tokens and the class tokens.", + "type": "text" + } + ], + "index": 24 + } + ], + "index": 22.5, + "bbox_fs": [ + 105, + 241, + 505, + 286 + ] + }, + { + "type": "text", + "bbox": [ + 106, + 290, + 505, + 368 + ], + "lines": [ + { + "bbox": [ + 106, + 290, + 505, + 303 + ], + "spans": [ + { + "bbox": [ + 106, + 290, + 505, + 303 + ], + "score": 1.0, + "content": "Token Labeling: Different from the classification loss as formulated in Eqn. (1) that measures the", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 105, + 300, + 505, + 315 + ], + "spans": [ + { + "bbox": [ + 105, + 300, + 505, + 315 + ], + "score": 1.0, + "content": "distance between the single class token (representing the whole input image) and the corresponding", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 106, + 312, + 506, + 325 + ], + "spans": [ + { + "bbox": [ + 106, + 312, + 506, + 325 + ], + "score": 1.0, + "content": "image-level label, token labeling emphasizes the importance of all output tokens and advocates that", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 106, + 324, + 505, + 335 + ], + "spans": [ + { + "bbox": [ + 106, + 324, + 505, + 335 + ], + "score": 1.0, + "content": "each output token should be associated with an individual location-specific label. Therefore, in our", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 334, + 505, + 346 + ], + "spans": [ + { + "bbox": [ + 105, + 334, + 383, + 346 + ], + "score": 1.0, + "content": "method, the ground truth for an input image involves not only a single", + "type": "text" + }, + { + "bbox": [ + 383, + 335, + 393, + 345 + ], + "score": 0.82, + "content": "K", + "type": "inline_equation" + }, + { + "bbox": [ + 393, + 334, + 472, + 346 + ], + "score": 1.0, + "content": "-dimensional vector", + "type": "text" + }, + { + "bbox": [ + 472, + 334, + 489, + 346 + ], + "score": 0.92, + "content": "y ^ { c l s }", + "type": "inline_equation" + }, + { + "bbox": [ + 489, + 334, + 505, + 346 + ], + "score": 1.0, + "content": "but", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 104, + 344, + 504, + 359 + ], + "spans": [ + { + "bbox": [ + 104, + 344, + 132, + 359 + ], + "score": 1.0, + "content": "also a", + "type": "text" + }, + { + "bbox": [ + 132, + 346, + 164, + 356 + ], + "score": 0.88, + "content": "K \\times N", + "type": "inline_equation" + }, + { + "bbox": [ + 164, + 344, + 237, + 359 + ], + "score": 1.0, + "content": "matrix or called a", + "type": "text" + }, + { + "bbox": [ + 238, + 346, + 248, + 355 + ], + "score": 0.84, + "content": "K", + "type": "inline_equation" + }, + { + "bbox": [ + 248, + 344, + 416, + 359 + ], + "score": 1.0, + "content": "-dimensional score map as represented by", + "type": "text" + }, + { + "bbox": [ + 417, + 345, + 462, + 357 + ], + "score": 0.94, + "content": "[ y ^ { 1 } , . . . , y ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 463, + 344, + 493, + 359 + ], + "score": 1.0, + "content": ", where", + "type": "text" + }, + { + "bbox": [ + 494, + 346, + 504, + 355 + ], + "score": 0.82, + "content": "N", + "type": "inline_equation" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 356, + 272, + 369 + ], + "spans": [ + { + "bbox": [ + 105, + 356, + 272, + 369 + ], + "score": 1.0, + "content": "is the number of the output patch tokens.", + "type": "text" + } + ], + "index": 31 + } + ], + "index": 28, + "bbox_fs": [ + 104, + 290, + 506, + 369 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 372, + 505, + 428 + ], + "lines": [ + { + "bbox": [ + 106, + 372, + 505, + 385 + ], + "spans": [ + { + "bbox": [ + 106, + 372, + 505, + 385 + ], + "score": 1.0, + "content": "Specifically, we leverage a dense score map for each training image and use the cross-entropy loss", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 106, + 384, + 505, + 396 + ], + "spans": [ + { + "bbox": [ + 106, + 384, + 505, + 396 + ], + "score": 1.0, + "content": "between each output patch token and the corresponding aligned label in the dense score map as an", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 106, + 394, + 505, + 407 + ], + "spans": [ + { + "bbox": [ + 106, + 394, + 505, + 407 + ], + "score": 1.0, + "content": "auxiliary loss at the training phase. Figure 2 provides an intuitive interpretation. Given the output", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 105, + 403, + 506, + 419 + ], + "spans": [ + { + "bbox": [ + 105, + 403, + 159, + 419 + ], + "score": 1.0, + "content": "patch tokens", + "type": "text" + }, + { + "bbox": [ + 159, + 405, + 207, + 417 + ], + "score": 0.93, + "content": "X ^ { 1 } , . . . , X ^ { N }", + "type": "inline_equation" + }, + { + "bbox": [ + 208, + 403, + 326, + 419 + ], + "score": 1.0, + "content": "and the corresponding labels", + "type": "text" + }, + { + "bbox": [ + 326, + 405, + 372, + 417 + ], + "score": 0.94, + "content": "[ y ^ { 1 } , . . . , y ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 372, + 403, + 506, + 419 + ], + "score": 1.0, + "content": ", the token labeling objective can", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 106, + 416, + 161, + 428 + ], + "spans": [ + { + "bbox": [ + 106, + 416, + 161, + 428 + ], + "score": 1.0, + "content": "be defined as", + "type": "text" + } + ], + "index": 36 + } + ], + "index": 34, + "bbox_fs": [ + 105, + 372, + 506, + 428 + ] + }, + { + "type": "interline_equation", + "bbox": [ + 254, + 433, + 357, + 468 + ], + "lines": [ + { + "bbox": [ + 254, + 433, + 357, + 468 + ], + "spans": [ + { + "bbox": [ + 254, + 433, + 357, + 468 + ], + "score": 0.95, + "content": "L _ { t l } = \\frac { 1 } { N } \\sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) .", + "type": "interline_equation", + "image_path": "a9d480948e81c19661ff7ef1b5daada45648302552f0f35ccc9e1a4855e07862.jpg" + } + ] + } + ], + "index": 37.5, + "virtual_lines": [ + { + "bbox": [ + 254, + 433, + 357, + 450.5 + ], + "spans": [], + "index": 37 + }, + { + "bbox": [ + 254, + 450.5, + 357, + 468.0 + ], + "spans": [], + "index": 38 + } + ] + }, + { + "type": "text", + "bbox": [ + 105, + 478, + 468, + 491 + ], + "lines": [ + { + "bbox": [ + 106, + 478, + 468, + 491 + ], + "spans": [ + { + "bbox": [ + 106, + 478, + 152, + 491 + ], + "score": 1.0, + "content": "Recall that", + "type": "text" + }, + { + "bbox": [ + 152, + 480, + 162, + 489 + ], + "score": 0.84, + "content": "H", + "type": "inline_equation" + }, + { + "bbox": [ + 163, + 478, + 468, + 491 + ], + "score": 1.0, + "content": "is the cross-entropy loss. Therefore, the total loss function can be written as", + "type": "text" + } + ], + "index": 39 + } + ], + "index": 39, + "bbox_fs": [ + 106, + 478, + 468, + 491 + ] + }, + { + "type": "interline_equation", + "bbox": [ + 207, + 501, + 403, + 553 + ], + "lines": [ + { + "bbox": [ + 207, + 501, + 403, + 553 + ], + "spans": [ + { + "bbox": [ + 207, + 501, + 403, + 553 + ], + "score": 0.94, + "content": "\\begin{array} { l } { { { \\cal L } _ { t o t a l } = H ( X ^ { c l s } , y ^ { c l s } ) + \\beta \\cdot L _ { t l } , } } \\\\ { { { } } } \\\\ { { = H ( X ^ { c l s } , y ^ { c l s } ) + \\beta \\cdot \\displaystyle \\frac { 1 } { N } \\sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) , } } \\end{array}", + "type": "interline_equation", + "image_path": "c5da8361fd2faf0126b83503cc2cc4f9798363630c4f410a80eef81deafa2915.jpg" + } + ] + } + ], + "index": 41.5, + "virtual_lines": [ + { + "bbox": [ + 207, + 501, + 403, + 514.0 + ], + "spans": [], + "index": 40 + }, + { + "bbox": [ + 207, + 514.0, + 403, + 527.0 + ], + "spans": [], + "index": 41 + }, + { + "bbox": [ + 207, + 527.0, + 403, + 540.0 + ], + "spans": [], + "index": 42 + }, + { + "bbox": [ + 207, + 540.0, + 403, + 553.0 + ], + "spans": [], + "index": 43 + } + ] + }, + { + "type": "text", + "bbox": [ + 107, + 563, + 505, + 585 + ], + "lines": [ + { + "bbox": [ + 105, + 561, + 505, + 577 + ], + "spans": [ + { + "bbox": [ + 105, + 561, + 133, + 577 + ], + "score": 1.0, + "content": "where", + "type": "text" + }, + { + "bbox": [ + 133, + 564, + 141, + 575 + ], + "score": 0.85, + "content": "\\beta", + "type": "inline_equation" + }, + { + "bbox": [ + 141, + 561, + 505, + 577 + ], + "score": 1.0, + "content": "is a hyper-parameter to balance the two terms. In our experiment, we empirically set it to", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 104, + 572, + 126, + 587 + ], + "spans": [ + { + "bbox": [ + 104, + 572, + 126, + 587 + ], + "score": 1.0, + "content": "0.5.", + "type": "text" + } + ], + "index": 45 + } + ], + "index": 44.5, + "bbox_fs": [ + 104, + 561, + 505, + 587 + ] + }, + { + "type": "text", + "bbox": [ + 106, + 590, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 104, + 590, + 505, + 604 + ], + "spans": [ + { + "bbox": [ + 104, + 590, + 505, + 604 + ], + "score": 1.0, + "content": "Advantages: Our token labeling offers the following advantages. First of all, unlike knowledge", + "type": "text" + } + ], + "index": 46 + }, + { + "bbox": [ + 105, + 601, + 506, + 615 + ], + "spans": [ + { + "bbox": [ + 105, + 601, + 506, + 615 + ], + "score": 1.0, + "content": "distillation methods that require a teacher model to generate supervision labels online, token labeling", + "type": "text" + } + ], + "index": 47 + }, + { + "bbox": [ + 104, + 612, + 506, + 626 + ], + "spans": [ + { + "bbox": [ + 104, + 612, + 506, + 626 + ], + "score": 1.0, + "content": "is a cheap operation. The dense score map can be generated by a pretrained model in advance (e.g.,", + "type": "text" + } + ], + "index": 48 + }, + { + "bbox": [ + 106, + 624, + 505, + 636 + ], + "spans": [ + { + "bbox": [ + 106, + 624, + 505, + 636 + ], + "score": 1.0, + "content": "EfficientNet [34] or NFNet [3]). During training, we only need to crop the score map and perform", + "type": "text" + } + ], + "index": 49 + }, + { + "bbox": [ + 105, + 634, + 506, + 647 + ], + "spans": [ + { + "bbox": [ + 105, + 634, + 506, + 647 + ], + "score": 1.0, + "content": "interpolation to make it aligned with the cropped image in the spatial coordinate. Thus, the additional", + "type": "text" + } + ], + "index": 50 + }, + { + "bbox": [ + 105, + 645, + 506, + 658 + ], + "spans": [ + { + "bbox": [ + 105, + 645, + 506, + 658 + ], + "score": 1.0, + "content": "computations are negligible. Second, rather than utilizing a single label vector as supervision as done", + "type": "text" + } + ], + "index": 51 + }, + { + "bbox": [ + 105, + 656, + 505, + 669 + ], + "spans": [ + { + "bbox": [ + 105, + 656, + 505, + 669 + ], + "score": 1.0, + "content": "in most classification models and the ReLabel strategy [49], we also harness score maps to supervise", + "type": "text" + } + ], + "index": 52 + }, + { + "bbox": [ + 106, + 667, + 505, + 679 + ], + "spans": [ + { + "bbox": [ + 106, + 667, + 505, + 679 + ], + "score": 1.0, + "content": "the models in a dense manner and thereby the label for each patch token provides location-specific", + "type": "text" + } + ], + "index": 53 + }, + { + "bbox": [ + 106, + 678, + 505, + 690 + ], + "spans": [ + { + "bbox": [ + 106, + 678, + 505, + 690 + ], + "score": 1.0, + "content": "information, which can aid the training models to easily discover the target objects and improve the", + "type": "text" + } + ], + "index": 54 + }, + { + "bbox": [ + 105, + 689, + 505, + 701 + ], + "spans": [ + { + "bbox": [ + 105, + 689, + 505, + 701 + ], + "score": 1.0, + "content": "recognition accuracy. Last but not the least, as dense supervision is adopted in training, we found", + "type": "text" + } + ], + "index": 55 + }, + { + "bbox": [ + 106, + 700, + 505, + 713 + ], + "spans": [ + { + "bbox": [ + 106, + 700, + 505, + 713 + ], + "score": 1.0, + "content": "that the pretrained models with token labeling benefit downstream tasks with dense prediction, like", + "type": "text" + } + ], + "index": 56 + }, + { + "bbox": [ + 106, + 712, + 201, + 722 + ], + "spans": [ + { + "bbox": [ + 106, + 712, + 201, + 722 + ], + "score": 1.0, + "content": "semantic segmentation.", + "type": "text" + } + ], + "index": 57 + } + ], + "index": 51.5, + "bbox_fs": [ + 104, + 590, + 506, + 722 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "title", + "bbox": [ + 108, + 72, + 264, + 84 + ], + "lines": [ + { + "bbox": [ + 104, + 71, + 265, + 87 + ], + "spans": [ + { + "bbox": [ + 104, + 71, + 265, + 87 + ], + "score": 1.0, + "content": "3.3 Token Labeling with MixToken", + "type": "text" + } + ], + "index": 0 + } + ], + "index": 0 + }, + { + "type": "text", + "bbox": [ + 106, + 93, + 505, + 204 + ], + "lines": [ + { + "bbox": [ + 106, + 94, + 506, + 106 + ], + "spans": [ + { + "bbox": [ + 106, + 94, + 506, + 106 + ], + "score": 1.0, + "content": "While training vision transformer, previous studies [36, 46] have shown that augmentation methods,", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 106, + 105, + 505, + 117 + ], + "spans": [ + { + "bbox": [ + 106, + 105, + 505, + 117 + ], + "score": 1.0, + "content": "like MixUp [52] and CutMix [48], can effectively boost the performance and robustness of the", + "type": "text" + } + ], + "index": 2 + }, + { + "bbox": [ + 105, + 115, + 506, + 129 + ], + "spans": [ + { + "bbox": [ + 105, + 115, + 506, + 129 + ], + "score": 1.0, + "content": "models. However, vision transformers rely on patch-based tokenization to map each input image to a", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 104, + 126, + 506, + 140 + ], + "spans": [ + { + "bbox": [ + 104, + 126, + 506, + 140 + ], + "score": 1.0, + "content": "sequence of tokens and our token labeling strategy also operates on patch-based token labels. If we", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 137, + 506, + 151 + ], + "spans": [ + { + "bbox": [ + 105, + 137, + 506, + 151 + ], + "score": 1.0, + "content": "apply CutMix directly on the raw image, some of the resulting patches may contain content from two", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 106, + 148, + 505, + 161 + ], + "spans": [ + { + "bbox": [ + 106, + 148, + 505, + 161 + ], + "score": 1.0, + "content": "images, leading to mixed regions within a small patch as shown in Figure 3. When performing token", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 105, + 159, + 506, + 172 + ], + "spans": [ + { + "bbox": [ + 105, + 159, + 506, + 172 + ], + "score": 1.0, + "content": "labeling, it is difficult to assign each output token a clean and correct label. Taking this situation into", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 106, + 171, + 505, + 182 + ], + "spans": [ + { + "bbox": [ + 106, + 171, + 505, + 182 + ], + "score": 1.0, + "content": "account, we rethink the CutMix augmentation method and present MixToken, which can be viewed", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 105, + 181, + 505, + 194 + ], + "spans": [ + { + "bbox": [ + 105, + 181, + 505, + 194 + ], + "score": 1.0, + "content": "as a modified version of CutMix operating on the tokens after patch embedding as illustrated in the", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 105, + 192, + 195, + 205 + ], + "spans": [ + { + "bbox": [ + 105, + 192, + 195, + 205 + ], + "score": 1.0, + "content": "right part of Figure 3.", + "type": "text" + } + ], + "index": 10 + } + ], + "index": 5.5 + }, + { + "type": "text", + "bbox": [ + 107, + 208, + 505, + 253 + ], + "lines": [ + { + "bbox": [ + 104, + 207, + 504, + 222 + ], + "spans": [ + { + "bbox": [ + 104, + 207, + 267, + 222 + ], + "score": 1.0, + "content": "To be specific, for two images denoted as", + "type": "text" + }, + { + "bbox": [ + 267, + 209, + 290, + 220 + ], + "score": 0.93, + "content": "I _ { 1 } , I _ { 2 }", + "type": "inline_equation" + }, + { + "bbox": [ + 290, + 207, + 434, + 222 + ], + "score": 1.0, + "content": "and their corresponding token labels", + "type": "text" + }, + { + "bbox": [ + 434, + 207, + 504, + 221 + ], + "score": 0.93, + "content": "Y _ { 1 } = [ y _ { 1 } ^ { 1 } , . . . , y _ { 1 } ^ { N } ]", + "type": "inline_equation" + } + ], + "index": 11 + }, + { + "bbox": [ + 104, + 218, + 506, + 232 + ], + "spans": [ + { + "bbox": [ + 104, + 219, + 146, + 232 + ], + "score": 1.0, + "content": "as well as", + "type": "text" + }, + { + "bbox": [ + 146, + 218, + 216, + 231 + ], + "score": 0.93, + "content": "Y _ { 2 } = [ y _ { 2 } ^ { 1 } , . . . , y _ { 2 } ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 216, + 219, + 506, + 232 + ], + "score": 1.0, + "content": ", we first feed the two images into the patch embedding module to tokenize", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 228, + 507, + 244 + ], + "spans": [ + { + "bbox": [ + 105, + 228, + 270, + 244 + ], + "score": 1.0, + "content": "each as a sequence of tokens, resulting in", + "type": "text" + }, + { + "bbox": [ + 270, + 230, + 336, + 243 + ], + "score": 0.93, + "content": "T _ { 1 } = [ t _ { 1 } ^ { 1 } , . . . , t _ { 1 } ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 337, + 228, + 354, + 244 + ], + "score": 1.0, + "content": "and", + "type": "text" + }, + { + "bbox": [ + 355, + 230, + 421, + 243 + ], + "score": 0.93, + "content": "\\bar { T _ { 2 } } = [ t _ { 2 } ^ { 1 } , . . . , t _ { 2 } ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 421, + 228, + 507, + 244 + ], + "score": 1.0, + "content": ". Then, we produce a", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 241, + 439, + 254 + ], + "spans": [ + { + "bbox": [ + 105, + 241, + 379, + 254 + ], + "score": 1.0, + "content": "new sequence of tokens by applying MixToken using a binary mask", + "type": "text" + }, + { + "bbox": [ + 380, + 243, + 391, + 251 + ], + "score": 0.83, + "content": "M", + "type": "inline_equation" + }, + { + "bbox": [ + 392, + 241, + 439, + 254 + ], + "score": 1.0, + "content": "as follows:", + "type": "text" + } + ], + "index": 14 + } + ], + "index": 12.5 + }, + { + "type": "interline_equation", + "bbox": [ + 240, + 261, + 370, + 277 + ], + "lines": [ + { + "bbox": [ + 240, + 261, + 370, + 277 + ], + "spans": [ + { + "bbox": [ + 240, + 261, + 370, + 277 + ], + "score": 0.91, + "content": "\\hat { T } = T _ { 1 } \\odot M + T _ { 2 } \\odot ( 1 - M ) ,", + "type": "interline_equation", + "image_path": "cdf5e3aab4339f76c0c2c44221d0f16054703d72f085272d761449ddf6e388dc.jpg" + } + ] + } + ], + "index": 15, + "virtual_lines": [ + { + "bbox": [ + 240, + 261, + 370, + 277 + ], + "spans": [], + "index": 15 + } + ] + }, + { + "type": "text", + "bbox": [ + 106, + 284, + 504, + 307 + ], + "lines": [ + { + "bbox": [ + 106, + 285, + 504, + 296 + ], + "spans": [ + { + "bbox": [ + 106, + 285, + 133, + 296 + ], + "score": 1.0, + "content": "where", + "type": "text" + }, + { + "bbox": [ + 134, + 286, + 143, + 295 + ], + "score": 0.83, + "content": "\\odot", + "type": "inline_equation" + }, + { + "bbox": [ + 143, + 285, + 450, + 296 + ], + "score": 1.0, + "content": "is element-wise multiplication. We use the same way to generate the mask", + "type": "text" + }, + { + "bbox": [ + 450, + 285, + 462, + 294 + ], + "score": 0.82, + "content": "M", + "type": "inline_equation" + }, + { + "bbox": [ + 462, + 285, + 504, + 296 + ], + "score": 1.0, + "content": "as in [48].", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 295, + 424, + 308 + ], + "spans": [ + { + "bbox": [ + 105, + 295, + 408, + 308 + ], + "score": 1.0, + "content": "For the corresponding token labels, we also mix them using the same mask", + "type": "text" + }, + { + "bbox": [ + 408, + 296, + 420, + 306 + ], + "score": 0.83, + "content": "M", + "type": "inline_equation" + }, + { + "bbox": [ + 420, + 295, + 424, + 308 + ], + "score": 1.0, + "content": ":", + "type": "text" + } + ], + "index": 17 + } + ], + "index": 16.5 + }, + { + "type": "interline_equation", + "bbox": [ + 240, + 315, + 371, + 331 + ], + "lines": [ + { + "bbox": [ + 240, + 315, + 371, + 331 + ], + "spans": [ + { + "bbox": [ + 240, + 315, + 371, + 331 + ], + "score": 0.92, + "content": "\\hat { Y } = Y _ { 1 } \\odot M + Y _ { 2 } \\odot ( 1 - M ) .", + "type": "interline_equation", + "image_path": "fa81133b791d6c30dfc1c80f85c7a073a5241a9fada2c61ef2a3b397042bb25e.jpg" + } + ] + } + ], + "index": 18, + "virtual_lines": [ + { + "bbox": [ + 240, + 315, + 371, + 331 + ], + "spans": [], + "index": 18 + } + ] + }, + { + "type": "text", + "bbox": [ + 108, + 339, + 290, + 350 + ], + "lines": [ + { + "bbox": [ + 106, + 338, + 291, + 351 + ], + "spans": [ + { + "bbox": [ + 106, + 338, + 291, + 351 + ], + "score": 1.0, + "content": "The label for the class token can be written as", + "type": "text" + } + ], + "index": 19 + } + ], + "index": 19 + }, + { + "type": "interline_equation", + "bbox": [ + 243, + 356, + 367, + 372 + ], + "lines": [ + { + "bbox": [ + 243, + 356, + 367, + 372 + ], + "spans": [ + { + "bbox": [ + 243, + 356, + 367, + 372 + ], + "score": 0.92, + "content": "y ^ { \\hat { c } l s } = \\bar { M } y _ { 1 } ^ { c l s } + ( 1 - \\bar { M } ) y _ { 2 } ^ { c l s } ,", + "type": "interline_equation", + "image_path": "1bb8081440df26328f22eeb83ce87e3030fb4e4f726532138a58c694ac3a6d6a.jpg" + } + ] + } + ], + "index": 20, + "virtual_lines": [ + { + "bbox": [ + 243, + 356, + 367, + 372 + ], + "spans": [], + "index": 20 + } + ] + }, + { + "type": "text", + "bbox": [ + 106, + 380, + 314, + 392 + ], + "lines": [ + { + "bbox": [ + 106, + 380, + 315, + 393 + ], + "spans": [ + { + "bbox": [ + 106, + 380, + 133, + 393 + ], + "score": 1.0, + "content": "where", + "type": "text" + }, + { + "bbox": [ + 133, + 380, + 145, + 390 + ], + "score": 0.87, + "content": "\\bar { M }", + "type": "inline_equation" + }, + { + "bbox": [ + 145, + 380, + 299, + 393 + ], + "score": 1.0, + "content": "is the average of all element values of", + "type": "text" + }, + { + "bbox": [ + 299, + 381, + 311, + 390 + ], + "score": 0.86, + "content": "M", + "type": "inline_equation" + }, + { + "bbox": [ + 311, + 380, + 315, + 393 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 21 + }, + { + "type": "title", + "bbox": [ + 107, + 410, + 191, + 424 + ], + "lines": [ + { + "bbox": [ + 104, + 408, + 193, + 427 + ], + "spans": [ + { + "bbox": [ + 104, + 408, + 193, + 427 + ], + "score": 1.0, + "content": "4 Experiments", + "type": "text" + } + ], + "index": 22 + } + ], + "index": 22 + }, + { + "type": "title", + "bbox": [ + 107, + 437, + 208, + 449 + ], + "lines": [ + { + "bbox": [ + 105, + 434, + 209, + 453 + ], + "spans": [ + { + "bbox": [ + 105, + 434, + 209, + 453 + ], + "score": 1.0, + "content": "4.1 Experiment Setup", + "type": "text" + } + ], + "index": 23 + } + ], + "index": 23 + }, + { + "type": "text", + "bbox": [ + 106, + 458, + 505, + 525 + ], + "lines": [ + { + "bbox": [ + 105, + 457, + 505, + 471 + ], + "spans": [ + { + "bbox": [ + 105, + 457, + 505, + 471 + ], + "score": 1.0, + "content": "We evaluate our method on the ImageNet [13] dataset. All experiments are built and conducted upon", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 106, + 470, + 506, + 482 + ], + "spans": [ + { + "bbox": [ + 106, + 470, + 506, + 482 + ], + "score": 1.0, + "content": "PyTorch [29] and the timm [42] library. We follow the standard training schedule and train our", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 106, + 481, + 505, + 493 + ], + "spans": [ + { + "bbox": [ + 106, + 481, + 505, + 493 + ], + "score": 1.0, + "content": "models on the ImageNet dataset for 300 epochs. Besides normal augmentations like CutOut [57] and", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 105, + 491, + 506, + 505 + ], + "spans": [ + { + "bbox": [ + 105, + 491, + 506, + 505 + ], + "score": 1.0, + "content": "RandAug [10], we also explore the effect of applying MixUp [52] and CutMix [48] together with our", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 105, + 501, + 505, + 515 + ], + "spans": [ + { + "bbox": [ + 105, + 501, + 505, + 515 + ], + "score": 1.0, + "content": "proposed token labeling. Empirically, we have found that using MixUp together with token labeling", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 513, + 446, + 527 + ], + "spans": [ + { + "bbox": [ + 105, + 513, + 446, + 527 + ], + "score": 1.0, + "content": "brings no benefit to the performance, and thus we do not apply it in our experiments.", + "type": "text" + } + ], + "index": 29 + } + ], + "index": 26.5 + }, + { + "type": "text", + "bbox": [ + 107, + 529, + 505, + 596 + ], + "lines": [ + { + "bbox": [ + 105, + 528, + 505, + 543 + ], + "spans": [ + { + "bbox": [ + 105, + 528, + 505, + 543 + ], + "score": 1.0, + "content": "For optimization, by default, we use the AdamW optimizer [27] with a linear learning rate scaling", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 103, + 536, + 507, + 557 + ], + "spans": [ + { + "bbox": [ + 103, + 536, + 141, + 557 + ], + "score": 1.0, + "content": "strategy", + "type": "text" + }, + { + "bbox": [ + 141, + 540, + 237, + 555 + ], + "score": 0.93, + "content": "\\begin{array} { r } { l r = 1 0 ^ { - 3 } \\times \\frac { b a t c h \\_ s i z e } { 6 4 0 } } \\end{array}", + "type": "inline_equation" + }, + { + "bbox": [ + 237, + 536, + 256, + 557 + ], + "score": 1.0, + "content": "and", + "type": "text" + }, + { + "bbox": [ + 256, + 541, + 295, + 552 + ], + "score": 0.91, + "content": "5 \\times 1 0 ^ { - 2 }", + "type": "inline_equation" + }, + { + "bbox": [ + 296, + 536, + 507, + 557 + ], + "score": 1.0, + "content": "weight decay rate. For Dropout regularization, we", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 552, + 505, + 565 + ], + "spans": [ + { + "bbox": [ + 105, + 552, + 505, + 565 + ], + "score": 1.0, + "content": "observe that for small models, using Dropout hurts the performance. This has also been observed in", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 105, + 564, + 505, + 577 + ], + "spans": [ + { + "bbox": [ + 105, + 564, + 505, + 577 + ], + "score": 1.0, + "content": "a few other works related to training vision transformers [36, 37, 46]. As a result, we do not apply", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 572, + 506, + 588 + ], + "spans": [ + { + "bbox": [ + 105, + 572, + 506, + 588 + ], + "score": 1.0, + "content": "Dropout [32] and use Stochastic Depth [23] instead. More details on hyper-parameters and finetuning", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 106, + 586, + 289, + 597 + ], + "spans": [ + { + "bbox": [ + 106, + 586, + 289, + 597 + ], + "score": 1.0, + "content": "can be found in our supplementary materials.", + "type": "text" + } + ], + "index": 35 + } + ], + "index": 32.5 + }, + { + "type": "text", + "bbox": [ + 106, + 601, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 105, + 601, + 505, + 615 + ], + "spans": [ + { + "bbox": [ + 105, + 601, + 342, + 615 + ], + "score": 1.0, + "content": "We use the NFNet-F6 [3] trained on ImageNet with an", + "type": "text" + }, + { + "bbox": [ + 343, + 601, + 370, + 613 + ], + "score": 0.87, + "content": "8 6 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 370, + 601, + 505, + 615 + ], + "score": 1.0, + "content": "Top-1 accuracy as the machine", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 105, + 613, + 505, + 625 + ], + "spans": [ + { + "bbox": [ + 105, + 613, + 505, + 625 + ], + "score": 1.0, + "content": "annotator to generate dense score maps for the ImageNet dataset, yielding a 1000-dimensional score", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 105, + 623, + 506, + 637 + ], + "spans": [ + { + "bbox": [ + 105, + 623, + 506, + 637 + ], + "score": 1.0, + "content": "map for each image for training. The score map generation procedure is similar to [49], but we", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 105, + 634, + 505, + 647 + ], + "spans": [ + { + "bbox": [ + 105, + 634, + 505, + 647 + ], + "score": 1.0, + "content": "limit our experiment setting by training all models from scratch on ImageNet without extra data", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 104, + 644, + 506, + 659 + ], + "spans": [ + { + "bbox": [ + 104, + 644, + 506, + 659 + ], + "score": 1.0, + "content": "support, such as JFT-300M and ImageNet-22K. This is different from the original ReLabel paper", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 106, + 656, + 505, + 669 + ], + "spans": [ + { + "bbox": [ + 106, + 656, + 505, + 669 + ], + "score": 1.0, + "content": "[49], in which the EfficientNet-L2 model pretrained on JFT-300M is used. The input resolution for", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 104, + 666, + 505, + 681 + ], + "spans": [ + { + "bbox": [ + 104, + 666, + 160, + 681 + ], + "score": 1.0, + "content": "NFNet-F6 is", + "type": "text" + }, + { + "bbox": [ + 160, + 667, + 203, + 677 + ], + "score": 0.89, + "content": "5 7 6 \\times 5 7 6", + "type": "inline_equation" + }, + { + "bbox": [ + 203, + 666, + 505, + 681 + ], + "score": 1.0, + "content": ", and the dimension of the corresponding output score map for each image", + "type": "text" + } + ], + "index": 42 + }, + { + "bbox": [ + 104, + 675, + 506, + 693 + ], + "spans": [ + { + "bbox": [ + 104, + 675, + 116, + 693 + ], + "score": 1.0, + "content": "is", + "type": "text" + }, + { + "bbox": [ + 117, + 677, + 190, + 688 + ], + "score": 0.93, + "content": "L \\in \\mathbb { R } ^ { 1 8 \\times 1 8 \\times 1 0 0 0 }", + "type": "inline_equation" + }, + { + "bbox": [ + 191, + 675, + 506, + 693 + ], + "score": 1.0, + "content": ". During training, the target labels for the tokens are generated by applying", + "type": "text" + } + ], + "index": 43 + }, + { + "bbox": [ + 105, + 688, + 506, + 702 + ], + "spans": [ + { + "bbox": [ + 105, + 688, + 506, + 702 + ], + "score": 1.0, + "content": "RoIAlign [17] on the corresponding score map. In practice, we only store the top-5 score maps for", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 105, + 699, + 506, + 713 + ], + "spans": [ + { + "bbox": [ + 105, + 699, + 506, + 713 + ], + "score": 1.0, + "content": "each position in half-precision to save space as storing the entire score maps for all the images results", + "type": "text" + } + ], + "index": 45 + }, + { + "bbox": [ + 105, + 710, + 480, + 725 + ], + "spans": [ + { + "bbox": [ + 105, + 710, + 480, + 725 + ], + "score": 1.0, + "content": "in 2TB storage. In our experiment, we only need 10GB of storage to store all the score maps.", + "type": "text" + } + ], + "index": 46 + } + ], + "index": 41 + } + ], + "page_idx": 4, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 302, + 741, + 309, + 750 + ], + "lines": [ + { + "bbox": [ + 301, + 740, + 310, + 753 + ], + "spans": [ + { + "bbox": [ + 301, + 740, + 310, + 753 + ], + "score": 1.0, + "content": "5", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "title", + "bbox": [ + 108, + 72, + 264, + 84 + ], + "lines": [ + { + "bbox": [ + 104, + 71, + 265, + 87 + ], + "spans": [ + { + "bbox": [ + 104, + 71, + 265, + 87 + ], + "score": 1.0, + "content": "3.3 Token Labeling with MixToken", + "type": "text" + } + ], + "index": 0 + } + ], + "index": 0 + }, + { + "type": "text", + "bbox": [ + 106, + 93, + 505, + 204 + ], + "lines": [ + { + "bbox": [ + 106, + 94, + 506, + 106 + ], + "spans": [ + { + "bbox": [ + 106, + 94, + 506, + 106 + ], + "score": 1.0, + "content": "While training vision transformer, previous studies [36, 46] have shown that augmentation methods,", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 106, + 105, + 505, + 117 + ], + "spans": [ + { + "bbox": [ + 106, + 105, + 505, + 117 + ], + "score": 1.0, + "content": "like MixUp [52] and CutMix [48], can effectively boost the performance and robustness of the", + "type": "text" + } + ], + "index": 2 + }, + { + "bbox": [ + 105, + 115, + 506, + 129 + ], + "spans": [ + { + "bbox": [ + 105, + 115, + 506, + 129 + ], + "score": 1.0, + "content": "models. However, vision transformers rely on patch-based tokenization to map each input image to a", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 104, + 126, + 506, + 140 + ], + "spans": [ + { + "bbox": [ + 104, + 126, + 506, + 140 + ], + "score": 1.0, + "content": "sequence of tokens and our token labeling strategy also operates on patch-based token labels. If we", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 137, + 506, + 151 + ], + "spans": [ + { + "bbox": [ + 105, + 137, + 506, + 151 + ], + "score": 1.0, + "content": "apply CutMix directly on the raw image, some of the resulting patches may contain content from two", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 106, + 148, + 505, + 161 + ], + "spans": [ + { + "bbox": [ + 106, + 148, + 505, + 161 + ], + "score": 1.0, + "content": "images, leading to mixed regions within a small patch as shown in Figure 3. When performing token", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 105, + 159, + 506, + 172 + ], + "spans": [ + { + "bbox": [ + 105, + 159, + 506, + 172 + ], + "score": 1.0, + "content": "labeling, it is difficult to assign each output token a clean and correct label. Taking this situation into", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 106, + 171, + 505, + 182 + ], + "spans": [ + { + "bbox": [ + 106, + 171, + 505, + 182 + ], + "score": 1.0, + "content": "account, we rethink the CutMix augmentation method and present MixToken, which can be viewed", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 105, + 181, + 505, + 194 + ], + "spans": [ + { + "bbox": [ + 105, + 181, + 505, + 194 + ], + "score": 1.0, + "content": "as a modified version of CutMix operating on the tokens after patch embedding as illustrated in the", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 105, + 192, + 195, + 205 + ], + "spans": [ + { + "bbox": [ + 105, + 192, + 195, + 205 + ], + "score": 1.0, + "content": "right part of Figure 3.", + "type": "text" + } + ], + "index": 10 + } + ], + "index": 5.5, + "bbox_fs": [ + 104, + 94, + 506, + 205 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 208, + 505, + 253 + ], + "lines": [ + { + "bbox": [ + 104, + 207, + 504, + 222 + ], + "spans": [ + { + "bbox": [ + 104, + 207, + 267, + 222 + ], + "score": 1.0, + "content": "To be specific, for two images denoted as", + "type": "text" + }, + { + "bbox": [ + 267, + 209, + 290, + 220 + ], + "score": 0.93, + "content": "I _ { 1 } , I _ { 2 }", + "type": "inline_equation" + }, + { + "bbox": [ + 290, + 207, + 434, + 222 + ], + "score": 1.0, + "content": "and their corresponding token labels", + "type": "text" + }, + { + "bbox": [ + 434, + 207, + 504, + 221 + ], + "score": 0.93, + "content": "Y _ { 1 } = [ y _ { 1 } ^ { 1 } , . . . , y _ { 1 } ^ { N } ]", + "type": "inline_equation" + } + ], + "index": 11 + }, + { + "bbox": [ + 104, + 218, + 506, + 232 + ], + "spans": [ + { + "bbox": [ + 104, + 219, + 146, + 232 + ], + "score": 1.0, + "content": "as well as", + "type": "text" + }, + { + "bbox": [ + 146, + 218, + 216, + 231 + ], + "score": 0.93, + "content": "Y _ { 2 } = [ y _ { 2 } ^ { 1 } , . . . , y _ { 2 } ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 216, + 219, + 506, + 232 + ], + "score": 1.0, + "content": ", we first feed the two images into the patch embedding module to tokenize", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 228, + 507, + 244 + ], + "spans": [ + { + "bbox": [ + 105, + 228, + 270, + 244 + ], + "score": 1.0, + "content": "each as a sequence of tokens, resulting in", + "type": "text" + }, + { + "bbox": [ + 270, + 230, + 336, + 243 + ], + "score": 0.93, + "content": "T _ { 1 } = [ t _ { 1 } ^ { 1 } , . . . , t _ { 1 } ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 337, + 228, + 354, + 244 + ], + "score": 1.0, + "content": "and", + "type": "text" + }, + { + "bbox": [ + 355, + 230, + 421, + 243 + ], + "score": 0.93, + "content": "\\bar { T _ { 2 } } = [ t _ { 2 } ^ { 1 } , . . . , t _ { 2 } ^ { N } ]", + "type": "inline_equation" + }, + { + "bbox": [ + 421, + 228, + 507, + 244 + ], + "score": 1.0, + "content": ". Then, we produce a", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 241, + 439, + 254 + ], + "spans": [ + { + "bbox": [ + 105, + 241, + 379, + 254 + ], + "score": 1.0, + "content": "new sequence of tokens by applying MixToken using a binary mask", + "type": "text" + }, + { + "bbox": [ + 380, + 243, + 391, + 251 + ], + "score": 0.83, + "content": "M", + "type": "inline_equation" + }, + { + "bbox": [ + 392, + 241, + 439, + 254 + ], + "score": 1.0, + "content": "as follows:", + "type": "text" + } + ], + "index": 14 + } + ], + "index": 12.5, + "bbox_fs": [ + 104, + 207, + 507, + 254 + ] + }, + { + "type": "interline_equation", + "bbox": [ + 240, + 261, + 370, + 277 + ], + "lines": [ + { + "bbox": [ + 240, + 261, + 370, + 277 + ], + "spans": [ + { + "bbox": [ + 240, + 261, + 370, + 277 + ], + "score": 0.91, + "content": "\\hat { T } = T _ { 1 } \\odot M + T _ { 2 } \\odot ( 1 - M ) ,", + "type": "interline_equation", + "image_path": "cdf5e3aab4339f76c0c2c44221d0f16054703d72f085272d761449ddf6e388dc.jpg" + } + ] + } + ], + "index": 15, + "virtual_lines": [ + { + "bbox": [ + 240, + 261, + 370, + 277 + ], + "spans": [], + "index": 15 + } + ] + }, + { + "type": "text", + "bbox": [ + 106, + 284, + 504, + 307 + ], + "lines": [ + { + "bbox": [ + 106, + 285, + 504, + 296 + ], + "spans": [ + { + "bbox": [ + 106, + 285, + 133, + 296 + ], + "score": 1.0, + "content": "where", + "type": "text" + }, + { + "bbox": [ + 134, + 286, + 143, + 295 + ], + "score": 0.83, + "content": "\\odot", + "type": "inline_equation" + }, + { + "bbox": [ + 143, + 285, + 450, + 296 + ], + "score": 1.0, + "content": "is element-wise multiplication. We use the same way to generate the mask", + "type": "text" + }, + { + "bbox": [ + 450, + 285, + 462, + 294 + ], + "score": 0.82, + "content": "M", + "type": "inline_equation" + }, + { + "bbox": [ + 462, + 285, + 504, + 296 + ], + "score": 1.0, + "content": "as in [48].", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 295, + 424, + 308 + ], + "spans": [ + { + "bbox": [ + 105, + 295, + 408, + 308 + ], + "score": 1.0, + "content": "For the corresponding token labels, we also mix them using the same mask", + "type": "text" + }, + { + "bbox": [ + 408, + 296, + 420, + 306 + ], + "score": 0.83, + "content": "M", + "type": "inline_equation" + }, + { + "bbox": [ + 420, + 295, + 424, + 308 + ], + "score": 1.0, + "content": ":", + "type": "text" + } + ], + "index": 17 + } + ], + "index": 16.5, + "bbox_fs": [ + 105, + 285, + 504, + 308 + ] + }, + { + "type": "interline_equation", + "bbox": [ + 240, + 315, + 371, + 331 + ], + "lines": [ + { + "bbox": [ + 240, + 315, + 371, + 331 + ], + "spans": [ + { + "bbox": [ + 240, + 315, + 371, + 331 + ], + "score": 0.92, + "content": "\\hat { Y } = Y _ { 1 } \\odot M + Y _ { 2 } \\odot ( 1 - M ) .", + "type": "interline_equation", + "image_path": "fa81133b791d6c30dfc1c80f85c7a073a5241a9fada2c61ef2a3b397042bb25e.jpg" + } + ] + } + ], + "index": 18, + "virtual_lines": [ + { + "bbox": [ + 240, + 315, + 371, + 331 + ], + "spans": [], + "index": 18 + } + ] + }, + { + "type": "text", + "bbox": [ + 108, + 339, + 290, + 350 + ], + "lines": [ + { + "bbox": [ + 106, + 338, + 291, + 351 + ], + "spans": [ + { + "bbox": [ + 106, + 338, + 291, + 351 + ], + "score": 1.0, + "content": "The label for the class token can be written as", + "type": "text" + } + ], + "index": 19 + } + ], + "index": 19, + "bbox_fs": [ + 106, + 338, + 291, + 351 + ] + }, + { + "type": "interline_equation", + "bbox": [ + 243, + 356, + 367, + 372 + ], + "lines": [ + { + "bbox": [ + 243, + 356, + 367, + 372 + ], + "spans": [ + { + "bbox": [ + 243, + 356, + 367, + 372 + ], + "score": 0.92, + "content": "y ^ { \\hat { c } l s } = \\bar { M } y _ { 1 } ^ { c l s } + ( 1 - \\bar { M } ) y _ { 2 } ^ { c l s } ,", + "type": "interline_equation", + "image_path": "1bb8081440df26328f22eeb83ce87e3030fb4e4f726532138a58c694ac3a6d6a.jpg" + } + ] + } + ], + "index": 20, + "virtual_lines": [ + { + "bbox": [ + 243, + 356, + 367, + 372 + ], + "spans": [], + "index": 20 + } + ] + }, + { + "type": "text", + "bbox": [ + 106, + 380, + 314, + 392 + ], + "lines": [ + { + "bbox": [ + 106, + 380, + 315, + 393 + ], + "spans": [ + { + "bbox": [ + 106, + 380, + 133, + 393 + ], + "score": 1.0, + "content": "where", + "type": "text" + }, + { + "bbox": [ + 133, + 380, + 145, + 390 + ], + "score": 0.87, + "content": "\\bar { M }", + "type": "inline_equation" + }, + { + "bbox": [ + 145, + 380, + 299, + 393 + ], + "score": 1.0, + "content": "is the average of all element values of", + "type": "text" + }, + { + "bbox": [ + 299, + 381, + 311, + 390 + ], + "score": 0.86, + "content": "M", + "type": "inline_equation" + }, + { + "bbox": [ + 311, + 380, + 315, + 393 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 21, + "bbox_fs": [ + 106, + 380, + 315, + 393 + ] + }, + { + "type": "title", + "bbox": [ + 107, + 410, + 191, + 424 + ], + "lines": [ + { + "bbox": [ + 104, + 408, + 193, + 427 + ], + "spans": [ + { + "bbox": [ + 104, + 408, + 193, + 427 + ], + "score": 1.0, + "content": "4 Experiments", + "type": "text" + } + ], + "index": 22 + } + ], + "index": 22 + }, + { + "type": "title", + "bbox": [ + 107, + 437, + 208, + 449 + ], + "lines": [ + { + "bbox": [ + 105, + 434, + 209, + 453 + ], + "spans": [ + { + "bbox": [ + 105, + 434, + 209, + 453 + ], + "score": 1.0, + "content": "4.1 Experiment Setup", + "type": "text" + } + ], + "index": 23 + } + ], + "index": 23 + }, + { + "type": "text", + "bbox": [ + 106, + 458, + 505, + 525 + ], + "lines": [ + { + "bbox": [ + 105, + 457, + 505, + 471 + ], + "spans": [ + { + "bbox": [ + 105, + 457, + 505, + 471 + ], + "score": 1.0, + "content": "We evaluate our method on the ImageNet [13] dataset. All experiments are built and conducted upon", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 106, + 470, + 506, + 482 + ], + "spans": [ + { + "bbox": [ + 106, + 470, + 506, + 482 + ], + "score": 1.0, + "content": "PyTorch [29] and the timm [42] library. We follow the standard training schedule and train our", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 106, + 481, + 505, + 493 + ], + "spans": [ + { + "bbox": [ + 106, + 481, + 505, + 493 + ], + "score": 1.0, + "content": "models on the ImageNet dataset for 300 epochs. Besides normal augmentations like CutOut [57] and", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 105, + 491, + 506, + 505 + ], + "spans": [ + { + "bbox": [ + 105, + 491, + 506, + 505 + ], + "score": 1.0, + "content": "RandAug [10], we also explore the effect of applying MixUp [52] and CutMix [48] together with our", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 105, + 501, + 505, + 515 + ], + "spans": [ + { + "bbox": [ + 105, + 501, + 505, + 515 + ], + "score": 1.0, + "content": "proposed token labeling. Empirically, we have found that using MixUp together with token labeling", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 513, + 446, + 527 + ], + "spans": [ + { + "bbox": [ + 105, + 513, + 446, + 527 + ], + "score": 1.0, + "content": "brings no benefit to the performance, and thus we do not apply it in our experiments.", + "type": "text" + } + ], + "index": 29 + } + ], + "index": 26.5, + "bbox_fs": [ + 105, + 457, + 506, + 527 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 529, + 505, + 596 + ], + "lines": [ + { + "bbox": [ + 105, + 528, + 505, + 543 + ], + "spans": [ + { + "bbox": [ + 105, + 528, + 505, + 543 + ], + "score": 1.0, + "content": "For optimization, by default, we use the AdamW optimizer [27] with a linear learning rate scaling", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 103, + 536, + 507, + 557 + ], + "spans": [ + { + "bbox": [ + 103, + 536, + 141, + 557 + ], + "score": 1.0, + "content": "strategy", + "type": "text" + }, + { + "bbox": [ + 141, + 540, + 237, + 555 + ], + "score": 0.93, + "content": "\\begin{array} { r } { l r = 1 0 ^ { - 3 } \\times \\frac { b a t c h \\_ s i z e } { 6 4 0 } } \\end{array}", + "type": "inline_equation" + }, + { + "bbox": [ + 237, + 536, + 256, + 557 + ], + "score": 1.0, + "content": "and", + "type": "text" + }, + { + "bbox": [ + 256, + 541, + 295, + 552 + ], + "score": 0.91, + "content": "5 \\times 1 0 ^ { - 2 }", + "type": "inline_equation" + }, + { + "bbox": [ + 296, + 536, + 507, + 557 + ], + "score": 1.0, + "content": "weight decay rate. For Dropout regularization, we", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 552, + 505, + 565 + ], + "spans": [ + { + "bbox": [ + 105, + 552, + 505, + 565 + ], + "score": 1.0, + "content": "observe that for small models, using Dropout hurts the performance. This has also been observed in", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 105, + 564, + 505, + 577 + ], + "spans": [ + { + "bbox": [ + 105, + 564, + 505, + 577 + ], + "score": 1.0, + "content": "a few other works related to training vision transformers [36, 37, 46]. As a result, we do not apply", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 572, + 506, + 588 + ], + "spans": [ + { + "bbox": [ + 105, + 572, + 506, + 588 + ], + "score": 1.0, + "content": "Dropout [32] and use Stochastic Depth [23] instead. More details on hyper-parameters and finetuning", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 106, + 586, + 289, + 597 + ], + "spans": [ + { + "bbox": [ + 106, + 586, + 289, + 597 + ], + "score": 1.0, + "content": "can be found in our supplementary materials.", + "type": "text" + } + ], + "index": 35 + } + ], + "index": 32.5, + "bbox_fs": [ + 103, + 528, + 507, + 597 + ] + }, + { + "type": "text", + "bbox": [ + 106, + 601, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 105, + 601, + 505, + 615 + ], + "spans": [ + { + "bbox": [ + 105, + 601, + 342, + 615 + ], + "score": 1.0, + "content": "We use the NFNet-F6 [3] trained on ImageNet with an", + "type": "text" + }, + { + "bbox": [ + 343, + 601, + 370, + 613 + ], + "score": 0.87, + "content": "8 6 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 370, + 601, + 505, + 615 + ], + "score": 1.0, + "content": "Top-1 accuracy as the machine", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 105, + 613, + 505, + 625 + ], + "spans": [ + { + "bbox": [ + 105, + 613, + 505, + 625 + ], + "score": 1.0, + "content": "annotator to generate dense score maps for the ImageNet dataset, yielding a 1000-dimensional score", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 105, + 623, + 506, + 637 + ], + "spans": [ + { + "bbox": [ + 105, + 623, + 506, + 637 + ], + "score": 1.0, + "content": "map for each image for training. The score map generation procedure is similar to [49], but we", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 105, + 634, + 505, + 647 + ], + "spans": [ + { + "bbox": [ + 105, + 634, + 505, + 647 + ], + "score": 1.0, + "content": "limit our experiment setting by training all models from scratch on ImageNet without extra data", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 104, + 644, + 506, + 659 + ], + "spans": [ + { + "bbox": [ + 104, + 644, + 506, + 659 + ], + "score": 1.0, + "content": "support, such as JFT-300M and ImageNet-22K. This is different from the original ReLabel paper", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 106, + 656, + 505, + 669 + ], + "spans": [ + { + "bbox": [ + 106, + 656, + 505, + 669 + ], + "score": 1.0, + "content": "[49], in which the EfficientNet-L2 model pretrained on JFT-300M is used. The input resolution for", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 104, + 666, + 505, + 681 + ], + "spans": [ + { + "bbox": [ + 104, + 666, + 160, + 681 + ], + "score": 1.0, + "content": "NFNet-F6 is", + "type": "text" + }, + { + "bbox": [ + 160, + 667, + 203, + 677 + ], + "score": 0.89, + "content": "5 7 6 \\times 5 7 6", + "type": "inline_equation" + }, + { + "bbox": [ + 203, + 666, + 505, + 681 + ], + "score": 1.0, + "content": ", and the dimension of the corresponding output score map for each image", + "type": "text" + } + ], + "index": 42 + }, + { + "bbox": [ + 104, + 675, + 506, + 693 + ], + "spans": [ + { + "bbox": [ + 104, + 675, + 116, + 693 + ], + "score": 1.0, + "content": "is", + "type": "text" + }, + { + "bbox": [ + 117, + 677, + 190, + 688 + ], + "score": 0.93, + "content": "L \\in \\mathbb { R } ^ { 1 8 \\times 1 8 \\times 1 0 0 0 }", + "type": "inline_equation" + }, + { + "bbox": [ + 191, + 675, + 506, + 693 + ], + "score": 1.0, + "content": ". During training, the target labels for the tokens are generated by applying", + "type": "text" + } + ], + "index": 43 + }, + { + "bbox": [ + 105, + 688, + 506, + 702 + ], + "spans": [ + { + "bbox": [ + 105, + 688, + 506, + 702 + ], + "score": 1.0, + "content": "RoIAlign [17] on the corresponding score map. In practice, we only store the top-5 score maps for", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 105, + 699, + 506, + 713 + ], + "spans": [ + { + "bbox": [ + 105, + 699, + 506, + 713 + ], + "score": 1.0, + "content": "each position in half-precision to save space as storing the entire score maps for all the images results", + "type": "text" + } + ], + "index": 45 + }, + { + "bbox": [ + 105, + 710, + 480, + 725 + ], + "spans": [ + { + "bbox": [ + 105, + 710, + 480, + 725 + ], + "score": 1.0, + "content": "in 2TB storage. In our experiment, we only need 10GB of storage to store all the score maps.", + "type": "text" + } + ], + "index": 46 + } + ], + "index": 41, + "bbox_fs": [ + 104, + 601, + 506, + 725 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "table", + "bbox": [ + 106, + 110, + 505, + 175 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 77, + 505, + 110 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 104, + 76, + 505, + 90 + ], + "spans": [ + { + "bbox": [ + 104, + 76, + 505, + 90 + ], + "score": 1.0, + "content": "Table 1: Performance of the proposed LV-ViT with different model sizes. Here, ‘depth’ denotes", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 105, + 88, + 505, + 100 + ], + "spans": [ + { + "bbox": [ + 105, + 88, + 505, + 100 + ], + "score": 1.0, + "content": "the number of transformer blocks used in different models. By default, the test resolution is set to", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 106, + 99, + 310, + 111 + ], + "spans": [ + { + "bbox": [ + 106, + 99, + 150, + 110 + ], + "score": 0.8, + "content": "2 2 4 \\times 2 2 4", + "type": "inline_equation" + }, + { + "bbox": [ + 150, + 99, + 263, + 111 + ], + "score": 1.0, + "content": "except the last one which is", + "type": "text" + }, + { + "bbox": [ + 263, + 99, + 306, + 110 + ], + "score": 0.88, + "content": "2 8 8 \\times 2 8 8", + "type": "inline_equation" + }, + { + "bbox": [ + 307, + 99, + 310, + 111 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 2 + } + ], + "index": 1 + }, + { + "type": "table_body", + "bbox": [ + 106, + 110, + 505, + 175 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 106, + 110, + 505, + 175 + ], + "spans": [ + { + "bbox": [ + 106, + 110, + 505, + 175 + ], + "score": 0.977, + "html": "
NameDepthEmbed dim.MLP Ratio#Heads#ParamsThroughput (im/s)Test sizeTop-1 Acc. (%)
LV-ViT-T122403.048.5M2032.622479.1
LV-ViT-S163843.0626M1018.222483.3
LV-ViT-M205123.0856M668.922484.1
LV-ViT-L247683.012150M204.828885.3
", + "type": "table", + "image_path": "8ab634b915f17e8f746292568270ed165dd5bacd94d1fed7927449e1328c4232.jpg" + } + ] + } + ], + "index": 4, + "virtual_lines": [ + { + "bbox": [ + 106, + 110, + 505, + 131.66666666666666 + ], + "spans": [], + "index": 3 + }, + { + "bbox": [ + 106, + 131.66666666666666, + 505, + 153.33333333333331 + ], + "spans": [], + "index": 4 + }, + { + "bbox": [ + 106, + 153.33333333333331, + 505, + 174.99999999999997 + ], + "spans": [], + "index": 5 + } + ] + } + ], + "index": 2.5 + }, + { + "type": "title", + "bbox": [ + 107, + 194, + 206, + 206 + ], + "lines": [ + { + "bbox": [ + 105, + 194, + 207, + 208 + ], + "spans": [ + { + "bbox": [ + 105, + 194, + 207, + 208 + ], + "score": 1.0, + "content": "4.2 Ablation Analysis", + "type": "text" + } + ], + "index": 6 + } + ], + "index": 6 + }, + { + "type": "text", + "bbox": [ + 106, + 214, + 505, + 313 + ], + "lines": [ + { + "bbox": [ + 105, + 213, + 506, + 227 + ], + "spans": [ + { + "bbox": [ + 105, + 213, + 506, + 227 + ], + "score": 1.0, + "content": "Model Settings: The default settings of the proposed LV-ViT are given in Table 1, where both token", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 106, + 226, + 505, + 237 + ], + "spans": [ + { + "bbox": [ + 106, + 226, + 505, + 237 + ], + "score": 1.0, + "content": "labeling and MixToken are used. A slight architecture modification to ViT [15] is that we replace the", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 105, + 236, + 506, + 250 + ], + "spans": [ + { + "bbox": [ + 105, + 236, + 506, + 250 + ], + "score": 1.0, + "content": "patch embedding module with a 4-layer convolution to better tokenize the input image and integrate", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 105, + 246, + 505, + 261 + ], + "spans": [ + { + "bbox": [ + 105, + 246, + 505, + 261 + ], + "score": 1.0, + "content": "local information. Detailed ablation about patch embedding can be found in our supplementary", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 258, + 506, + 272 + ], + "spans": [ + { + "bbox": [ + 105, + 258, + 318, + 272 + ], + "score": 1.0, + "content": "materials. As can be seen, our LV-ViT-T with only", + "type": "text" + }, + { + "bbox": [ + 319, + 258, + 343, + 269 + ], + "score": 0.38, + "content": "8 . 5 { \\bf M }", + "type": "inline_equation" + }, + { + "bbox": [ + 343, + 258, + 506, + 272 + ], + "score": 1.0, + "content": "parameters can already achieve a top-1", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 269, + 506, + 282 + ], + "spans": [ + { + "bbox": [ + 105, + 270, + 154, + 282 + ], + "score": 1.0, + "content": "accuracy of", + "type": "text" + }, + { + "bbox": [ + 154, + 269, + 181, + 280 + ], + "score": 0.87, + "content": "7 9 . 1 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 181, + 270, + 506, + 282 + ], + "score": 1.0, + "content": "on ImageNet. Increasing the embedding dimension and network depth can further", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 280, + 506, + 293 + ], + "spans": [ + { + "bbox": [ + 105, + 280, + 506, + 293 + ], + "score": 1.0, + "content": "boost the performance. More experiments compared to other methods can be found in Sec. 4.3. In", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 106, + 290, + 506, + 304 + ], + "spans": [ + { + "bbox": [ + 106, + 290, + 506, + 304 + ], + "score": 1.0, + "content": "the following ablation experiments, we will set our LV-ViT-S as baseline and show the advantages of", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 105, + 302, + 318, + 315 + ], + "spans": [ + { + "bbox": [ + 105, + 302, + 318, + 315 + ], + "score": 1.0, + "content": "the proposed token labeling and MixToken methods.", + "type": "text" + } + ], + "index": 15 + } + ], + "index": 11 + }, + { + "type": "text", + "bbox": [ + 106, + 318, + 505, + 384 + ], + "lines": [ + { + "bbox": [ + 106, + 318, + 506, + 330 + ], + "spans": [ + { + "bbox": [ + 106, + 318, + 506, + 330 + ], + "score": 1.0, + "content": "MixToken: We use MixToken as a substitution for CutMix while applying token labeling. Our", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 329, + 506, + 341 + ], + "spans": [ + { + "bbox": [ + 105, + 329, + 506, + 341 + ], + "score": 1.0, + "content": "experiments show that MixToken performs better than CutMix for token-based transformer models.", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 340, + 505, + 352 + ], + "spans": [ + { + "bbox": [ + 105, + 340, + 482, + 352 + ], + "score": 1.0, + "content": "As shown in Table 2, when training with the original ImageNet labels, using MixToken is", + "type": "text" + }, + { + "bbox": [ + 482, + 340, + 505, + 351 + ], + "score": 0.87, + "content": "0 . 1 \\%", + "type": "inline_equation" + } + ], + "index": 18 + }, + { + "bbox": [ + 105, + 350, + 506, + 363 + ], + "spans": [ + { + "bbox": [ + 105, + 350, + 506, + 363 + ], + "score": 1.0, + "content": "higher than using CutMix. When using the ReLabel supervision, we can also see an advantage of", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 106, + 361, + 506, + 375 + ], + "spans": [ + { + "bbox": [ + 106, + 362, + 129, + 373 + ], + "score": 0.87, + "content": "0 . { \\bar { 2 } } \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 129, + 361, + 506, + 375 + ], + "score": 1.0, + "content": "over the CutMix baseline. Combining with our token labeling, the performance can be further", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 105, + 373, + 174, + 384 + ], + "spans": [ + { + "bbox": [ + 105, + 373, + 143, + 384 + ], + "score": 1.0, + "content": "raised to", + "type": "text" + }, + { + "bbox": [ + 143, + 373, + 170, + 384 + ], + "score": 0.89, + "content": "8 3 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 171, + 373, + 174, + 384 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 18.5 + }, + { + "type": "table", + "bbox": [ + 106, + 444, + 289, + 518 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 399, + 288, + 444 + ], + "group_id": 1, + "lines": [ + { + "bbox": [ + 106, + 399, + 287, + 411 + ], + "spans": [ + { + "bbox": [ + 106, + 399, + 287, + 411 + ], + "score": 1.0, + "content": "Table 2: Ablation on the proposed MixToken", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 106, + 411, + 286, + 422 + ], + "spans": [ + { + "bbox": [ + 106, + 411, + 286, + 422 + ], + "score": 1.0, + "content": "and token labeling augmentations. We also", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 106, + 421, + 287, + 433 + ], + "spans": [ + { + "bbox": [ + 106, + 421, + 287, + 433 + ], + "score": 1.0, + "content": "show results with either the ImageNet hard", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 106, + 432, + 276, + 444 + ], + "spans": [ + { + "bbox": [ + 106, + 432, + 276, + 444 + ], + "score": 1.0, + "content": "label and the ReLabel [49] as supervision.", + "type": "text" + } + ], + "index": 28 + } + ], + "index": 25.0 + }, + { + "type": "table_body", + "bbox": [ + 106, + 444, + 289, + 518 + ], + "group_id": 1, + "lines": [ + { + "bbox": [ + 106, + 444, + 289, + 518 + ], + "spans": [ + { + "bbox": [ + 106, + 444, + 289, + 518 + ], + "score": 0.976, + "html": "
Aug. MethodSupervisionTop-1 Acc.
MixTokenToken labeling83.3
MixTokenReLabel83.0
CutMixReLabel82.8
MixtokenImageNet Label82.5
CutMixImageNet Label82.4
", + "type": "table", + "image_path": "ecc151313b3142fbf655b4065fd6e4b0c147ea2803b47139c3e90b472d0b2137.jpg" + } + ] + } + ], + "index": 34, + "virtual_lines": [ + { + "bbox": [ + 106, + 444, + 289, + 458.8 + ], + "spans": [], + "index": 30 + }, + { + "bbox": [ + 106, + 458.8, + 289, + 473.6 + ], + "spans": [], + "index": 32 + }, + { + "bbox": [ + 106, + 473.6, + 289, + 488.40000000000003 + ], + "spans": [], + "index": 34 + }, + { + "bbox": [ + 106, + 488.40000000000003, + 289, + 503.20000000000005 + ], + "spans": [], + "index": 36 + }, + { + "bbox": [ + 106, + 503.20000000000005, + 289, + 518.0 + ], + "spans": [], + "index": 38 + } + ] + } + ], + "index": 29.5 + }, + { + "type": "table", + "bbox": [ + 296, + 443, + 505, + 518 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 296, + 399, + 505, + 443 + ], + "group_id": 2, + "lines": [ + { + "bbox": [ + 295, + 398, + 506, + 412 + ], + "spans": [ + { + "bbox": [ + 295, + 398, + 506, + 412 + ], + "score": 1.0, + "content": "Table 3: Ablation on different widely-used data aug-", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 296, + 410, + 505, + 423 + ], + "spans": [ + { + "bbox": [ + 296, + 410, + 505, + 423 + ], + "score": 1.0, + "content": "mentations. We have empirically found our proposed", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 297, + 422, + 505, + 433 + ], + "spans": [ + { + "bbox": [ + 297, + 422, + 505, + 433 + ], + "score": 1.0, + "content": "MixToken performs even better than the combination", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 296, + 433, + 482, + 444 + ], + "spans": [ + { + "bbox": [ + 296, + 433, + 482, + 444 + ], + "score": 1.0, + "content": "of MixUp and CutMix in vision transformers.", + "type": "text" + } + ], + "index": 29 + } + ], + "index": 26.0 + }, + { + "type": "table_body", + "bbox": [ + 296, + 443, + 505, + 518 + ], + "group_id": 2, + "lines": [ + { + "bbox": [ + 296, + 443, + 505, + 518 + ], + "spans": [ + { + "bbox": [ + 296, + 443, + 505, + 518 + ], + "score": 0.976, + "html": "
MixTokenMixUpCutOutRandAugTop-1 Acc.
83.3
81.3
>x>xx>>>>>>>83.1
X√83.0
广XXX82.8
", + "type": "table", + "image_path": "51895f2ce4fa7182e488e57176934e6ab370cc3160e81c0730e961fdd370158d.jpg" + } + ] + } + ], + "index": 35, + "virtual_lines": [ + { + "bbox": [ + 296, + 443, + 505, + 458.0 + ], + "spans": [], + "index": 31 + }, + { + "bbox": [ + 296, + 458.0, + 505, + 473.0 + ], + "spans": [], + "index": 33 + }, + { + "bbox": [ + 296, + 473.0, + 505, + 488.0 + ], + "spans": [], + "index": 35 + }, + { + "bbox": [ + 296, + 488.0, + 505, + 503.0 + ], + "spans": [], + "index": 37 + }, + { + "bbox": [ + 296, + 503.0, + 505, + 518.0 + ], + "spans": [], + "index": 39 + } + ] + } + ], + "index": 30.5 + }, + { + "type": "text", + "bbox": [ + 106, + 530, + 505, + 630 + ], + "lines": [ + { + "bbox": [ + 105, + 529, + 506, + 543 + ], + "spans": [ + { + "bbox": [ + 105, + 529, + 506, + 543 + ], + "score": 1.0, + "content": "Data Augmentation: Here, we study the compatibility of MixToken with other augmentation", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 105, + 540, + 506, + 554 + ], + "spans": [ + { + "bbox": [ + 105, + 540, + 506, + 554 + ], + "score": 1.0, + "content": "techniques, such as MixUp [52], CutOut [57] and RandAug [10]. The ablation results are shown in", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 105, + 552, + 506, + 565 + ], + "spans": [ + { + "bbox": [ + 105, + 552, + 468, + 565 + ], + "score": 1.0, + "content": "Table 3. We can see when all the four augmentation methods are used, a top-1 accuracy of", + "type": "text" + }, + { + "bbox": [ + 468, + 552, + 495, + 563 + ], + "score": 0.88, + "content": "8 3 . 1 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 495, + 552, + 506, + 565 + ], + "score": 1.0, + "content": "is", + "type": "text" + } + ], + "index": 42 + }, + { + "bbox": [ + 105, + 563, + 506, + 576 + ], + "spans": [ + { + "bbox": [ + 105, + 563, + 506, + 576 + ], + "score": 1.0, + "content": "achieved. Interestingly, when the MixUp augmentation is removed, the performance can be improved", + "type": "text" + } + ], + "index": 43 + }, + { + "bbox": [ + 105, + 573, + 505, + 587 + ], + "spans": [ + { + "bbox": [ + 105, + 573, + 117, + 587 + ], + "score": 1.0, + "content": "to", + "type": "text" + }, + { + "bbox": [ + 118, + 574, + 145, + 585 + ], + "score": 0.87, + "content": "8 3 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 145, + 573, + 505, + 587 + ], + "score": 1.0, + "content": ". This may be explained as, using MixToken and MixUp at the same time would bring", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 106, + 586, + 505, + 597 + ], + "spans": [ + { + "bbox": [ + 106, + 586, + 505, + 597 + ], + "score": 1.0, + "content": "too much noise in the label, and consequently cause confusion of the model. Moreover, the CutOut", + "type": "text" + } + ], + "index": 45 + }, + { + "bbox": [ + 105, + 596, + 506, + 609 + ], + "spans": [ + { + "bbox": [ + 105, + 596, + 506, + 609 + ], + "score": 1.0, + "content": "augmentation, which randomly erases some parts of the image, is also effective and removing it", + "type": "text" + } + ], + "index": 46 + }, + { + "bbox": [ + 106, + 606, + 506, + 620 + ], + "spans": [ + { + "bbox": [ + 106, + 606, + 228, + 620 + ], + "score": 1.0, + "content": "brings a performance drop of", + "type": "text" + }, + { + "bbox": [ + 228, + 607, + 251, + 618 + ], + "score": 0.89, + "content": "\\dot { 0 } . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 251, + 606, + 506, + 620 + ], + "score": 1.0, + "content": ". Similarly, the RandAug augmentation also contributes to the", + "type": "text" + } + ], + "index": 47 + }, + { + "bbox": [ + 105, + 618, + 339, + 630 + ], + "spans": [ + { + "bbox": [ + 105, + 618, + 313, + 630 + ], + "score": 1.0, + "content": "performance and using it brings an improvement of", + "type": "text" + }, + { + "bbox": [ + 314, + 618, + 336, + 629 + ], + "score": 0.88, + "content": "0 . 5 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 336, + 618, + 339, + 630 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 48 + } + ], + "index": 44 + }, + { + "type": "text", + "bbox": [ + 106, + 634, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 105, + 633, + 505, + 647 + ], + "spans": [ + { + "bbox": [ + 105, + 633, + 505, + 647 + ], + "score": 1.0, + "content": "All Tokens Matter: To show the importance of involving all tokens in our token labeling method, we", + "type": "text" + } + ], + "index": 49 + }, + { + "bbox": [ + 105, + 645, + 505, + 658 + ], + "spans": [ + { + "bbox": [ + 105, + 645, + 505, + 658 + ], + "score": 1.0, + "content": "attempt to randomly drop some tokens and use the remaining ones for computing the token labeling", + "type": "text" + } + ], + "index": 50 + }, + { + "bbox": [ + 106, + 656, + 505, + 668 + ], + "spans": [ + { + "bbox": [ + 106, + 656, + 505, + 668 + ], + "score": 1.0, + "content": "loss. The percentage of the remaining tokens is denoted as Token Participation Rate. As shown in", + "type": "text" + } + ], + "index": 51 + }, + { + "bbox": [ + 106, + 667, + 506, + 680 + ], + "spans": [ + { + "bbox": [ + 106, + 667, + 506, + 680 + ], + "score": 1.0, + "content": "Figure 4 (Left), we conduct experiments on two models: LV-ViT-S and LV-ViT-M. As can be seen,", + "type": "text" + } + ], + "index": 52 + }, + { + "bbox": [ + 106, + 678, + 505, + 690 + ], + "spans": [ + { + "bbox": [ + 106, + 678, + 150, + 690 + ], + "score": 1.0, + "content": "using only", + "type": "text" + }, + { + "bbox": [ + 151, + 678, + 170, + 689 + ], + "score": 0.89, + "content": "2 0 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 171, + 678, + 471, + 690 + ], + "score": 1.0, + "content": "of the tokens to compute the token labeling loss decreases the performance", + "type": "text" + }, + { + "bbox": [ + 471, + 678, + 505, + 689 + ], + "score": 0.87, + "content": "( - 0 . 5 \\%", + "type": "inline_equation" + } + ], + "index": 53 + }, + { + "bbox": [ + 105, + 688, + 505, + 701 + ], + "spans": [ + { + "bbox": [ + 105, + 688, + 179, + 701 + ], + "score": 1.0, + "content": "for LV-ViT-S and", + "type": "text" + }, + { + "bbox": [ + 180, + 689, + 209, + 700 + ], + "score": 0.91, + "content": "- 0 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 209, + 688, + 505, + 701 + ], + "score": 1.0, + "content": "for LV-ViT-M). Involving more tokens for loss computation consistently", + "type": "text" + } + ], + "index": 54 + }, + { + "bbox": [ + 105, + 699, + 505, + 713 + ], + "spans": [ + { + "bbox": [ + 105, + 699, + 505, + 713 + ], + "score": 1.0, + "content": "leads to better performance. Since involving all tokens brings negligible computation cost and gives", + "type": "text" + } + ], + "index": 55 + }, + { + "bbox": [ + 106, + 711, + 506, + 723 + ], + "spans": [ + { + "bbox": [ + 106, + 711, + 365, + 723 + ], + "score": 1.0, + "content": "the best performance, we always set the token participation rate as", + "type": "text" + }, + { + "bbox": [ + 365, + 711, + 390, + 721 + ], + "score": 0.89, + "content": "\\mathrm { \\bar { 1 0 0 \\% } }", + "type": "inline_equation" + }, + { + "bbox": [ + 390, + 711, + 506, + 723 + ], + "score": 1.0, + "content": "in the following experiments.", + "type": "text" + } + ], + "index": 56 + } + ], + "index": 52.5 + } + ], + "page_idx": 5, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 302, + 742, + 309, + 750 + ], + "lines": [ + { + "bbox": [ + 302, + 741, + 310, + 752 + ], + "spans": [ + { + "bbox": [ + 302, + 741, + 310, + 752 + ], + "score": 1.0, + "content": "6", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "table", + "bbox": [ + 106, + 110, + 505, + 175 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 77, + 505, + 110 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 104, + 76, + 505, + 90 + ], + "spans": [ + { + "bbox": [ + 104, + 76, + 505, + 90 + ], + "score": 1.0, + "content": "Table 1: Performance of the proposed LV-ViT with different model sizes. Here, ‘depth’ denotes", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 105, + 88, + 505, + 100 + ], + "spans": [ + { + "bbox": [ + 105, + 88, + 505, + 100 + ], + "score": 1.0, + "content": "the number of transformer blocks used in different models. By default, the test resolution is set to", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 106, + 99, + 310, + 111 + ], + "spans": [ + { + "bbox": [ + 106, + 99, + 150, + 110 + ], + "score": 0.8, + "content": "2 2 4 \\times 2 2 4", + "type": "inline_equation" + }, + { + "bbox": [ + 150, + 99, + 263, + 111 + ], + "score": 1.0, + "content": "except the last one which is", + "type": "text" + }, + { + "bbox": [ + 263, + 99, + 306, + 110 + ], + "score": 0.88, + "content": "2 8 8 \\times 2 8 8", + "type": "inline_equation" + }, + { + "bbox": [ + 307, + 99, + 310, + 111 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 2 + } + ], + "index": 1 + }, + { + "type": "table_body", + "bbox": [ + 106, + 110, + 505, + 175 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 106, + 110, + 505, + 175 + ], + "spans": [ + { + "bbox": [ + 106, + 110, + 505, + 175 + ], + "score": 0.977, + "html": "
NameDepthEmbed dim.MLP Ratio#Heads#ParamsThroughput (im/s)Test sizeTop-1 Acc. (%)
LV-ViT-T122403.048.5M2032.622479.1
LV-ViT-S163843.0626M1018.222483.3
LV-ViT-M205123.0856M668.922484.1
LV-ViT-L247683.012150M204.828885.3
", + "type": "table", + "image_path": "8ab634b915f17e8f746292568270ed165dd5bacd94d1fed7927449e1328c4232.jpg" + } + ] + } + ], + "index": 4, + "virtual_lines": [ + { + "bbox": [ + 106, + 110, + 505, + 131.66666666666666 + ], + "spans": [], + "index": 3 + }, + { + "bbox": [ + 106, + 131.66666666666666, + 505, + 153.33333333333331 + ], + "spans": [], + "index": 4 + }, + { + "bbox": [ + 106, + 153.33333333333331, + 505, + 174.99999999999997 + ], + "spans": [], + "index": 5 + } + ] + } + ], + "index": 2.5 + }, + { + "type": "title", + "bbox": [ + 107, + 194, + 206, + 206 + ], + "lines": [ + { + "bbox": [ + 105, + 194, + 207, + 208 + ], + "spans": [ + { + "bbox": [ + 105, + 194, + 207, + 208 + ], + "score": 1.0, + "content": "4.2 Ablation Analysis", + "type": "text" + } + ], + "index": 6 + } + ], + "index": 6 + }, + { + "type": "text", + "bbox": [ + 106, + 214, + 505, + 313 + ], + "lines": [ + { + "bbox": [ + 105, + 213, + 506, + 227 + ], + "spans": [ + { + "bbox": [ + 105, + 213, + 506, + 227 + ], + "score": 1.0, + "content": "Model Settings: The default settings of the proposed LV-ViT are given in Table 1, where both token", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 106, + 226, + 505, + 237 + ], + "spans": [ + { + "bbox": [ + 106, + 226, + 505, + 237 + ], + "score": 1.0, + "content": "labeling and MixToken are used. A slight architecture modification to ViT [15] is that we replace the", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 105, + 236, + 506, + 250 + ], + "spans": [ + { + "bbox": [ + 105, + 236, + 506, + 250 + ], + "score": 1.0, + "content": "patch embedding module with a 4-layer convolution to better tokenize the input image and integrate", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 105, + 246, + 505, + 261 + ], + "spans": [ + { + "bbox": [ + 105, + 246, + 505, + 261 + ], + "score": 1.0, + "content": "local information. Detailed ablation about patch embedding can be found in our supplementary", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 258, + 506, + 272 + ], + "spans": [ + { + "bbox": [ + 105, + 258, + 318, + 272 + ], + "score": 1.0, + "content": "materials. As can be seen, our LV-ViT-T with only", + "type": "text" + }, + { + "bbox": [ + 319, + 258, + 343, + 269 + ], + "score": 0.38, + "content": "8 . 5 { \\bf M }", + "type": "inline_equation" + }, + { + "bbox": [ + 343, + 258, + 506, + 272 + ], + "score": 1.0, + "content": "parameters can already achieve a top-1", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 269, + 506, + 282 + ], + "spans": [ + { + "bbox": [ + 105, + 270, + 154, + 282 + ], + "score": 1.0, + "content": "accuracy of", + "type": "text" + }, + { + "bbox": [ + 154, + 269, + 181, + 280 + ], + "score": 0.87, + "content": "7 9 . 1 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 181, + 270, + 506, + 282 + ], + "score": 1.0, + "content": "on ImageNet. Increasing the embedding dimension and network depth can further", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 280, + 506, + 293 + ], + "spans": [ + { + "bbox": [ + 105, + 280, + 506, + 293 + ], + "score": 1.0, + "content": "boost the performance. More experiments compared to other methods can be found in Sec. 4.3. In", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 106, + 290, + 506, + 304 + ], + "spans": [ + { + "bbox": [ + 106, + 290, + 506, + 304 + ], + "score": 1.0, + "content": "the following ablation experiments, we will set our LV-ViT-S as baseline and show the advantages of", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 105, + 302, + 318, + 315 + ], + "spans": [ + { + "bbox": [ + 105, + 302, + 318, + 315 + ], + "score": 1.0, + "content": "the proposed token labeling and MixToken methods.", + "type": "text" + } + ], + "index": 15 + } + ], + "index": 11, + "bbox_fs": [ + 105, + 213, + 506, + 315 + ] + }, + { + "type": "text", + "bbox": [ + 106, + 318, + 505, + 384 + ], + "lines": [ + { + "bbox": [ + 106, + 318, + 506, + 330 + ], + "spans": [ + { + "bbox": [ + 106, + 318, + 506, + 330 + ], + "score": 1.0, + "content": "MixToken: We use MixToken as a substitution for CutMix while applying token labeling. Our", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 329, + 506, + 341 + ], + "spans": [ + { + "bbox": [ + 105, + 329, + 506, + 341 + ], + "score": 1.0, + "content": "experiments show that MixToken performs better than CutMix for token-based transformer models.", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 340, + 505, + 352 + ], + "spans": [ + { + "bbox": [ + 105, + 340, + 482, + 352 + ], + "score": 1.0, + "content": "As shown in Table 2, when training with the original ImageNet labels, using MixToken is", + "type": "text" + }, + { + "bbox": [ + 482, + 340, + 505, + 351 + ], + "score": 0.87, + "content": "0 . 1 \\%", + "type": "inline_equation" + } + ], + "index": 18 + }, + { + "bbox": [ + 105, + 350, + 506, + 363 + ], + "spans": [ + { + "bbox": [ + 105, + 350, + 506, + 363 + ], + "score": 1.0, + "content": "higher than using CutMix. When using the ReLabel supervision, we can also see an advantage of", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 106, + 361, + 506, + 375 + ], + "spans": [ + { + "bbox": [ + 106, + 362, + 129, + 373 + ], + "score": 0.87, + "content": "0 . { \\bar { 2 } } \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 129, + 361, + 506, + 375 + ], + "score": 1.0, + "content": "over the CutMix baseline. Combining with our token labeling, the performance can be further", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 105, + 373, + 174, + 384 + ], + "spans": [ + { + "bbox": [ + 105, + 373, + 143, + 384 + ], + "score": 1.0, + "content": "raised to", + "type": "text" + }, + { + "bbox": [ + 143, + 373, + 170, + 384 + ], + "score": 0.89, + "content": "8 3 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 171, + 373, + 174, + 384 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 18.5, + "bbox_fs": [ + 105, + 318, + 506, + 384 + ] + }, + { + "type": "table", + "bbox": [ + 106, + 444, + 289, + 518 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 399, + 288, + 444 + ], + "group_id": 1, + "lines": [ + { + "bbox": [ + 106, + 399, + 287, + 411 + ], + "spans": [ + { + "bbox": [ + 106, + 399, + 287, + 411 + ], + "score": 1.0, + "content": "Table 2: Ablation on the proposed MixToken", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 106, + 411, + 286, + 422 + ], + "spans": [ + { + "bbox": [ + 106, + 411, + 286, + 422 + ], + "score": 1.0, + "content": "and token labeling augmentations. We also", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 106, + 421, + 287, + 433 + ], + "spans": [ + { + "bbox": [ + 106, + 421, + 287, + 433 + ], + "score": 1.0, + "content": "show results with either the ImageNet hard", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 106, + 432, + 276, + 444 + ], + "spans": [ + { + "bbox": [ + 106, + 432, + 276, + 444 + ], + "score": 1.0, + "content": "label and the ReLabel [49] as supervision.", + "type": "text" + } + ], + "index": 28 + } + ], + "index": 25.0 + }, + { + "type": "table_body", + "bbox": [ + 106, + 444, + 289, + 518 + ], + "group_id": 1, + "lines": [ + { + "bbox": [ + 106, + 444, + 289, + 518 + ], + "spans": [ + { + "bbox": [ + 106, + 444, + 289, + 518 + ], + "score": 0.976, + "html": "
Aug. MethodSupervisionTop-1 Acc.
MixTokenToken labeling83.3
MixTokenReLabel83.0
CutMixReLabel82.8
MixtokenImageNet Label82.5
CutMixImageNet Label82.4
", + "type": "table", + "image_path": "ecc151313b3142fbf655b4065fd6e4b0c147ea2803b47139c3e90b472d0b2137.jpg" + } + ] + } + ], + "index": 34, + "virtual_lines": [ + { + "bbox": [ + 106, + 444, + 289, + 458.8 + ], + "spans": [], + "index": 30 + }, + { + "bbox": [ + 106, + 458.8, + 289, + 473.6 + ], + "spans": [], + "index": 32 + }, + { + "bbox": [ + 106, + 473.6, + 289, + 488.40000000000003 + ], + "spans": [], + "index": 34 + }, + { + "bbox": [ + 106, + 488.40000000000003, + 289, + 503.20000000000005 + ], + "spans": [], + "index": 36 + }, + { + "bbox": [ + 106, + 503.20000000000005, + 289, + 518.0 + ], + "spans": [], + "index": 38 + } + ] + } + ], + "index": 29.5 + }, + { + "type": "table", + "bbox": [ + 296, + 443, + 505, + 518 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 296, + 399, + 505, + 443 + ], + "group_id": 2, + "lines": [ + { + "bbox": [ + 295, + 398, + 506, + 412 + ], + "spans": [ + { + "bbox": [ + 295, + 398, + 506, + 412 + ], + "score": 1.0, + "content": "Table 3: Ablation on different widely-used data aug-", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 296, + 410, + 505, + 423 + ], + "spans": [ + { + "bbox": [ + 296, + 410, + 505, + 423 + ], + "score": 1.0, + "content": "mentations. We have empirically found our proposed", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 297, + 422, + 505, + 433 + ], + "spans": [ + { + "bbox": [ + 297, + 422, + 505, + 433 + ], + "score": 1.0, + "content": "MixToken performs even better than the combination", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 296, + 433, + 482, + 444 + ], + "spans": [ + { + "bbox": [ + 296, + 433, + 482, + 444 + ], + "score": 1.0, + "content": "of MixUp and CutMix in vision transformers.", + "type": "text" + } + ], + "index": 29 + } + ], + "index": 26.0 + }, + { + "type": "table_body", + "bbox": [ + 296, + 443, + 505, + 518 + ], + "group_id": 2, + "lines": [ + { + "bbox": [ + 296, + 443, + 505, + 518 + ], + "spans": [ + { + "bbox": [ + 296, + 443, + 505, + 518 + ], + "score": 0.976, + "html": "
MixTokenMixUpCutOutRandAugTop-1 Acc.
83.3
81.3
>x>xx>>>>>>>83.1
X√83.0
广XXX82.8
", + "type": "table", + "image_path": "51895f2ce4fa7182e488e57176934e6ab370cc3160e81c0730e961fdd370158d.jpg" + } + ] + } + ], + "index": 35, + "virtual_lines": [ + { + "bbox": [ + 296, + 443, + 505, + 458.0 + ], + "spans": [], + "index": 31 + }, + { + "bbox": [ + 296, + 458.0, + 505, + 473.0 + ], + "spans": [], + "index": 33 + }, + { + "bbox": [ + 296, + 473.0, + 505, + 488.0 + ], + "spans": [], + "index": 35 + }, + { + "bbox": [ + 296, + 488.0, + 505, + 503.0 + ], + "spans": [], + "index": 37 + }, + { + "bbox": [ + 296, + 503.0, + 505, + 518.0 + ], + "spans": [], + "index": 39 + } + ] + } + ], + "index": 30.5 + }, + { + "type": "text", + "bbox": [ + 106, + 530, + 505, + 630 + ], + "lines": [ + { + "bbox": [ + 105, + 529, + 506, + 543 + ], + "spans": [ + { + "bbox": [ + 105, + 529, + 506, + 543 + ], + "score": 1.0, + "content": "Data Augmentation: Here, we study the compatibility of MixToken with other augmentation", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 105, + 540, + 506, + 554 + ], + "spans": [ + { + "bbox": [ + 105, + 540, + 506, + 554 + ], + "score": 1.0, + "content": "techniques, such as MixUp [52], CutOut [57] and RandAug [10]. The ablation results are shown in", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 105, + 552, + 506, + 565 + ], + "spans": [ + { + "bbox": [ + 105, + 552, + 468, + 565 + ], + "score": 1.0, + "content": "Table 3. We can see when all the four augmentation methods are used, a top-1 accuracy of", + "type": "text" + }, + { + "bbox": [ + 468, + 552, + 495, + 563 + ], + "score": 0.88, + "content": "8 3 . 1 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 495, + 552, + 506, + 565 + ], + "score": 1.0, + "content": "is", + "type": "text" + } + ], + "index": 42 + }, + { + "bbox": [ + 105, + 563, + 506, + 576 + ], + "spans": [ + { + "bbox": [ + 105, + 563, + 506, + 576 + ], + "score": 1.0, + "content": "achieved. Interestingly, when the MixUp augmentation is removed, the performance can be improved", + "type": "text" + } + ], + "index": 43 + }, + { + "bbox": [ + 105, + 573, + 505, + 587 + ], + "spans": [ + { + "bbox": [ + 105, + 573, + 117, + 587 + ], + "score": 1.0, + "content": "to", + "type": "text" + }, + { + "bbox": [ + 118, + 574, + 145, + 585 + ], + "score": 0.87, + "content": "8 3 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 145, + 573, + 505, + 587 + ], + "score": 1.0, + "content": ". This may be explained as, using MixToken and MixUp at the same time would bring", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 106, + 586, + 505, + 597 + ], + "spans": [ + { + "bbox": [ + 106, + 586, + 505, + 597 + ], + "score": 1.0, + "content": "too much noise in the label, and consequently cause confusion of the model. Moreover, the CutOut", + "type": "text" + } + ], + "index": 45 + }, + { + "bbox": [ + 105, + 596, + 506, + 609 + ], + "spans": [ + { + "bbox": [ + 105, + 596, + 506, + 609 + ], + "score": 1.0, + "content": "augmentation, which randomly erases some parts of the image, is also effective and removing it", + "type": "text" + } + ], + "index": 46 + }, + { + "bbox": [ + 106, + 606, + 506, + 620 + ], + "spans": [ + { + "bbox": [ + 106, + 606, + 228, + 620 + ], + "score": 1.0, + "content": "brings a performance drop of", + "type": "text" + }, + { + "bbox": [ + 228, + 607, + 251, + 618 + ], + "score": 0.89, + "content": "\\dot { 0 } . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 251, + 606, + 506, + 620 + ], + "score": 1.0, + "content": ". Similarly, the RandAug augmentation also contributes to the", + "type": "text" + } + ], + "index": 47 + }, + { + "bbox": [ + 105, + 618, + 339, + 630 + ], + "spans": [ + { + "bbox": [ + 105, + 618, + 313, + 630 + ], + "score": 1.0, + "content": "performance and using it brings an improvement of", + "type": "text" + }, + { + "bbox": [ + 314, + 618, + 336, + 629 + ], + "score": 0.88, + "content": "0 . 5 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 336, + 618, + 339, + 630 + ], + "score": 1.0, + "content": ".", + "type": "text" + } + ], + "index": 48 + } + ], + "index": 44, + "bbox_fs": [ + 105, + 529, + 506, + 630 + ] + }, + { + "type": "text", + "bbox": [ + 106, + 634, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 105, + 633, + 505, + 647 + ], + "spans": [ + { + "bbox": [ + 105, + 633, + 505, + 647 + ], + "score": 1.0, + "content": "All Tokens Matter: To show the importance of involving all tokens in our token labeling method, we", + "type": "text" + } + ], + "index": 49 + }, + { + "bbox": [ + 105, + 645, + 505, + 658 + ], + "spans": [ + { + "bbox": [ + 105, + 645, + 505, + 658 + ], + "score": 1.0, + "content": "attempt to randomly drop some tokens and use the remaining ones for computing the token labeling", + "type": "text" + } + ], + "index": 50 + }, + { + "bbox": [ + 106, + 656, + 505, + 668 + ], + "spans": [ + { + "bbox": [ + 106, + 656, + 505, + 668 + ], + "score": 1.0, + "content": "loss. The percentage of the remaining tokens is denoted as Token Participation Rate. As shown in", + "type": "text" + } + ], + "index": 51 + }, + { + "bbox": [ + 106, + 667, + 506, + 680 + ], + "spans": [ + { + "bbox": [ + 106, + 667, + 506, + 680 + ], + "score": 1.0, + "content": "Figure 4 (Left), we conduct experiments on two models: LV-ViT-S and LV-ViT-M. As can be seen,", + "type": "text" + } + ], + "index": 52 + }, + { + "bbox": [ + 106, + 678, + 505, + 690 + ], + "spans": [ + { + "bbox": [ + 106, + 678, + 150, + 690 + ], + "score": 1.0, + "content": "using only", + "type": "text" + }, + { + "bbox": [ + 151, + 678, + 170, + 689 + ], + "score": 0.89, + "content": "2 0 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 171, + 678, + 471, + 690 + ], + "score": 1.0, + "content": "of the tokens to compute the token labeling loss decreases the performance", + "type": "text" + }, + { + "bbox": [ + 471, + 678, + 505, + 689 + ], + "score": 0.87, + "content": "( - 0 . 5 \\%", + "type": "inline_equation" + } + ], + "index": 53 + }, + { + "bbox": [ + 105, + 688, + 505, + 701 + ], + "spans": [ + { + "bbox": [ + 105, + 688, + 179, + 701 + ], + "score": 1.0, + "content": "for LV-ViT-S and", + "type": "text" + }, + { + "bbox": [ + 180, + 689, + 209, + 700 + ], + "score": 0.91, + "content": "- 0 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 209, + 688, + 505, + 701 + ], + "score": 1.0, + "content": "for LV-ViT-M). Involving more tokens for loss computation consistently", + "type": "text" + } + ], + "index": 54 + }, + { + "bbox": [ + 105, + 699, + 505, + 713 + ], + "spans": [ + { + "bbox": [ + 105, + 699, + 505, + 713 + ], + "score": 1.0, + "content": "leads to better performance. Since involving all tokens brings negligible computation cost and gives", + "type": "text" + } + ], + "index": 55 + }, + { + "bbox": [ + 106, + 711, + 506, + 723 + ], + "spans": [ + { + "bbox": [ + 106, + 711, + 365, + 723 + ], + "score": 1.0, + "content": "the best performance, we always set the token participation rate as", + "type": "text" + }, + { + "bbox": [ + 365, + 711, + 390, + 721 + ], + "score": 0.89, + "content": "\\mathrm { \\bar { 1 0 0 \\% } }", + "type": "inline_equation" + }, + { + "bbox": [ + 390, + 711, + 506, + 723 + ], + "score": 1.0, + "content": "in the following experiments.", + "type": "text" + } + ], + "index": 56 + } + ], + "index": 52.5, + "bbox_fs": [ + 105, + 633, + 506, + 723 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "image", + "bbox": [ + 109, + 82, + 484, + 218 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 109, + 82, + 484, + 218 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 109, + 82, + 484, + 218 + ], + "spans": [ + { + "bbox": [ + 109, + 82, + 484, + 218 + ], + "score": 0.916, + "type": "image", + "image_path": "810578c2e36580b46aa98e4ec01030b20f6e00994c33dafcaa4cde435beb4d46.jpg" + } + ] + } + ], + "index": 1, + "virtual_lines": [ + { + "bbox": [ + 109, + 82, + 484, + 127.33333333333334 + ], + "spans": [], + "index": 0 + }, + { + "bbox": [ + 109, + 127.33333333333334, + 484, + 172.66666666666669 + ], + "spans": [], + "index": 1 + }, + { + "bbox": [ + 109, + 172.66666666666669, + 484, + 218.00000000000003 + ], + "spans": [], + "index": 2 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 106, + 223, + 506, + 288 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 104, + 222, + 506, + 236 + ], + "spans": [ + { + "bbox": [ + 104, + 222, + 506, + 236 + ], + "score": 1.0, + "content": "Figure 4: Left: LV-ViT ImageNet Top-1 Accuracy w.r.t. the token participation rate while applying", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 105, + 233, + 506, + 247 + ], + "spans": [ + { + "bbox": [ + 105, + 233, + 506, + 247 + ], + "score": 1.0, + "content": "token labeling. Token participation rate indicates the percentage of patch tokens involved in computing", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 244, + 506, + 257 + ], + "spans": [ + { + "bbox": [ + 105, + 244, + 506, + 257 + ], + "score": 1.0, + "content": "the token labeling loss. This experiment reflects that all tokens matter for vision transformers. Right:", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 106, + 256, + 505, + 267 + ], + "spans": [ + { + "bbox": [ + 106, + 256, + 505, + 267 + ], + "score": 1.0, + "content": "LV-ViT-S ImageNet Top-1 Accuracy w.r.t. different annotator models. The point size indicates the", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 104, + 266, + 506, + 279 + ], + "spans": [ + { + "bbox": [ + 104, + 266, + 506, + 279 + ], + "score": 1.0, + "content": "parameter number of the annotator model. Clearly, our token labeling objective is robust to different", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 105, + 278, + 180, + 288 + ], + "spans": [ + { + "bbox": [ + 105, + 278, + 180, + 288 + ], + "score": 1.0, + "content": "annotator models.", + "type": "text" + } + ], + "index": 8 + } + ], + "index": 5.5 + } + ], + "index": 3.25 + }, + { + "type": "table", + "bbox": [ + 128, + 342, + 483, + 411 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 108, + 308, + 505, + 341 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 307, + 505, + 321 + ], + "spans": [ + { + "bbox": [ + 105, + 307, + 505, + 321 + ], + "score": 1.0, + "content": "Table 4: Comparison of token labeling (TL), knowledge distillation (KD) based method and ReLabel", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 106, + 319, + 505, + 331 + ], + "spans": [ + { + "bbox": [ + 106, + 319, + 505, + 331 + ], + "score": 1.0, + "content": "method based on utilized tokens, DeiT-S/LV-ViT-S Top-1 accuracy on ImageNet validation set and", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 106, + 330, + 278, + 343 + ], + "spans": [ + { + "bbox": [ + 106, + 330, + 278, + 343 + ], + "score": 1.0, + "content": "training time on a single V100 GPU node.", + "type": "text" + } + ], + "index": 11 + } + ], + "index": 10 + }, + { + "type": "table_body", + "bbox": [ + 128, + 342, + 483, + 411 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 128, + 342, + 483, + 411 + ], + "spans": [ + { + "bbox": [ + 128, + 342, + 483, + 411 + ], + "score": 0.979, + "html": "
MethodOnline KDOnline TLTLReLabelVanilla
Tokens Utilized2AllAll11
DeiT-S Acc. (%)81.281.881.080.479.9
LV-ViT-S Acc. (%)83.083.583.382.882.4
Training Time (8× V100)63 hrs63 hrs45 hrs45 hrs41 hrs
", + "type": "table", + "image_path": "ae9fdf37e5d602133531450f1b69751afa30b0104c5bff5b6433f7a76be69066.jpg" + } + ] + } + ], + "index": 13, + "virtual_lines": [ + { + "bbox": [ + 128, + 342, + 483, + 365.0 + ], + "spans": [], + "index": 12 + }, + { + "bbox": [ + 128, + 365.0, + 483, + 388.0 + ], + "spans": [], + "index": 13 + }, + { + "bbox": [ + 128, + 388.0, + 483, + 411.0 + ], + "spans": [], + "index": 14 + } + ] + } + ], + "index": 11.5 + }, + { + "type": "text", + "bbox": [ + 106, + 432, + 505, + 585 + ], + "lines": [ + { + "bbox": [ + 105, + 431, + 505, + 445 + ], + "spans": [ + { + "bbox": [ + 105, + 431, + 505, + 445 + ], + "score": 1.0, + "content": "Online Token Labeling: Unlike the online knowledge distillation method which generates labels by", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 105, + 443, + 505, + 456 + ], + "spans": [ + { + "bbox": [ + 105, + 443, + 505, + 456 + ], + "score": 1.0, + "content": "a teacher model online, our token labeling approach utilizes the dense label map generated in advance", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 106, + 455, + 505, + 467 + ], + "spans": [ + { + "bbox": [ + 106, + 455, + 505, + 467 + ], + "score": 1.0, + "content": "and directly applies the corresponding augmentation methods, such as random crop, on the label", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 464, + 506, + 478 + ], + "spans": [ + { + "bbox": [ + 105, + 464, + 506, + 478 + ], + "score": 1.0, + "content": "map to obtain token-level labels. To directly compare with the online knowledge distillation based", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 105, + 475, + 506, + 489 + ], + "spans": [ + { + "bbox": [ + 105, + 475, + 506, + 489 + ], + "score": 1.0, + "content": "method and validate the effectiveness of token-level supervision, we further conduct experiments on", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 105, + 486, + 505, + 501 + ], + "spans": [ + { + "bbox": [ + 105, + 486, + 505, + 501 + ], + "score": 1.0, + "content": "the online version of our token labeling method, which generates token-level labels online during", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 106, + 498, + 506, + 510 + ], + "spans": [ + { + "bbox": [ + 106, + 498, + 506, + 510 + ], + "score": 1.0, + "content": "training. Following DeiT [36], we use RegNetY-16GF [30] as the online teacher model. Results in", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 105, + 508, + 505, + 522 + ], + "spans": [ + { + "bbox": [ + 105, + 508, + 505, + 522 + ], + "score": 1.0, + "content": "terms of DeiT-S/LV-ViT-S Top-1 accuracy and training time for our token labeling, online knowledge", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 519, + 506, + 532 + ], + "spans": [ + { + "bbox": [ + 105, + 519, + 506, + 532 + ], + "score": 1.0, + "content": "distillation, and ReLabel [49] are listed in Table 4, with number of utilized tokens also included for", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 531, + 506, + 543 + ], + "spans": [ + { + "bbox": [ + 105, + 531, + 506, + 543 + ], + "score": 1.0, + "content": "clear comparison. As can be seen, for both online and offline cases, using token-level supervision can", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 105, + 541, + 506, + 555 + ], + "spans": [ + { + "bbox": [ + 105, + 541, + 506, + 555 + ], + "score": 1.0, + "content": "improve the overall performance with only negligible additional training cost. Meanwhile, compared", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 106, + 553, + 505, + 565 + ], + "spans": [ + { + "bbox": [ + 106, + 553, + 505, + 565 + ], + "score": 1.0, + "content": "to the vanilla training baseline, our proposed offline token labeling brings almost no additional", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 106, + 564, + 505, + 576 + ], + "spans": [ + { + "bbox": [ + 106, + 564, + 366, + 576 + ], + "score": 1.0, + "content": "training cost, and boosts the overall performance of LV-ViT-S by", + "type": "text" + }, + { + "bbox": [ + 366, + 564, + 388, + 574 + ], + "score": 0.86, + "content": "0 . 9 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 388, + 564, + 505, + 576 + ], + "score": 1.0, + "content": ", which well demonstrates its", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 106, + 574, + 221, + 587 + ], + "spans": [ + { + "bbox": [ + 106, + 574, + 221, + 587 + ], + "score": 1.0, + "content": "efficiency and effectiveness.", + "type": "text" + } + ], + "index": 28 + } + ], + "index": 21.5 + }, + { + "type": "text", + "bbox": [ + 107, + 591, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 106, + 591, + 505, + 603 + ], + "spans": [ + { + "bbox": [ + 106, + 591, + 505, + 603 + ], + "score": 1.0, + "content": "Robustness to Different Annotators: To evaluate the robustness of our token labeling method, we", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 106, + 602, + 505, + 613 + ], + "spans": [ + { + "bbox": [ + 106, + 602, + 505, + 613 + ], + "score": 1.0, + "content": "use different pretrained CNNs, including EfficientNet-B3,B4,B5,B6,B7,B8 [34], NFNet-F6 [3] and", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 612, + 505, + 625 + ], + "spans": [ + { + "bbox": [ + 105, + 612, + 505, + 625 + ], + "score": 1.0, + "content": "ResNest269E [51], as annotator models to provide dense supervision. Results are shown in the right", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 623, + 506, + 636 + ], + "spans": [ + { + "bbox": [ + 105, + 623, + 506, + 636 + ], + "score": 1.0, + "content": "part of Figure 4. We can see that, even if we use an annotator with relatively lower performance, such", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 106, + 635, + 505, + 647 + ], + "spans": [ + { + "bbox": [ + 106, + 635, + 284, + 647 + ], + "score": 1.0, + "content": "as EfficientNet-B3 whose Top-1 accuracy is", + "type": "text" + }, + { + "bbox": [ + 285, + 635, + 311, + 645 + ], + "score": 0.89, + "content": "8 1 . 6 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 312, + 635, + 505, + 647 + ], + "score": 1.0, + "content": ", it can still provide multi-label location-specific", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 646, + 505, + 657 + ], + "spans": [ + { + "bbox": [ + 105, + 646, + 505, + 657 + ], + "score": 1.0, + "content": "supervision and help improve the performance of our LV-ViT-S model. Meanwhile, annotator models", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 105, + 655, + 506, + 669 + ], + "spans": [ + { + "bbox": [ + 105, + 655, + 506, + 669 + ], + "score": 1.0, + "content": "with better performance can provide more accurate supervision, bringing even better performance, as", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 105, + 666, + 506, + 680 + ], + "spans": [ + { + "bbox": [ + 105, + 666, + 506, + 680 + ], + "score": 1.0, + "content": "stronger annotator models can generate better token-level labels. The largest annotator NFNet-F6 [3],", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 106, + 678, + 506, + 690 + ], + "spans": [ + { + "bbox": [ + 106, + 678, + 244, + 690 + ], + "score": 1.0, + "content": "which has the best performance of", + "type": "text" + }, + { + "bbox": [ + 244, + 678, + 271, + 689 + ], + "score": 0.89, + "content": "8 6 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 271, + 678, + 506, + 690 + ], + "score": 1.0, + "content": ", allows us to achieve the best result for LV-ViT-S, which is", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 106, + 689, + 506, + 701 + ], + "spans": [ + { + "bbox": [ + 106, + 689, + 134, + 700 + ], + "score": 0.87, + "content": "8 3 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 134, + 689, + 506, + 701 + ], + "score": 1.0, + "content": ". In addition, we also attempt to use a better model, EfficientNet-L2 pretrained on JFT-300M", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 105, + 699, + 505, + 712 + ], + "spans": [ + { + "bbox": [ + 105, + 699, + 230, + 712 + ], + "score": 1.0, + "content": "as described in [49] which has", + "type": "text" + }, + { + "bbox": [ + 230, + 700, + 257, + 711 + ], + "score": 0.88, + "content": "8 8 . 2 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 257, + 699, + 505, + 712 + ], + "score": 1.0, + "content": "Top-1 ImageNet accuracy, as our annotator. The performance", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 106, + 711, + 505, + 723 + ], + "spans": [ + { + "bbox": [ + 106, + 711, + 264, + 723 + ], + "score": 1.0, + "content": "of LV-ViT-S can be further improved to", + "type": "text" + }, + { + "bbox": [ + 265, + 711, + 292, + 721 + ], + "score": 0.88, + "content": "8 \\mathrm { { 3 . 5 \\% } }", + "type": "inline_equation" + }, + { + "bbox": [ + 292, + 711, + 505, + 723 + ], + "score": 1.0, + "content": ". However, to fairly compare with the models without", + "type": "text" + } + ], + "index": 40 + } + ], + "index": 34.5 + } + ], + "page_idx": 6, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 302, + 741, + 309, + 750 + ], + "lines": [ + { + "bbox": [ + 302, + 741, + 309, + 752 + ], + "spans": [ + { + "bbox": [ + 302, + 741, + 309, + 752 + ], + "score": 1.0, + "content": "7", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "image", + "bbox": [ + 109, + 82, + 484, + 218 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 109, + 82, + 484, + 218 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 109, + 82, + 484, + 218 + ], + "spans": [ + { + "bbox": [ + 109, + 82, + 484, + 218 + ], + "score": 0.916, + "type": "image", + "image_path": "810578c2e36580b46aa98e4ec01030b20f6e00994c33dafcaa4cde435beb4d46.jpg" + } + ] + } + ], + "index": 1, + "virtual_lines": [ + { + "bbox": [ + 109, + 82, + 484, + 127.33333333333334 + ], + "spans": [], + "index": 0 + }, + { + "bbox": [ + 109, + 127.33333333333334, + 484, + 172.66666666666669 + ], + "spans": [], + "index": 1 + }, + { + "bbox": [ + 109, + 172.66666666666669, + 484, + 218.00000000000003 + ], + "spans": [], + "index": 2 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 106, + 223, + 506, + 288 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 104, + 222, + 506, + 236 + ], + "spans": [ + { + "bbox": [ + 104, + 222, + 506, + 236 + ], + "score": 1.0, + "content": "Figure 4: Left: LV-ViT ImageNet Top-1 Accuracy w.r.t. the token participation rate while applying", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 105, + 233, + 506, + 247 + ], + "spans": [ + { + "bbox": [ + 105, + 233, + 506, + 247 + ], + "score": 1.0, + "content": "token labeling. Token participation rate indicates the percentage of patch tokens involved in computing", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 244, + 506, + 257 + ], + "spans": [ + { + "bbox": [ + 105, + 244, + 506, + 257 + ], + "score": 1.0, + "content": "the token labeling loss. This experiment reflects that all tokens matter for vision transformers. Right:", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 106, + 256, + 505, + 267 + ], + "spans": [ + { + "bbox": [ + 106, + 256, + 505, + 267 + ], + "score": 1.0, + "content": "LV-ViT-S ImageNet Top-1 Accuracy w.r.t. different annotator models. The point size indicates the", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 104, + 266, + 506, + 279 + ], + "spans": [ + { + "bbox": [ + 104, + 266, + 506, + 279 + ], + "score": 1.0, + "content": "parameter number of the annotator model. Clearly, our token labeling objective is robust to different", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 105, + 278, + 180, + 288 + ], + "spans": [ + { + "bbox": [ + 105, + 278, + 180, + 288 + ], + "score": 1.0, + "content": "annotator models.", + "type": "text" + } + ], + "index": 8 + } + ], + "index": 5.5 + } + ], + "index": 3.25 + }, + { + "type": "table", + "bbox": [ + 128, + 342, + 483, + 411 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 108, + 308, + 505, + 341 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 307, + 505, + 321 + ], + "spans": [ + { + "bbox": [ + 105, + 307, + 505, + 321 + ], + "score": 1.0, + "content": "Table 4: Comparison of token labeling (TL), knowledge distillation (KD) based method and ReLabel", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 106, + 319, + 505, + 331 + ], + "spans": [ + { + "bbox": [ + 106, + 319, + 505, + 331 + ], + "score": 1.0, + "content": "method based on utilized tokens, DeiT-S/LV-ViT-S Top-1 accuracy on ImageNet validation set and", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 106, + 330, + 278, + 343 + ], + "spans": [ + { + "bbox": [ + 106, + 330, + 278, + 343 + ], + "score": 1.0, + "content": "training time on a single V100 GPU node.", + "type": "text" + } + ], + "index": 11 + } + ], + "index": 10 + }, + { + "type": "table_body", + "bbox": [ + 128, + 342, + 483, + 411 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 128, + 342, + 483, + 411 + ], + "spans": [ + { + "bbox": [ + 128, + 342, + 483, + 411 + ], + "score": 0.979, + "html": "
MethodOnline KDOnline TLTLReLabelVanilla
Tokens Utilized2AllAll11
DeiT-S Acc. (%)81.281.881.080.479.9
LV-ViT-S Acc. (%)83.083.583.382.882.4
Training Time (8× V100)63 hrs63 hrs45 hrs45 hrs41 hrs
", + "type": "table", + "image_path": "ae9fdf37e5d602133531450f1b69751afa30b0104c5bff5b6433f7a76be69066.jpg" + } + ] + } + ], + "index": 13, + "virtual_lines": [ + { + "bbox": [ + 128, + 342, + 483, + 365.0 + ], + "spans": [], + "index": 12 + }, + { + "bbox": [ + 128, + 365.0, + 483, + 388.0 + ], + "spans": [], + "index": 13 + }, + { + "bbox": [ + 128, + 388.0, + 483, + 411.0 + ], + "spans": [], + "index": 14 + } + ] + } + ], + "index": 11.5 + }, + { + "type": "text", + "bbox": [ + 106, + 432, + 505, + 585 + ], + "lines": [ + { + "bbox": [ + 105, + 431, + 505, + 445 + ], + "spans": [ + { + "bbox": [ + 105, + 431, + 505, + 445 + ], + "score": 1.0, + "content": "Online Token Labeling: Unlike the online knowledge distillation method which generates labels by", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 105, + 443, + 505, + 456 + ], + "spans": [ + { + "bbox": [ + 105, + 443, + 505, + 456 + ], + "score": 1.0, + "content": "a teacher model online, our token labeling approach utilizes the dense label map generated in advance", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 106, + 455, + 505, + 467 + ], + "spans": [ + { + "bbox": [ + 106, + 455, + 505, + 467 + ], + "score": 1.0, + "content": "and directly applies the corresponding augmentation methods, such as random crop, on the label", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 464, + 506, + 478 + ], + "spans": [ + { + "bbox": [ + 105, + 464, + 506, + 478 + ], + "score": 1.0, + "content": "map to obtain token-level labels. To directly compare with the online knowledge distillation based", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 105, + 475, + 506, + 489 + ], + "spans": [ + { + "bbox": [ + 105, + 475, + 506, + 489 + ], + "score": 1.0, + "content": "method and validate the effectiveness of token-level supervision, we further conduct experiments on", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 105, + 486, + 505, + 501 + ], + "spans": [ + { + "bbox": [ + 105, + 486, + 505, + 501 + ], + "score": 1.0, + "content": "the online version of our token labeling method, which generates token-level labels online during", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 106, + 498, + 506, + 510 + ], + "spans": [ + { + "bbox": [ + 106, + 498, + 506, + 510 + ], + "score": 1.0, + "content": "training. Following DeiT [36], we use RegNetY-16GF [30] as the online teacher model. Results in", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 105, + 508, + 505, + 522 + ], + "spans": [ + { + "bbox": [ + 105, + 508, + 505, + 522 + ], + "score": 1.0, + "content": "terms of DeiT-S/LV-ViT-S Top-1 accuracy and training time for our token labeling, online knowledge", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 519, + 506, + 532 + ], + "spans": [ + { + "bbox": [ + 105, + 519, + 506, + 532 + ], + "score": 1.0, + "content": "distillation, and ReLabel [49] are listed in Table 4, with number of utilized tokens also included for", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 531, + 506, + 543 + ], + "spans": [ + { + "bbox": [ + 105, + 531, + 506, + 543 + ], + "score": 1.0, + "content": "clear comparison. As can be seen, for both online and offline cases, using token-level supervision can", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 105, + 541, + 506, + 555 + ], + "spans": [ + { + "bbox": [ + 105, + 541, + 506, + 555 + ], + "score": 1.0, + "content": "improve the overall performance with only negligible additional training cost. Meanwhile, compared", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 106, + 553, + 505, + 565 + ], + "spans": [ + { + "bbox": [ + 106, + 553, + 505, + 565 + ], + "score": 1.0, + "content": "to the vanilla training baseline, our proposed offline token labeling brings almost no additional", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 106, + 564, + 505, + 576 + ], + "spans": [ + { + "bbox": [ + 106, + 564, + 366, + 576 + ], + "score": 1.0, + "content": "training cost, and boosts the overall performance of LV-ViT-S by", + "type": "text" + }, + { + "bbox": [ + 366, + 564, + 388, + 574 + ], + "score": 0.86, + "content": "0 . 9 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 388, + 564, + 505, + 576 + ], + "score": 1.0, + "content": ", which well demonstrates its", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 106, + 574, + 221, + 587 + ], + "spans": [ + { + "bbox": [ + 106, + 574, + 221, + 587 + ], + "score": 1.0, + "content": "efficiency and effectiveness.", + "type": "text" + } + ], + "index": 28 + } + ], + "index": 21.5, + "bbox_fs": [ + 105, + 431, + 506, + 587 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 591, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 106, + 591, + 505, + 603 + ], + "spans": [ + { + "bbox": [ + 106, + 591, + 505, + 603 + ], + "score": 1.0, + "content": "Robustness to Different Annotators: To evaluate the robustness of our token labeling method, we", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 106, + 602, + 505, + 613 + ], + "spans": [ + { + "bbox": [ + 106, + 602, + 505, + 613 + ], + "score": 1.0, + "content": "use different pretrained CNNs, including EfficientNet-B3,B4,B5,B6,B7,B8 [34], NFNet-F6 [3] and", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 612, + 505, + 625 + ], + "spans": [ + { + "bbox": [ + 105, + 612, + 505, + 625 + ], + "score": 1.0, + "content": "ResNest269E [51], as annotator models to provide dense supervision. Results are shown in the right", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 623, + 506, + 636 + ], + "spans": [ + { + "bbox": [ + 105, + 623, + 506, + 636 + ], + "score": 1.0, + "content": "part of Figure 4. We can see that, even if we use an annotator with relatively lower performance, such", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 106, + 635, + 505, + 647 + ], + "spans": [ + { + "bbox": [ + 106, + 635, + 284, + 647 + ], + "score": 1.0, + "content": "as EfficientNet-B3 whose Top-1 accuracy is", + "type": "text" + }, + { + "bbox": [ + 285, + 635, + 311, + 645 + ], + "score": 0.89, + "content": "8 1 . 6 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 312, + 635, + 505, + 647 + ], + "score": 1.0, + "content": ", it can still provide multi-label location-specific", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 646, + 505, + 657 + ], + "spans": [ + { + "bbox": [ + 105, + 646, + 505, + 657 + ], + "score": 1.0, + "content": "supervision and help improve the performance of our LV-ViT-S model. Meanwhile, annotator models", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 105, + 655, + 506, + 669 + ], + "spans": [ + { + "bbox": [ + 105, + 655, + 506, + 669 + ], + "score": 1.0, + "content": "with better performance can provide more accurate supervision, bringing even better performance, as", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 105, + 666, + 506, + 680 + ], + "spans": [ + { + "bbox": [ + 105, + 666, + 506, + 680 + ], + "score": 1.0, + "content": "stronger annotator models can generate better token-level labels. The largest annotator NFNet-F6 [3],", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 106, + 678, + 506, + 690 + ], + "spans": [ + { + "bbox": [ + 106, + 678, + 244, + 690 + ], + "score": 1.0, + "content": "which has the best performance of", + "type": "text" + }, + { + "bbox": [ + 244, + 678, + 271, + 689 + ], + "score": 0.89, + "content": "8 6 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 271, + 678, + 506, + 690 + ], + "score": 1.0, + "content": ", allows us to achieve the best result for LV-ViT-S, which is", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 106, + 689, + 506, + 701 + ], + "spans": [ + { + "bbox": [ + 106, + 689, + 134, + 700 + ], + "score": 0.87, + "content": "8 3 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 134, + 689, + 506, + 701 + ], + "score": 1.0, + "content": ". In addition, we also attempt to use a better model, EfficientNet-L2 pretrained on JFT-300M", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 105, + 699, + 505, + 712 + ], + "spans": [ + { + "bbox": [ + 105, + 699, + 230, + 712 + ], + "score": 1.0, + "content": "as described in [49] which has", + "type": "text" + }, + { + "bbox": [ + 230, + 700, + 257, + 711 + ], + "score": 0.88, + "content": "8 8 . 2 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 257, + 699, + 505, + 712 + ], + "score": 1.0, + "content": "Top-1 ImageNet accuracy, as our annotator. The performance", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 106, + 711, + 505, + 723 + ], + "spans": [ + { + "bbox": [ + 106, + 711, + 264, + 723 + ], + "score": 1.0, + "content": "of LV-ViT-S can be further improved to", + "type": "text" + }, + { + "bbox": [ + 265, + 711, + 292, + 721 + ], + "score": 0.88, + "content": "8 \\mathrm { { 3 . 5 \\% } }", + "type": "inline_equation" + }, + { + "bbox": [ + 292, + 711, + 505, + 723 + ], + "score": 1.0, + "content": ". However, to fairly compare with the models without", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 105, + 72, + 506, + 86 + ], + "spans": [ + { + "bbox": [ + 105, + 72, + 506, + 86 + ], + "score": 1.0, + "content": "extra training data, we only report results based on dense supervision produced by NFNet-F6 [3] that", + "type": "text", + "cross_page": true + } + ], + "index": 0 + }, + { + "bbox": [ + 105, + 84, + 243, + 96 + ], + "spans": [ + { + "bbox": [ + 105, + 84, + 243, + 96 + ], + "score": 1.0, + "content": "uses only ImageNet training data.", + "type": "text", + "cross_page": true + } + ], + "index": 1 + } + ], + "index": 34.5, + "bbox_fs": [ + 105, + 591, + 506, + 723 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "text", + "bbox": [ + 107, + 73, + 505, + 95 + ], + "lines": [ + { + "bbox": [ + 105, + 72, + 506, + 86 + ], + "spans": [ + { + "bbox": [ + 105, + 72, + 506, + 86 + ], + "score": 1.0, + "content": "extra training data, we only report results based on dense supervision produced by NFNet-F6 [3] that", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 105, + 84, + 243, + 96 + ], + "spans": [ + { + "bbox": [ + 105, + 84, + 243, + 96 + ], + "score": 1.0, + "content": "uses only ImageNet training data.", + "type": "text" + } + ], + "index": 1 + } + ], + "index": 0.5 + }, + { + "type": "image", + "bbox": [ + 110, + 114, + 505, + 206 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 110, + 114, + 505, + 206 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 110, + 114, + 505, + 206 + ], + "spans": [ + { + "bbox": [ + 110, + 114, + 505, + 206 + ], + "score": 0.966, + "type": "image", + "image_path": "a3d559defddf629a8a69e1ea7f71228daa68e20f028756c9b3671f019cf5ad40.jpg" + } + ] + } + ], + "index": 3, + "virtual_lines": [ + { + "bbox": [ + 110, + 114, + 505, + 144.66666666666666 + ], + "spans": [], + "index": 2 + }, + { + "bbox": [ + 110, + 144.66666666666666, + 505, + 175.33333333333331 + ], + "spans": [], + "index": 3 + }, + { + "bbox": [ + 110, + 175.33333333333331, + 505, + 205.99999999999997 + ], + "spans": [], + "index": 4 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 107, + 214, + 506, + 247 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 213, + 507, + 226 + ], + "spans": [ + { + "bbox": [ + 105, + 213, + 507, + 226 + ], + "score": 1.0, + "content": "Figure 5: Performance of the proposed token labeling objective on three different vision transformers:", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 106, + 225, + 505, + 237 + ], + "spans": [ + { + "bbox": [ + 106, + 225, + 505, + 237 + ], + "score": 1.0, + "content": "DeiT [36] (Left), T2T-ViT [46] (Middle), and LV-ViT (Right). Our method has a consistent", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 105, + 236, + 283, + 247 + ], + "spans": [ + { + "bbox": [ + 105, + 236, + 283, + 247 + ], + "score": 1.0, + "content": "improvement on all 7 different ViT models.", + "type": "text" + } + ], + "index": 7 + } + ], + "index": 6 + } + ], + "index": 4.5 + }, + { + "type": "text", + "bbox": [ + 106, + 267, + 505, + 356 + ], + "lines": [ + { + "bbox": [ + 105, + 267, + 505, + 280 + ], + "spans": [ + { + "bbox": [ + 105, + 267, + 505, + 280 + ], + "score": 1.0, + "content": "Robustness to Different ViT Variants: To further evaluate the robustness of our token labeling, we", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 105, + 278, + 506, + 290 + ], + "spans": [ + { + "bbox": [ + 105, + 278, + 506, + 290 + ], + "score": 1.0, + "content": "train different transformer-based networks, including DeiT [36], T2T-ViT [3] and our model LV-ViT,", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 106, + 290, + 505, + 302 + ], + "spans": [ + { + "bbox": [ + 106, + 290, + 505, + 302 + ], + "score": 1.0, + "content": "with the proposed training objective. Results are shown in Figure 5. It can be found that, all the", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 300, + 505, + 313 + ], + "spans": [ + { + "bbox": [ + 105, + 300, + 505, + 313 + ], + "score": 1.0, + "content": "models trained with token labeling consistently outperform their vanilla counterparts, demonstrating", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 311, + 506, + 324 + ], + "spans": [ + { + "bbox": [ + 105, + 311, + 506, + 324 + ], + "score": 1.0, + "content": "the robustness of token labeling with respect to different variants of patch-based vision transformers.", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 322, + 505, + 335 + ], + "spans": [ + { + "bbox": [ + 105, + 322, + 505, + 335 + ], + "score": 1.0, + "content": "Meanwhile, for different scales of the models, the improvement is also consistent. Interestingly, we", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 332, + 506, + 346 + ], + "spans": [ + { + "bbox": [ + 105, + 332, + 506, + 346 + ], + "score": 1.0, + "content": "observe larger improvements for larger models. These indicate that our proposed token labeling", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 105, + 344, + 457, + 356 + ], + "spans": [ + { + "bbox": [ + 105, + 344, + 457, + 356 + ], + "score": 1.0, + "content": "method is widely applicable to a large range of patch-based vision transformer variants.", + "type": "text" + } + ], + "index": 15 + } + ], + "index": 11.5 + }, + { + "type": "text", + "bbox": [ + 106, + 360, + 505, + 426 + ], + "lines": [ + { + "bbox": [ + 105, + 360, + 506, + 373 + ], + "spans": [ + { + "bbox": [ + 105, + 360, + 506, + 373 + ], + "score": 1.0, + "content": "Beyond Vision Transformers: We further explore the performance of token labeling on other", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 370, + 505, + 384 + ], + "spans": [ + { + "bbox": [ + 105, + 370, + 505, + 384 + ], + "score": 1.0, + "content": "CNN-based and MLP-based models. Results are shown in Table 5. Besides our re-implementation", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 106, + 383, + 505, + 394 + ], + "spans": [ + { + "bbox": [ + 106, + 383, + 505, + 394 + ], + "score": 1.0, + "content": "with more data augmentation and regularization techniques, we also provide the results from the", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 104, + 392, + 505, + 407 + ], + "spans": [ + { + "bbox": [ + 104, + 392, + 505, + 407 + ], + "score": 1.0, + "content": "original papers. It can be found that for both MLP-based and CNN-based models, our token labeling", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 106, + 404, + 505, + 417 + ], + "spans": [ + { + "bbox": [ + 106, + 404, + 505, + 417 + ], + "score": 1.0, + "content": "objective can also improve the performance over strong baselines by providing location-specific", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 106, + 415, + 182, + 427 + ], + "spans": [ + { + "bbox": [ + 106, + 415, + 182, + 427 + ], + "score": 1.0, + "content": "dense supervision.", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 18.5 + }, + { + "type": "table", + "bbox": [ + 111, + 486, + 499, + 545 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 452, + 505, + 486 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 104, + 451, + 505, + 465 + ], + "spans": [ + { + "bbox": [ + 104, + 451, + 505, + 465 + ], + "score": 1.0, + "content": "Table 5: Performance of the proposed token labeling objective on representative CNN-based", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 106, + 464, + 505, + 476 + ], + "spans": [ + { + "bbox": [ + 106, + 464, + 505, + 476 + ], + "score": 1.0, + "content": "(ResNeSt) and MLP-based (Mixer-MLP) models. Our method has a consistent improvement on all", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 473, + 383, + 488 + ], + "spans": [ + { + "bbox": [ + 105, + 473, + 383, + 488 + ], + "score": 1.0, + "content": "different models. Here † indicates results reported in original papers.", + "type": "text" + } + ], + "index": 24 + } + ], + "index": 23 + }, + { + "type": "table_body", + "bbox": [ + 111, + 486, + 499, + 545 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 111, + 486, + 499, + 545 + ], + "spans": [ + { + "bbox": [ + 111, + 486, + 499, + 545 + ], + "score": 0.978, + "html": "
ModelMixer-S/16 [35]Mixer-B/16 [35]Mixer-L/16 [35]ResNeSt-50 [51]
Token LabelingXX√×X√××√××√
Parameters18M18M18M59M59M 59M207M207M207M27M27M 27M
Top-1 Acc. (%)73.8t75.676.176.478.379.571.6t77.780.181.1t80.981.5
", + "type": "table", + "image_path": "163ebee6ecc8c4b165c6f3e5a941947668dd1d76a42075b0fcd4b5d7c14a71c5.jpg" + } + ] + } + ], + "index": 26, + "virtual_lines": [ + { + "bbox": [ + 111, + 486, + 499, + 505.6666666666667 + ], + "spans": [], + "index": 25 + }, + { + "bbox": [ + 111, + 505.6666666666667, + 499, + 525.3333333333334 + ], + "spans": [], + "index": 26 + }, + { + "bbox": [ + 111, + 525.3333333333334, + 499, + 545.0 + ], + "spans": [], + "index": 27 + } + ] + } + ], + "index": 24.5 + }, + { + "type": "title", + "bbox": [ + 107, + 588, + 261, + 600 + ], + "lines": [ + { + "bbox": [ + 105, + 588, + 262, + 602 + ], + "spans": [ + { + "bbox": [ + 105, + 588, + 262, + 602 + ], + "score": 1.0, + "content": "4.3 Comparison to Other Methods", + "type": "text" + } + ], + "index": 28 + } + ], + "index": 28 + }, + { + "type": "text", + "bbox": [ + 106, + 612, + 506, + 722 + ], + "lines": [ + { + "bbox": [ + 106, + 612, + 506, + 623 + ], + "spans": [ + { + "bbox": [ + 106, + 612, + 506, + 623 + ], + "score": 1.0, + "content": "We compare our proposed model LV-ViT with other state-of-the-art methods in Table 6. For small-", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 623, + 506, + 636 + ], + "spans": [ + { + "bbox": [ + 105, + 623, + 287, + 636 + ], + "score": 1.0, + "content": "sized models, when the test resolution is set to", + "type": "text" + }, + { + "bbox": [ + 287, + 624, + 327, + 634 + ], + "score": 0.9, + "content": "2 2 4 \\times 2 2 4", + "type": "inline_equation" + }, + { + "bbox": [ + 328, + 623, + 387, + 636 + ], + "score": 1.0, + "content": ", we achieve an", + "type": "text" + }, + { + "bbox": [ + 388, + 623, + 415, + 634 + ], + "score": 0.88, + "content": "8 3 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 415, + 623, + 506, + 636 + ], + "score": 1.0, + "content": "accuracy on ImageNet", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 633, + 506, + 647 + ], + "spans": [ + { + "bbox": [ + 105, + 633, + 252, + 647 + ], + "score": 1.0, + "content": "with only 26M parameters, which is", + "type": "text" + }, + { + "bbox": [ + 252, + 634, + 274, + 645 + ], + "score": 0.89, + "content": "3 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 275, + 633, + 506, + 647 + ], + "score": 1.0, + "content": "higher than the strong baseline DeiT-S [36]. For medium-", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 106, + 645, + 506, + 657 + ], + "spans": [ + { + "bbox": [ + 106, + 645, + 300, + 657 + ], + "score": 1.0, + "content": "sized models, when the test resolution is set to", + "type": "text" + }, + { + "bbox": [ + 300, + 645, + 345, + 656 + ], + "score": 0.9, + "content": "3 8 4 \\times 3 8 4", + "type": "inline_equation" + }, + { + "bbox": [ + 345, + 645, + 475, + 657 + ], + "score": 1.0, + "content": "we achieve the performance of", + "type": "text" + }, + { + "bbox": [ + 475, + 645, + 502, + 656 + ], + "score": 0.9, + "content": "8 5 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 503, + 645, + 506, + 657 + ], + "score": 1.0, + "content": ",", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 106, + 657, + 505, + 668 + ], + "spans": [ + { + "bbox": [ + 106, + 657, + 505, + 668 + ], + "score": 1.0, + "content": "the same as CaiT-S36 [37], but with much less computational cost and parameters. Note that both", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 667, + 506, + 680 + ], + "spans": [ + { + "bbox": [ + 105, + 667, + 506, + 680 + ], + "score": 1.0, + "content": "DeiT and CaiT use knowledge distillation to improve their models, which introduce much more", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 106, + 679, + 505, + 690 + ], + "spans": [ + { + "bbox": [ + 106, + 679, + 505, + 690 + ], + "score": 1.0, + "content": "computations in training. However, we do not require any extra computations in training and only", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 105, + 689, + 505, + 700 + ], + "spans": [ + { + "bbox": [ + 105, + 689, + 505, + 700 + ], + "score": 1.0, + "content": "have to compute and store the dense score maps in advance. For large-sized models, our LV-ViT-L", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 105, + 700, + 506, + 713 + ], + "spans": [ + { + "bbox": [ + 105, + 700, + 209, + 713 + ], + "score": 1.0, + "content": "with a test resolution of", + "type": "text" + }, + { + "bbox": [ + 210, + 700, + 255, + 711 + ], + "score": 0.89, + "content": "4 4 8 \\times 4 4 8", + "type": "inline_equation" + }, + { + "bbox": [ + 255, + 700, + 308, + 713 + ], + "score": 1.0, + "content": "achieves an", + "type": "text" + }, + { + "bbox": [ + 308, + 700, + 336, + 711 + ], + "score": 0.87, + "content": "8 6 . 2 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 336, + 700, + 506, + 713 + ], + "score": 1.0, + "content": "top-1 accuracy, which is comparable to", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 106, + 710, + 339, + 723 + ], + "spans": [ + { + "bbox": [ + 106, + 710, + 339, + 723 + ], + "score": 1.0, + "content": "CaiT-M36 [37] but with far fewer FLOPs and parameters.", + "type": "text" + } + ], + "index": 38 + } + ], + "index": 33.5 + } + ], + "page_idx": 7, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 302, + 742, + 308, + 750 + ], + "lines": [ + { + "bbox": [ + 301, + 740, + 310, + 752 + ], + "spans": [ + { + "bbox": [ + 301, + 740, + 310, + 752 + ], + "score": 1.0, + "content": "", + "type": "text", + "height": 12, + "width": 9 + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "text", + "bbox": [ + 107, + 73, + 505, + 95 + ], + "lines": [], + "index": 0.5, + "bbox_fs": [ + 105, + 72, + 506, + 96 + ], + "lines_deleted": true + }, + { + "type": "image", + "bbox": [ + 110, + 114, + 505, + 206 + ], + "blocks": [ + { + "type": "image_body", + "bbox": [ + 110, + 114, + 505, + 206 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 110, + 114, + 505, + 206 + ], + "spans": [ + { + "bbox": [ + 110, + 114, + 505, + 206 + ], + "score": 0.966, + "type": "image", + "image_path": "a3d559defddf629a8a69e1ea7f71228daa68e20f028756c9b3671f019cf5ad40.jpg" + } + ] + } + ], + "index": 3, + "virtual_lines": [ + { + "bbox": [ + 110, + 114, + 505, + 144.66666666666666 + ], + "spans": [], + "index": 2 + }, + { + "bbox": [ + 110, + 144.66666666666666, + 505, + 175.33333333333331 + ], + "spans": [], + "index": 3 + }, + { + "bbox": [ + 110, + 175.33333333333331, + 505, + 205.99999999999997 + ], + "spans": [], + "index": 4 + } + ] + }, + { + "type": "image_caption", + "bbox": [ + 107, + 214, + 506, + 247 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 213, + 507, + 226 + ], + "spans": [ + { + "bbox": [ + 105, + 213, + 507, + 226 + ], + "score": 1.0, + "content": "Figure 5: Performance of the proposed token labeling objective on three different vision transformers:", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 106, + 225, + 505, + 237 + ], + "spans": [ + { + "bbox": [ + 106, + 225, + 505, + 237 + ], + "score": 1.0, + "content": "DeiT [36] (Left), T2T-ViT [46] (Middle), and LV-ViT (Right). Our method has a consistent", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 105, + 236, + 283, + 247 + ], + "spans": [ + { + "bbox": [ + 105, + 236, + 283, + 247 + ], + "score": 1.0, + "content": "improvement on all 7 different ViT models.", + "type": "text" + } + ], + "index": 7 + } + ], + "index": 6 + } + ], + "index": 4.5 + }, + { + "type": "text", + "bbox": [ + 106, + 267, + 505, + 356 + ], + "lines": [ + { + "bbox": [ + 105, + 267, + 505, + 280 + ], + "spans": [ + { + "bbox": [ + 105, + 267, + 505, + 280 + ], + "score": 1.0, + "content": "Robustness to Different ViT Variants: To further evaluate the robustness of our token labeling, we", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 105, + 278, + 506, + 290 + ], + "spans": [ + { + "bbox": [ + 105, + 278, + 506, + 290 + ], + "score": 1.0, + "content": "train different transformer-based networks, including DeiT [36], T2T-ViT [3] and our model LV-ViT,", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 106, + 290, + 505, + 302 + ], + "spans": [ + { + "bbox": [ + 106, + 290, + 505, + 302 + ], + "score": 1.0, + "content": "with the proposed training objective. Results are shown in Figure 5. It can be found that, all the", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 300, + 505, + 313 + ], + "spans": [ + { + "bbox": [ + 105, + 300, + 505, + 313 + ], + "score": 1.0, + "content": "models trained with token labeling consistently outperform their vanilla counterparts, demonstrating", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 311, + 506, + 324 + ], + "spans": [ + { + "bbox": [ + 105, + 311, + 506, + 324 + ], + "score": 1.0, + "content": "the robustness of token labeling with respect to different variants of patch-based vision transformers.", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 322, + 505, + 335 + ], + "spans": [ + { + "bbox": [ + 105, + 322, + 505, + 335 + ], + "score": 1.0, + "content": "Meanwhile, for different scales of the models, the improvement is also consistent. Interestingly, we", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 332, + 506, + 346 + ], + "spans": [ + { + "bbox": [ + 105, + 332, + 506, + 346 + ], + "score": 1.0, + "content": "observe larger improvements for larger models. These indicate that our proposed token labeling", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 105, + 344, + 457, + 356 + ], + "spans": [ + { + "bbox": [ + 105, + 344, + 457, + 356 + ], + "score": 1.0, + "content": "method is widely applicable to a large range of patch-based vision transformer variants.", + "type": "text" + } + ], + "index": 15 + } + ], + "index": 11.5, + "bbox_fs": [ + 105, + 267, + 506, + 356 + ] + }, + { + "type": "text", + "bbox": [ + 106, + 360, + 505, + 426 + ], + "lines": [ + { + "bbox": [ + 105, + 360, + 506, + 373 + ], + "spans": [ + { + "bbox": [ + 105, + 360, + 506, + 373 + ], + "score": 1.0, + "content": "Beyond Vision Transformers: We further explore the performance of token labeling on other", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 370, + 505, + 384 + ], + "spans": [ + { + "bbox": [ + 105, + 370, + 505, + 384 + ], + "score": 1.0, + "content": "CNN-based and MLP-based models. Results are shown in Table 5. Besides our re-implementation", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 106, + 383, + 505, + 394 + ], + "spans": [ + { + "bbox": [ + 106, + 383, + 505, + 394 + ], + "score": 1.0, + "content": "with more data augmentation and regularization techniques, we also provide the results from the", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 104, + 392, + 505, + 407 + ], + "spans": [ + { + "bbox": [ + 104, + 392, + 505, + 407 + ], + "score": 1.0, + "content": "original papers. It can be found that for both MLP-based and CNN-based models, our token labeling", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 106, + 404, + 505, + 417 + ], + "spans": [ + { + "bbox": [ + 106, + 404, + 505, + 417 + ], + "score": 1.0, + "content": "objective can also improve the performance over strong baselines by providing location-specific", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 106, + 415, + 182, + 427 + ], + "spans": [ + { + "bbox": [ + 106, + 415, + 182, + 427 + ], + "score": 1.0, + "content": "dense supervision.", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 18.5, + "bbox_fs": [ + 104, + 360, + 506, + 427 + ] + }, + { + "type": "table", + "bbox": [ + 111, + 486, + 499, + 545 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 452, + 505, + 486 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 104, + 451, + 505, + 465 + ], + "spans": [ + { + "bbox": [ + 104, + 451, + 505, + 465 + ], + "score": 1.0, + "content": "Table 5: Performance of the proposed token labeling objective on representative CNN-based", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 106, + 464, + 505, + 476 + ], + "spans": [ + { + "bbox": [ + 106, + 464, + 505, + 476 + ], + "score": 1.0, + "content": "(ResNeSt) and MLP-based (Mixer-MLP) models. Our method has a consistent improvement on all", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 473, + 383, + 488 + ], + "spans": [ + { + "bbox": [ + 105, + 473, + 383, + 488 + ], + "score": 1.0, + "content": "different models. Here † indicates results reported in original papers.", + "type": "text" + } + ], + "index": 24 + } + ], + "index": 23 + }, + { + "type": "table_body", + "bbox": [ + 111, + 486, + 499, + 545 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 111, + 486, + 499, + 545 + ], + "spans": [ + { + "bbox": [ + 111, + 486, + 499, + 545 + ], + "score": 0.978, + "html": "
ModelMixer-S/16 [35]Mixer-B/16 [35]Mixer-L/16 [35]ResNeSt-50 [51]
Token LabelingXX√×X√××√××√
Parameters18M18M18M59M59M 59M207M207M207M27M27M 27M
Top-1 Acc. (%)73.8t75.676.176.478.379.571.6t77.780.181.1t80.981.5
", + "type": "table", + "image_path": "163ebee6ecc8c4b165c6f3e5a941947668dd1d76a42075b0fcd4b5d7c14a71c5.jpg" + } + ] + } + ], + "index": 26, + "virtual_lines": [ + { + "bbox": [ + 111, + 486, + 499, + 505.6666666666667 + ], + "spans": [], + "index": 25 + }, + { + "bbox": [ + 111, + 505.6666666666667, + 499, + 525.3333333333334 + ], + "spans": [], + "index": 26 + }, + { + "bbox": [ + 111, + 525.3333333333334, + 499, + 545.0 + ], + "spans": [], + "index": 27 + } + ] + } + ], + "index": 24.5 + }, + { + "type": "title", + "bbox": [ + 107, + 588, + 261, + 600 + ], + "lines": [ + { + "bbox": [ + 105, + 588, + 262, + 602 + ], + "spans": [ + { + "bbox": [ + 105, + 588, + 262, + 602 + ], + "score": 1.0, + "content": "4.3 Comparison to Other Methods", + "type": "text" + } + ], + "index": 28 + } + ], + "index": 28 + }, + { + "type": "text", + "bbox": [ + 106, + 612, + 506, + 722 + ], + "lines": [ + { + "bbox": [ + 106, + 612, + 506, + 623 + ], + "spans": [ + { + "bbox": [ + 106, + 612, + 506, + 623 + ], + "score": 1.0, + "content": "We compare our proposed model LV-ViT with other state-of-the-art methods in Table 6. For small-", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 623, + 506, + 636 + ], + "spans": [ + { + "bbox": [ + 105, + 623, + 287, + 636 + ], + "score": 1.0, + "content": "sized models, when the test resolution is set to", + "type": "text" + }, + { + "bbox": [ + 287, + 624, + 327, + 634 + ], + "score": 0.9, + "content": "2 2 4 \\times 2 2 4", + "type": "inline_equation" + }, + { + "bbox": [ + 328, + 623, + 387, + 636 + ], + "score": 1.0, + "content": ", we achieve an", + "type": "text" + }, + { + "bbox": [ + 388, + 623, + 415, + 634 + ], + "score": 0.88, + "content": "8 3 . 3 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 415, + 623, + 506, + 636 + ], + "score": 1.0, + "content": "accuracy on ImageNet", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 633, + 506, + 647 + ], + "spans": [ + { + "bbox": [ + 105, + 633, + 252, + 647 + ], + "score": 1.0, + "content": "with only 26M parameters, which is", + "type": "text" + }, + { + "bbox": [ + 252, + 634, + 274, + 645 + ], + "score": 0.89, + "content": "3 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 275, + 633, + 506, + 647 + ], + "score": 1.0, + "content": "higher than the strong baseline DeiT-S [36]. For medium-", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 106, + 645, + 506, + 657 + ], + "spans": [ + { + "bbox": [ + 106, + 645, + 300, + 657 + ], + "score": 1.0, + "content": "sized models, when the test resolution is set to", + "type": "text" + }, + { + "bbox": [ + 300, + 645, + 345, + 656 + ], + "score": 0.9, + "content": "3 8 4 \\times 3 8 4", + "type": "inline_equation" + }, + { + "bbox": [ + 345, + 645, + 475, + 657 + ], + "score": 1.0, + "content": "we achieve the performance of", + "type": "text" + }, + { + "bbox": [ + 475, + 645, + 502, + 656 + ], + "score": 0.9, + "content": "8 5 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 503, + 645, + 506, + 657 + ], + "score": 1.0, + "content": ",", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 106, + 657, + 505, + 668 + ], + "spans": [ + { + "bbox": [ + 106, + 657, + 505, + 668 + ], + "score": 1.0, + "content": "the same as CaiT-S36 [37], but with much less computational cost and parameters. Note that both", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 667, + 506, + 680 + ], + "spans": [ + { + "bbox": [ + 105, + 667, + 506, + 680 + ], + "score": 1.0, + "content": "DeiT and CaiT use knowledge distillation to improve their models, which introduce much more", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 106, + 679, + 505, + 690 + ], + "spans": [ + { + "bbox": [ + 106, + 679, + 505, + 690 + ], + "score": 1.0, + "content": "computations in training. However, we do not require any extra computations in training and only", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 105, + 689, + 505, + 700 + ], + "spans": [ + { + "bbox": [ + 105, + 689, + 505, + 700 + ], + "score": 1.0, + "content": "have to compute and store the dense score maps in advance. For large-sized models, our LV-ViT-L", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 105, + 700, + 506, + 713 + ], + "spans": [ + { + "bbox": [ + 105, + 700, + 209, + 713 + ], + "score": 1.0, + "content": "with a test resolution of", + "type": "text" + }, + { + "bbox": [ + 210, + 700, + 255, + 711 + ], + "score": 0.89, + "content": "4 4 8 \\times 4 4 8", + "type": "inline_equation" + }, + { + "bbox": [ + 255, + 700, + 308, + 713 + ], + "score": 1.0, + "content": "achieves an", + "type": "text" + }, + { + "bbox": [ + 308, + 700, + 336, + 711 + ], + "score": 0.87, + "content": "8 6 . 2 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 336, + 700, + 506, + 713 + ], + "score": 1.0, + "content": "top-1 accuracy, which is comparable to", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 106, + 710, + 339, + 723 + ], + "spans": [ + { + "bbox": [ + 106, + 710, + 339, + 723 + ], + "score": 1.0, + "content": "CaiT-M36 [37] but with far fewer FLOPs and parameters.", + "type": "text" + } + ], + "index": 38 + } + ], + "index": 33.5, + "bbox_fs": [ + 105, + 612, + 506, + 723 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "table", + "bbox": [ + 107, + 119, + 505, + 479 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 77, + 506, + 121 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 76, + 507, + 92 + ], + "spans": [ + { + "bbox": [ + 105, + 76, + 507, + 92 + ], + "score": 1.0, + "content": "Table 6: Top-1 accuracy comparison with other methods on ImageNet [13] and ImageNet Real [2].", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 105, + 88, + 506, + 101 + ], + "spans": [ + { + "bbox": [ + 105, + 88, + 506, + 101 + ], + "score": 1.0, + "content": "All models are trained without external data. With the same computation and parameter constraint,", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 105, + 98, + 506, + 113 + ], + "spans": [ + { + "bbox": [ + 105, + 98, + 506, + 113 + ], + "score": 1.0, + "content": "our model consistently outperforms other CNN-based and transformer-based counterparts. The", + "type": "text" + } + ], + "index": 2 + }, + { + "bbox": [ + 105, + 109, + 311, + 123 + ], + "spans": [ + { + "bbox": [ + 105, + 109, + 311, + 123 + ], + "score": 1.0, + "content": "results of CNNs and ViT are referenced from [37].", + "type": "text" + } + ], + "index": 3 + } + ], + "index": 1.5 + }, + { + "type": "table_body", + "bbox": [ + 107, + 119, + 505, + 479 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 107, + 119, + 505, + 479 + ], + "spans": [ + { + "bbox": [ + 107, + 119, + 505, + 479 + ], + "score": 0.983, + "html": "
NetworkParamsFLOPsTrain sizeTest sizeTop-1(%)Real Top-1 (%)
EfficientNet-B5 [34] SNNO30M9.9B45645683.688.3
EfficientNet-B7 [34]66M37.0B60060084.3
Fix-EfficientNet-B8 [34,38]87M89.5B67280085.790.0
NFNet-F3 [3]255M114.8B32041685.789.4
NFNet-F4 [3]316M215.3B38451285.989.4
NFNet-F5[3]377M289.8B41654486.089.2
ViT-B/16 [15]86M55.4B22438477.983.6
ViT-L/16[15]307M190.7B22438476.582.2
T2T-ViT-14 [46]22M5.2B22422481.5
T2T-ViT-14↑384 [46]22M17.1B22438483.31
Cross ViT [6]45M56.6B22448084.1一
Swin-B[25]88M47.0B22438484.2
TNT-B[16]66M14.1B22422482.8
iriirrrrrs DeepViT-S [59]27M6.2B22422482.3
DeepViT-L [59]55M12.5B22422483.1
DeiT-S[36]22M4.6B22422479.985.7
Distilled DeiT-S [36]22M4.6B22422481.286.8
DeiT-B [36]86M17.5B22422481.886.7
DeiT-B↑384 [36]86M55.4B22438483.187.7
Distilled DeiT-B [36]87M17.5B22422483.488.3
BoTNet-S1-128 [31]79.1M19.3B25625684.2
BoTNet-S1-128↑384 [31]79.1M45.8B25638484.7-
CaiT-S36↑384 [37]68M48.0B22438485.4- 89.8
CaiT-M36[37]271M53.7B22422485.189.3
CaiT-M36↑448 [37]271M247.8B22444886.390.2
LV-ViT-S26M6.6B22422483.3
江 LV-ViT-S↑38426M22.2B22438484.488.1 88.9
LV-ViT-M56M16.0B22422484.188.4
W LV-ViT-M↑38456M42.2B22438485.489.5
LV-ViT-L150M59.0B28828885.389.3
0 LV-ViT-L↑448150M157.2B28844885.989.7
LV-ViT-L↑448150M157.2B44844886.289.9
LV-ViT-L↑512151M214.8B44851286.490.1
", + "type": "table", + "image_path": "1ffe2c2d40127fa5107472db1c5e025901208024a3c26b0c26ca4a1ac6751875.jpg" + } + ] + } + ], + "index": 5, + "virtual_lines": [ + { + "bbox": [ + 107, + 119, + 505, + 239.0 + ], + "spans": [], + "index": 4 + }, + { + "bbox": [ + 107, + 239.0, + 505, + 359.0 + ], + "spans": [], + "index": 5 + }, + { + "bbox": [ + 107, + 359.0, + 505, + 479.0 + ], + "spans": [], + "index": 6 + } + ] + } + ], + "index": 3.25 + }, + { + "type": "title", + "bbox": [ + 107, + 503, + 284, + 515 + ], + "lines": [ + { + "bbox": [ + 105, + 502, + 286, + 517 + ], + "spans": [ + { + "bbox": [ + 105, + 502, + 286, + 517 + ], + "score": 1.0, + "content": "4.4 Semantic Segmentation on ADE20K", + "type": "text" + } + ], + "index": 7 + } + ], + "index": 7 + }, + { + "type": "text", + "bbox": [ + 107, + 525, + 505, + 569 + ], + "lines": [ + { + "bbox": [ + 105, + 525, + 505, + 537 + ], + "spans": [ + { + "bbox": [ + 105, + 525, + 505, + 537 + ], + "score": 1.0, + "content": "It has been shown in [19] that different training techniques for pretrained models have different", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 106, + 537, + 505, + 548 + ], + "spans": [ + { + "bbox": [ + 106, + 537, + 505, + 548 + ], + "score": 1.0, + "content": "impacts on downstream tasks with dense prediction, like semantic segmentation. To demonstrate", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 105, + 547, + 506, + 560 + ], + "spans": [ + { + "bbox": [ + 105, + 547, + 506, + 560 + ], + "score": 1.0, + "content": "the advantage of the proposed token labeling objective on tasks with dense prediction, we apply our", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 558, + 401, + 570 + ], + "spans": [ + { + "bbox": [ + 105, + 558, + 401, + 570 + ], + "score": 1.0, + "content": "pretrained LV-ViT with token labeling to the semantic segmentation task.", + "type": "text" + } + ], + "index": 11 + } + ], + "index": 9.5 + }, + { + "type": "text", + "bbox": [ + 107, + 574, + 505, + 684 + ], + "lines": [ + { + "bbox": [ + 105, + 574, + 506, + 587 + ], + "spans": [ + { + "bbox": [ + 105, + 574, + 506, + 587 + ], + "score": 1.0, + "content": "Similar to previous work [25], we run experiments on the widely-used ADE20K [58] dataset.", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 585, + 506, + 598 + ], + "spans": [ + { + "bbox": [ + 105, + 585, + 506, + 598 + ], + "score": 1.0, + "content": "ADE20K contains 25K images in total, including 20K images for training, 2K images for validation", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 595, + 506, + 610 + ], + "spans": [ + { + "bbox": [ + 105, + 595, + 506, + 610 + ], + "score": 1.0, + "content": "and 3K images for test, and covering 150 different foreground categories. We take both FCN [26]", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 105, + 606, + 506, + 620 + ], + "spans": [ + { + "bbox": [ + 105, + 606, + 506, + 620 + ], + "score": 1.0, + "content": "and UperNet [44] as our segmentation frameworks and use the mmseg toolbox to implement. During", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 105, + 618, + 506, + 631 + ], + "spans": [ + { + "bbox": [ + 105, + 618, + 506, + 631 + ], + "score": 1.0, + "content": "training, following [25], we use the AdamW optimizer with an initial learning rate of 6e-5 and", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 629, + 506, + 642 + ], + "spans": [ + { + "bbox": [ + 105, + 629, + 506, + 642 + ], + "score": 1.0, + "content": "a weight decay of 0.01. We also use a linear learning schedule with a minimum learning rate of", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 639, + 506, + 652 + ], + "spans": [ + { + "bbox": [ + 105, + 639, + 506, + 652 + ], + "score": 1.0, + "content": "5e-6. All models are trained on 8 GPUs and with a batch size of 16 (i.e., 2 images on each GPU).", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 105, + 650, + 506, + 663 + ], + "spans": [ + { + "bbox": [ + 105, + 650, + 224, + 663 + ], + "score": 1.0, + "content": "The input resolution is set to", + "type": "text" + }, + { + "bbox": [ + 225, + 651, + 268, + 661 + ], + "score": 0.9, + "content": "5 1 2 \\times 5 1 2", + "type": "inline_equation" + }, + { + "bbox": [ + 268, + 650, + 506, + 663 + ], + "score": 1.0, + "content": ". In inference, a multi-scale test with interpolation rates of", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 105, + 662, + 505, + 675 + ], + "spans": [ + { + "bbox": [ + 105, + 662, + 505, + 675 + ], + "score": 1.0, + "content": "[0.75, 1.0, 1.25, 1.5, 1.75] is used. As suggested by [58], we report results in terms of both mean", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 105, + 673, + 412, + 685 + ], + "spans": [ + { + "bbox": [ + 105, + 673, + 412, + 685 + ], + "score": 1.0, + "content": "intersection-over-union (mIoU) and the average pixel accuracy (Pixel Acc.).", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 16.5 + }, + { + "type": "text", + "bbox": [ + 108, + 689, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 105, + 688, + 505, + 702 + ], + "spans": [ + { + "bbox": [ + 105, + 688, + 505, + 702 + ], + "score": 1.0, + "content": "In Table 7, we test the performance of token labeling on both FCN and UperNet frameworks. The", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 699, + 505, + 713 + ], + "spans": [ + { + "bbox": [ + 105, + 699, + 505, + 713 + ], + "score": 1.0, + "content": "FCN framework has a light convolutional head and can directly reflect the performance of the", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 711, + 505, + 723 + ], + "spans": [ + { + "bbox": [ + 105, + 711, + 505, + 723 + ], + "score": 1.0, + "content": "pretrained models in terms of transferable capability. As can be seen, pretrained models with token", + "type": "text" + } + ], + "index": 24 + } + ], + "index": 23 + } + ], + "page_idx": 8, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 302, + 741, + 308, + 750 + ], + "lines": [ + { + "bbox": [ + 302, + 741, + 309, + 752 + ], + "spans": [ + { + "bbox": [ + 302, + 741, + 309, + 752 + ], + "score": 1.0, + "content": "9", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "table", + "bbox": [ + 107, + 119, + 505, + 479 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 77, + 506, + 121 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 76, + 507, + 92 + ], + "spans": [ + { + "bbox": [ + 105, + 76, + 507, + 92 + ], + "score": 1.0, + "content": "Table 6: Top-1 accuracy comparison with other methods on ImageNet [13] and ImageNet Real [2].", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 105, + 88, + 506, + 101 + ], + "spans": [ + { + "bbox": [ + 105, + 88, + 506, + 101 + ], + "score": 1.0, + "content": "All models are trained without external data. With the same computation and parameter constraint,", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 105, + 98, + 506, + 113 + ], + "spans": [ + { + "bbox": [ + 105, + 98, + 506, + 113 + ], + "score": 1.0, + "content": "our model consistently outperforms other CNN-based and transformer-based counterparts. The", + "type": "text" + } + ], + "index": 2 + }, + { + "bbox": [ + 105, + 109, + 311, + 123 + ], + "spans": [ + { + "bbox": [ + 105, + 109, + 311, + 123 + ], + "score": 1.0, + "content": "results of CNNs and ViT are referenced from [37].", + "type": "text" + } + ], + "index": 3 + } + ], + "index": 1.5 + }, + { + "type": "table_body", + "bbox": [ + 107, + 119, + 505, + 479 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 107, + 119, + 505, + 479 + ], + "spans": [ + { + "bbox": [ + 107, + 119, + 505, + 479 + ], + "score": 0.983, + "html": "
NetworkParamsFLOPsTrain sizeTest sizeTop-1(%)Real Top-1 (%)
EfficientNet-B5 [34] SNNO30M9.9B45645683.688.3
EfficientNet-B7 [34]66M37.0B60060084.3
Fix-EfficientNet-B8 [34,38]87M89.5B67280085.790.0
NFNet-F3 [3]255M114.8B32041685.789.4
NFNet-F4 [3]316M215.3B38451285.989.4
NFNet-F5[3]377M289.8B41654486.089.2
ViT-B/16 [15]86M55.4B22438477.983.6
ViT-L/16[15]307M190.7B22438476.582.2
T2T-ViT-14 [46]22M5.2B22422481.5
T2T-ViT-14↑384 [46]22M17.1B22438483.31
Cross ViT [6]45M56.6B22448084.1一
Swin-B[25]88M47.0B22438484.2
TNT-B[16]66M14.1B22422482.8
iriirrrrrs DeepViT-S [59]27M6.2B22422482.3
DeepViT-L [59]55M12.5B22422483.1
DeiT-S[36]22M4.6B22422479.985.7
Distilled DeiT-S [36]22M4.6B22422481.286.8
DeiT-B [36]86M17.5B22422481.886.7
DeiT-B↑384 [36]86M55.4B22438483.187.7
Distilled DeiT-B [36]87M17.5B22422483.488.3
BoTNet-S1-128 [31]79.1M19.3B25625684.2
BoTNet-S1-128↑384 [31]79.1M45.8B25638484.7-
CaiT-S36↑384 [37]68M48.0B22438485.4- 89.8
CaiT-M36[37]271M53.7B22422485.189.3
CaiT-M36↑448 [37]271M247.8B22444886.390.2
LV-ViT-S26M6.6B22422483.3
江 LV-ViT-S↑38426M22.2B22438484.488.1 88.9
LV-ViT-M56M16.0B22422484.188.4
W LV-ViT-M↑38456M42.2B22438485.489.5
LV-ViT-L150M59.0B28828885.389.3
0 LV-ViT-L↑448150M157.2B28844885.989.7
LV-ViT-L↑448150M157.2B44844886.289.9
LV-ViT-L↑512151M214.8B44851286.490.1
", + "type": "table", + "image_path": "1ffe2c2d40127fa5107472db1c5e025901208024a3c26b0c26ca4a1ac6751875.jpg" + } + ] + } + ], + "index": 5, + "virtual_lines": [ + { + "bbox": [ + 107, + 119, + 505, + 239.0 + ], + "spans": [], + "index": 4 + }, + { + "bbox": [ + 107, + 239.0, + 505, + 359.0 + ], + "spans": [], + "index": 5 + }, + { + "bbox": [ + 107, + 359.0, + 505, + 479.0 + ], + "spans": [], + "index": 6 + } + ] + } + ], + "index": 3.25 + }, + { + "type": "title", + "bbox": [ + 107, + 503, + 284, + 515 + ], + "lines": [ + { + "bbox": [ + 105, + 502, + 286, + 517 + ], + "spans": [ + { + "bbox": [ + 105, + 502, + 286, + 517 + ], + "score": 1.0, + "content": "4.4 Semantic Segmentation on ADE20K", + "type": "text" + } + ], + "index": 7 + } + ], + "index": 7 + }, + { + "type": "text", + "bbox": [ + 107, + 525, + 505, + 569 + ], + "lines": [ + { + "bbox": [ + 105, + 525, + 505, + 537 + ], + "spans": [ + { + "bbox": [ + 105, + 525, + 505, + 537 + ], + "score": 1.0, + "content": "It has been shown in [19] that different training techniques for pretrained models have different", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 106, + 537, + 505, + 548 + ], + "spans": [ + { + "bbox": [ + 106, + 537, + 505, + 548 + ], + "score": 1.0, + "content": "impacts on downstream tasks with dense prediction, like semantic segmentation. To demonstrate", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 105, + 547, + 506, + 560 + ], + "spans": [ + { + "bbox": [ + 105, + 547, + 506, + 560 + ], + "score": 1.0, + "content": "the advantage of the proposed token labeling objective on tasks with dense prediction, we apply our", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 558, + 401, + 570 + ], + "spans": [ + { + "bbox": [ + 105, + 558, + 401, + 570 + ], + "score": 1.0, + "content": "pretrained LV-ViT with token labeling to the semantic segmentation task.", + "type": "text" + } + ], + "index": 11 + } + ], + "index": 9.5, + "bbox_fs": [ + 105, + 525, + 506, + 570 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 574, + 505, + 684 + ], + "lines": [ + { + "bbox": [ + 105, + 574, + 506, + 587 + ], + "spans": [ + { + "bbox": [ + 105, + 574, + 506, + 587 + ], + "score": 1.0, + "content": "Similar to previous work [25], we run experiments on the widely-used ADE20K [58] dataset.", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 585, + 506, + 598 + ], + "spans": [ + { + "bbox": [ + 105, + 585, + 506, + 598 + ], + "score": 1.0, + "content": "ADE20K contains 25K images in total, including 20K images for training, 2K images for validation", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 105, + 595, + 506, + 610 + ], + "spans": [ + { + "bbox": [ + 105, + 595, + 506, + 610 + ], + "score": 1.0, + "content": "and 3K images for test, and covering 150 different foreground categories. We take both FCN [26]", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 105, + 606, + 506, + 620 + ], + "spans": [ + { + "bbox": [ + 105, + 606, + 506, + 620 + ], + "score": 1.0, + "content": "and UperNet [44] as our segmentation frameworks and use the mmseg toolbox to implement. During", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 105, + 618, + 506, + 631 + ], + "spans": [ + { + "bbox": [ + 105, + 618, + 506, + 631 + ], + "score": 1.0, + "content": "training, following [25], we use the AdamW optimizer with an initial learning rate of 6e-5 and", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 629, + 506, + 642 + ], + "spans": [ + { + "bbox": [ + 105, + 629, + 506, + 642 + ], + "score": 1.0, + "content": "a weight decay of 0.01. We also use a linear learning schedule with a minimum learning rate of", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 639, + 506, + 652 + ], + "spans": [ + { + "bbox": [ + 105, + 639, + 506, + 652 + ], + "score": 1.0, + "content": "5e-6. All models are trained on 8 GPUs and with a batch size of 16 (i.e., 2 images on each GPU).", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 105, + 650, + 506, + 663 + ], + "spans": [ + { + "bbox": [ + 105, + 650, + 224, + 663 + ], + "score": 1.0, + "content": "The input resolution is set to", + "type": "text" + }, + { + "bbox": [ + 225, + 651, + 268, + 661 + ], + "score": 0.9, + "content": "5 1 2 \\times 5 1 2", + "type": "inline_equation" + }, + { + "bbox": [ + 268, + 650, + 506, + 663 + ], + "score": 1.0, + "content": ". In inference, a multi-scale test with interpolation rates of", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 105, + 662, + 505, + 675 + ], + "spans": [ + { + "bbox": [ + 105, + 662, + 505, + 675 + ], + "score": 1.0, + "content": "[0.75, 1.0, 1.25, 1.5, 1.75] is used. As suggested by [58], we report results in terms of both mean", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 105, + 673, + 412, + 685 + ], + "spans": [ + { + "bbox": [ + 105, + 673, + 412, + 685 + ], + "score": 1.0, + "content": "intersection-over-union (mIoU) and the average pixel accuracy (Pixel Acc.).", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 16.5, + "bbox_fs": [ + 105, + 574, + 506, + 685 + ] + }, + { + "type": "text", + "bbox": [ + 108, + 689, + 505, + 722 + ], + "lines": [ + { + "bbox": [ + 105, + 688, + 505, + 702 + ], + "spans": [ + { + "bbox": [ + 105, + 688, + 505, + 702 + ], + "score": 1.0, + "content": "In Table 7, we test the performance of token labeling on both FCN and UperNet frameworks. The", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 699, + 505, + 713 + ], + "spans": [ + { + "bbox": [ + 105, + 699, + 505, + 713 + ], + "score": 1.0, + "content": "FCN framework has a light convolutional head and can directly reflect the performance of the", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 711, + 505, + 723 + ], + "spans": [ + { + "bbox": [ + 105, + 711, + 505, + 723 + ], + "score": 1.0, + "content": "pretrained models in terms of transferable capability. As can be seen, pretrained models with token", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 105, + 201, + 505, + 215 + ], + "spans": [ + { + "bbox": [ + 105, + 201, + 505, + 215 + ], + "score": 1.0, + "content": "labeling perform better than those without token labeling. This indicates token labeling is indeed", + "type": "text", + "cross_page": true + } + ], + "index": 6 + }, + { + "bbox": [ + 106, + 214, + 253, + 225 + ], + "spans": [ + { + "bbox": [ + 106, + 214, + 253, + 225 + ], + "score": 1.0, + "content": "beneficial to semantic segmentation.", + "type": "text", + "cross_page": true + } + ], + "index": 7 + } + ], + "index": 23, + "bbox_fs": [ + 105, + 688, + 505, + 723 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "table", + "bbox": [ + 109, + 110, + 502, + 181 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 77, + 505, + 110 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 77, + 505, + 89 + ], + "spans": [ + { + "bbox": [ + 105, + 77, + 505, + 89 + ], + "score": 1.0, + "content": "Table 7: Transfer performance of the proposed LV-ViT in semantic segmentation. We take two classic", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 106, + 88, + 505, + 100 + ], + "spans": [ + { + "bbox": [ + 106, + 88, + 505, + 100 + ], + "score": 1.0, + "content": "methods, FCN and UperNet, as segmentation architectures and show both single-scale (SS) and", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 106, + 99, + 291, + 111 + ], + "spans": [ + { + "bbox": [ + 106, + 99, + 291, + 111 + ], + "score": 1.0, + "content": "multi-scale (MS) results on the validation set.", + "type": "text" + } + ], + "index": 2 + } + ], + "index": 1 + }, + { + "type": "table_body", + "bbox": [ + 109, + 110, + 502, + 181 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 109, + 110, + 502, + 181 + ], + "spans": [ + { + "bbox": [ + 109, + 110, + 502, + 181 + ], + "score": 0.977, + "html": "
MethodToken LabelingModel SizemIoU (SS)P. Acc. (SS)mIoU (MS)P. Acc. (MS)
LV-ViT-S + FCN×30M46.181.947.382.6
LV-ViT-S + FCN30M47.282.448.483.0
LV-ViT-S + UperNet44M46.582.147.682.7
LV-ViT-S + UperNetX44M47.982.648.683.1
", + "type": "table", + "image_path": "70c1d4e41ceddea09bd2be3af5115392c9cbb1586a4c7cf96581f2f713f425f0.jpg" + } + ] + } + ], + "index": 4, + "virtual_lines": [ + { + "bbox": [ + 109, + 110, + 502, + 133.66666666666666 + ], + "spans": [], + "index": 3 + }, + { + "bbox": [ + 109, + 133.66666666666666, + 502, + 157.33333333333331 + ], + "spans": [], + "index": 4 + }, + { + "bbox": [ + 109, + 157.33333333333331, + 502, + 180.99999999999997 + ], + "spans": [], + "index": 5 + } + ] + } + ], + "index": 2.5 + }, + { + "type": "text", + "bbox": [ + 107, + 203, + 503, + 225 + ], + "lines": [ + { + "bbox": [ + 105, + 201, + 505, + 215 + ], + "spans": [ + { + "bbox": [ + 105, + 201, + 505, + 215 + ], + "score": 1.0, + "content": "labeling perform better than those without token labeling. This indicates token labeling is indeed", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 106, + 214, + 253, + 225 + ], + "spans": [ + { + "bbox": [ + 106, + 214, + 253, + 225 + ], + "score": 1.0, + "content": "beneficial to semantic segmentation.", + "type": "text" + } + ], + "index": 7 + } + ], + "index": 6.5 + }, + { + "type": "text", + "bbox": [ + 106, + 230, + 505, + 307 + ], + "lines": [ + { + "bbox": [ + 106, + 230, + 505, + 243 + ], + "spans": [ + { + "bbox": [ + 106, + 230, + 505, + 243 + ], + "score": 1.0, + "content": "We also compare our segmentation results with previous state-of-the-art segmentation methods in", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 104, + 240, + 506, + 254 + ], + "spans": [ + { + "bbox": [ + 104, + 240, + 506, + 254 + ], + "score": 1.0, + "content": "Table 8. Without pretraining on large-scale datasets such as ImageNet-22K, our LV-ViT-M with", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 105, + 251, + 507, + 265 + ], + "spans": [ + { + "bbox": [ + 105, + 251, + 507, + 265 + ], + "score": 1.0, + "content": "the UperNet segmentation architecture achieves an mIoU score of 50.6 with only 77M parameters.", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 262, + 507, + 276 + ], + "spans": [ + { + "bbox": [ + 105, + 262, + 507, + 276 + ], + "score": 1.0, + "content": "This result is much better than the previous CNN-based and transformer-based models. Furthermore,", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 273, + 505, + 286 + ], + "spans": [ + { + "bbox": [ + 105, + 273, + 505, + 286 + ], + "score": 1.0, + "content": "using our LV-ViT-L as the pretrained model yields a better result of 51.8 in terms of mIoU. As far as", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 285, + 505, + 297 + ], + "spans": [ + { + "bbox": [ + 105, + 285, + 505, + 297 + ], + "score": 1.0, + "content": "we know, this is the best result reported on ADE20K with no pretraining on ImageNet-22K or other", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 106, + 296, + 188, + 308 + ], + "spans": [ + { + "bbox": [ + 106, + 296, + 188, + 308 + ], + "score": 1.0, + "content": "large-scale datasets.", + "type": "text" + } + ], + "index": 14 + } + ], + "index": 11 + }, + { + "type": "table", + "bbox": [ + 106, + 360, + 505, + 536 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 326, + 506, + 359 + ], + "group_id": 1, + "lines": [ + { + "bbox": [ + 105, + 326, + 506, + 338 + ], + "spans": [ + { + "bbox": [ + 105, + 326, + 506, + 338 + ], + "score": 1.0, + "content": "Table 8: Comparison with previous work on ADE20K validation set. As far as we know, our LV-", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 106, + 337, + 505, + 349 + ], + "spans": [ + { + "bbox": [ + 106, + 337, + 133, + 349 + ], + "score": 1.0, + "content": "ViT-L", + "type": "text" + }, + { + "bbox": [ + 134, + 338, + 142, + 347 + ], + "score": 0.62, + "content": "^ +", + "type": "inline_equation" + }, + { + "bbox": [ + 142, + 337, + 505, + 349 + ], + "score": 1.0, + "content": "UperNet achieves the best result on ADE20K with only ImageNet-1K as training data in", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 348, + 281, + 360 + ], + "spans": [ + { + "bbox": [ + 105, + 348, + 281, + 360 + ], + "score": 1.0, + "content": "pretraining. †Pretrained on ImageNet-22K.", + "type": "text" + } + ], + "index": 17 + } + ], + "index": 16 + }, + { + "type": "table_body", + "bbox": [ + 106, + 360, + 505, + 536 + ], + "group_id": 1, + "lines": [ + { + "bbox": [ + 106, + 360, + 505, + 536 + ], + "spans": [ + { + "bbox": [ + 106, + 360, + 505, + 536 + ], + "score": 0.984, + "html": "
BackboneSegmentation ArchitectureModel SizemIoU (MS)Pixel Acc. (MS)
SNNOResNet-269PSPNet [54]44.981.7
ResNet-101UperNet [44]86M44.9-
ResNet-101Strip Pooling [22]45.682.1
ResNeSt200DeepLabV3+ [8]88M48.41
DeiT-SUperNet52M44.0-
TirriiirrssViT-LargetSETR [56]308M50.383.5
Swin-T[25]UperNet60M46.11
Swin-S [25]UperNet81M49.3=
Swin-B [25]UperNet121M49.7
Swin-B† [25]UperNet121M51.6-
LIA-ATLV-ViT-SFCN30M48.483.0
LV-ViT-SUperNet44M48.683.1
LV-ViT-MUperNet77M50.683.5
LV-ViT-LUperNet209M51.884.1
", + "type": "table", + "image_path": "7675f5b12a93d9e094cea27ffe903eaa1865bea2f706e0842c5d4c04af07ac46.jpg" + } + ] + } + ], + "index": 19, + "virtual_lines": [ + { + "bbox": [ + 106, + 360, + 505, + 418.6666666666667 + ], + "spans": [], + "index": 18 + }, + { + "bbox": [ + 106, + 418.6666666666667, + 505, + 477.33333333333337 + ], + "spans": [], + "index": 19 + }, + { + "bbox": [ + 106, + 477.33333333333337, + 505, + 536.0 + ], + "spans": [], + "index": 20 + } + ] + } + ], + "index": 17.5 + }, + { + "type": "title", + "bbox": [ + 107, + 561, + 267, + 576 + ], + "lines": [ + { + "bbox": [ + 104, + 560, + 268, + 577 + ], + "spans": [ + { + "bbox": [ + 104, + 560, + 268, + 577 + ], + "score": 1.0, + "content": "5 Conclusions and Discussion", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 21 + }, + { + "type": "text", + "bbox": [ + 107, + 587, + 505, + 643 + ], + "lines": [ + { + "bbox": [ + 105, + 587, + 505, + 601 + ], + "spans": [ + { + "bbox": [ + 105, + 587, + 505, + 601 + ], + "score": 1.0, + "content": "In this paper, we introduce a new token labeling method to help improve the performance of vision", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 598, + 505, + 612 + ], + "spans": [ + { + "bbox": [ + 105, + 598, + 505, + 612 + ], + "score": 1.0, + "content": "transformers. We also analyze the effectiveness and robustness of our token labeling with respect", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 609, + 505, + 622 + ], + "spans": [ + { + "bbox": [ + 105, + 609, + 505, + 622 + ], + "score": 1.0, + "content": "to different annotators and different variants of patch-based vision transformers. By applying token", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 105, + 621, + 505, + 634 + ], + "spans": [ + { + "bbox": [ + 105, + 621, + 264, + 634 + ], + "score": 1.0, + "content": "labeling, our proposed LV-ViT achieves", + "type": "text" + }, + { + "bbox": [ + 264, + 621, + 291, + 631 + ], + "score": 0.86, + "content": "8 4 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 292, + 621, + 478, + 634 + ], + "score": 1.0, + "content": "Top-1 accuracy with only 26M parameters and", + "type": "text" + }, + { + "bbox": [ + 478, + 621, + 505, + 631 + ], + "score": 0.84, + "content": "8 6 . 4 \\%", + "type": "inline_equation" + } + ], + "index": 25 + }, + { + "bbox": [ + 105, + 631, + 381, + 645 + ], + "spans": [ + { + "bbox": [ + 105, + 631, + 381, + 645 + ], + "score": 1.0, + "content": "Top-1 accuracy with 150M parameters on ImageNet-1K benchmark.", + "type": "text" + } + ], + "index": 26 + } + ], + "index": 24 + }, + { + "type": "text", + "bbox": [ + 107, + 648, + 505, + 714 + ], + "lines": [ + { + "bbox": [ + 105, + 648, + 505, + 660 + ], + "spans": [ + { + "bbox": [ + 105, + 648, + 505, + 660 + ], + "score": 1.0, + "content": "Despite the effectiveness, token labeling has a limitation of requiring a pretrained model as the", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 105, + 658, + 506, + 672 + ], + "spans": [ + { + "bbox": [ + 105, + 658, + 506, + 672 + ], + "score": 1.0, + "content": "machine annotator. Fortunately, the machine annotating procedure can be done in advance to", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 670, + 505, + 682 + ], + "spans": [ + { + "bbox": [ + 105, + 670, + 505, + 682 + ], + "score": 1.0, + "content": "avoid introducing extra computational cost in training. This makes our method quite different from", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 680, + 506, + 694 + ], + "spans": [ + { + "bbox": [ + 105, + 680, + 506, + 694 + ], + "score": 1.0, + "content": "knowledge distillation methods that rely on online teaching. For users with limited machine resources", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 691, + 506, + 704 + ], + "spans": [ + { + "bbox": [ + 105, + 691, + 506, + 704 + ], + "score": 1.0, + "content": "on hand, our token labeling provides a promising training technique to improve the performance of", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 703, + 189, + 715 + ], + "spans": [ + { + "bbox": [ + 105, + 703, + 189, + 715 + ], + "score": 1.0, + "content": "vision transformers.", + "type": "text" + } + ], + "index": 32 + } + ], + "index": 29.5 + } + ], + "page_idx": 9, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 300, + 741, + 311, + 750 + ], + "lines": [ + { + "bbox": [ + 299, + 740, + 313, + 754 + ], + "spans": [ + { + "bbox": [ + 299, + 740, + 313, + 754 + ], + "score": 1.0, + "content": "10", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "table", + "bbox": [ + 109, + 110, + 502, + 181 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 77, + 505, + 110 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 105, + 77, + 505, + 89 + ], + "spans": [ + { + "bbox": [ + 105, + 77, + 505, + 89 + ], + "score": 1.0, + "content": "Table 7: Transfer performance of the proposed LV-ViT in semantic segmentation. We take two classic", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 106, + 88, + 505, + 100 + ], + "spans": [ + { + "bbox": [ + 106, + 88, + 505, + 100 + ], + "score": 1.0, + "content": "methods, FCN and UperNet, as segmentation architectures and show both single-scale (SS) and", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 106, + 99, + 291, + 111 + ], + "spans": [ + { + "bbox": [ + 106, + 99, + 291, + 111 + ], + "score": 1.0, + "content": "multi-scale (MS) results on the validation set.", + "type": "text" + } + ], + "index": 2 + } + ], + "index": 1 + }, + { + "type": "table_body", + "bbox": [ + 109, + 110, + 502, + 181 + ], + "group_id": 0, + "lines": [ + { + "bbox": [ + 109, + 110, + 502, + 181 + ], + "spans": [ + { + "bbox": [ + 109, + 110, + 502, + 181 + ], + "score": 0.977, + "html": "
MethodToken LabelingModel SizemIoU (SS)P. Acc. (SS)mIoU (MS)P. Acc. (MS)
LV-ViT-S + FCN×30M46.181.947.382.6
LV-ViT-S + FCN30M47.282.448.483.0
LV-ViT-S + UperNet44M46.582.147.682.7
LV-ViT-S + UperNetX44M47.982.648.683.1
", + "type": "table", + "image_path": "70c1d4e41ceddea09bd2be3af5115392c9cbb1586a4c7cf96581f2f713f425f0.jpg" + } + ] + } + ], + "index": 4, + "virtual_lines": [ + { + "bbox": [ + 109, + 110, + 502, + 133.66666666666666 + ], + "spans": [], + "index": 3 + }, + { + "bbox": [ + 109, + 133.66666666666666, + 502, + 157.33333333333331 + ], + "spans": [], + "index": 4 + }, + { + "bbox": [ + 109, + 157.33333333333331, + 502, + 180.99999999999997 + ], + "spans": [], + "index": 5 + } + ] + } + ], + "index": 2.5 + }, + { + "type": "text", + "bbox": [ + 107, + 203, + 503, + 225 + ], + "lines": [], + "index": 6.5, + "bbox_fs": [ + 105, + 201, + 505, + 225 + ], + "lines_deleted": true + }, + { + "type": "text", + "bbox": [ + 106, + 230, + 505, + 307 + ], + "lines": [ + { + "bbox": [ + 106, + 230, + 505, + 243 + ], + "spans": [ + { + "bbox": [ + 106, + 230, + 505, + 243 + ], + "score": 1.0, + "content": "We also compare our segmentation results with previous state-of-the-art segmentation methods in", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 104, + 240, + 506, + 254 + ], + "spans": [ + { + "bbox": [ + 104, + 240, + 506, + 254 + ], + "score": 1.0, + "content": "Table 8. Without pretraining on large-scale datasets such as ImageNet-22K, our LV-ViT-M with", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 105, + 251, + 507, + 265 + ], + "spans": [ + { + "bbox": [ + 105, + 251, + 507, + 265 + ], + "score": 1.0, + "content": "the UperNet segmentation architecture achieves an mIoU score of 50.6 with only 77M parameters.", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 105, + 262, + 507, + 276 + ], + "spans": [ + { + "bbox": [ + 105, + 262, + 507, + 276 + ], + "score": 1.0, + "content": "This result is much better than the previous CNN-based and transformer-based models. Furthermore,", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 273, + 505, + 286 + ], + "spans": [ + { + "bbox": [ + 105, + 273, + 505, + 286 + ], + "score": 1.0, + "content": "using our LV-ViT-L as the pretrained model yields a better result of 51.8 in terms of mIoU. As far as", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 105, + 285, + 505, + 297 + ], + "spans": [ + { + "bbox": [ + 105, + 285, + 505, + 297 + ], + "score": 1.0, + "content": "we know, this is the best result reported on ADE20K with no pretraining on ImageNet-22K or other", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 106, + 296, + 188, + 308 + ], + "spans": [ + { + "bbox": [ + 106, + 296, + 188, + 308 + ], + "score": 1.0, + "content": "large-scale datasets.", + "type": "text" + } + ], + "index": 14 + } + ], + "index": 11, + "bbox_fs": [ + 104, + 230, + 507, + 308 + ] + }, + { + "type": "table", + "bbox": [ + 106, + 360, + 505, + 536 + ], + "blocks": [ + { + "type": "table_caption", + "bbox": [ + 106, + 326, + 506, + 359 + ], + "group_id": 1, + "lines": [ + { + "bbox": [ + 105, + 326, + 506, + 338 + ], + "spans": [ + { + "bbox": [ + 105, + 326, + 506, + 338 + ], + "score": 1.0, + "content": "Table 8: Comparison with previous work on ADE20K validation set. As far as we know, our LV-", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 106, + 337, + 505, + 349 + ], + "spans": [ + { + "bbox": [ + 106, + 337, + 133, + 349 + ], + "score": 1.0, + "content": "ViT-L", + "type": "text" + }, + { + "bbox": [ + 134, + 338, + 142, + 347 + ], + "score": 0.62, + "content": "^ +", + "type": "inline_equation" + }, + { + "bbox": [ + 142, + 337, + 505, + 349 + ], + "score": 1.0, + "content": "UperNet achieves the best result on ADE20K with only ImageNet-1K as training data in", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 105, + 348, + 281, + 360 + ], + "spans": [ + { + "bbox": [ + 105, + 348, + 281, + 360 + ], + "score": 1.0, + "content": "pretraining. †Pretrained on ImageNet-22K.", + "type": "text" + } + ], + "index": 17 + } + ], + "index": 16 + }, + { + "type": "table_body", + "bbox": [ + 106, + 360, + 505, + 536 + ], + "group_id": 1, + "lines": [ + { + "bbox": [ + 106, + 360, + 505, + 536 + ], + "spans": [ + { + "bbox": [ + 106, + 360, + 505, + 536 + ], + "score": 0.984, + "html": "
BackboneSegmentation ArchitectureModel SizemIoU (MS)Pixel Acc. (MS)
SNNOResNet-269PSPNet [54]44.981.7
ResNet-101UperNet [44]86M44.9-
ResNet-101Strip Pooling [22]45.682.1
ResNeSt200DeepLabV3+ [8]88M48.41
DeiT-SUperNet52M44.0-
TirriiirrssViT-LargetSETR [56]308M50.383.5
Swin-T[25]UperNet60M46.11
Swin-S [25]UperNet81M49.3=
Swin-B [25]UperNet121M49.7
Swin-B† [25]UperNet121M51.6-
LIA-ATLV-ViT-SFCN30M48.483.0
LV-ViT-SUperNet44M48.683.1
LV-ViT-MUperNet77M50.683.5
LV-ViT-LUperNet209M51.884.1
", + "type": "table", + "image_path": "7675f5b12a93d9e094cea27ffe903eaa1865bea2f706e0842c5d4c04af07ac46.jpg" + } + ] + } + ], + "index": 19, + "virtual_lines": [ + { + "bbox": [ + 106, + 360, + 505, + 418.6666666666667 + ], + "spans": [], + "index": 18 + }, + { + "bbox": [ + 106, + 418.6666666666667, + 505, + 477.33333333333337 + ], + "spans": [], + "index": 19 + }, + { + "bbox": [ + 106, + 477.33333333333337, + 505, + 536.0 + ], + "spans": [], + "index": 20 + } + ] + } + ], + "index": 17.5 + }, + { + "type": "title", + "bbox": [ + 107, + 561, + 267, + 576 + ], + "lines": [ + { + "bbox": [ + 104, + 560, + 268, + 577 + ], + "spans": [ + { + "bbox": [ + 104, + 560, + 268, + 577 + ], + "score": 1.0, + "content": "5 Conclusions and Discussion", + "type": "text" + } + ], + "index": 21 + } + ], + "index": 21 + }, + { + "type": "text", + "bbox": [ + 107, + 587, + 505, + 643 + ], + "lines": [ + { + "bbox": [ + 105, + 587, + 505, + 601 + ], + "spans": [ + { + "bbox": [ + 105, + 587, + 505, + 601 + ], + "score": 1.0, + "content": "In this paper, we introduce a new token labeling method to help improve the performance of vision", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 598, + 505, + 612 + ], + "spans": [ + { + "bbox": [ + 105, + 598, + 505, + 612 + ], + "score": 1.0, + "content": "transformers. We also analyze the effectiveness and robustness of our token labeling with respect", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 105, + 609, + 505, + 622 + ], + "spans": [ + { + "bbox": [ + 105, + 609, + 505, + 622 + ], + "score": 1.0, + "content": "to different annotators and different variants of patch-based vision transformers. By applying token", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 105, + 621, + 505, + 634 + ], + "spans": [ + { + "bbox": [ + 105, + 621, + 264, + 634 + ], + "score": 1.0, + "content": "labeling, our proposed LV-ViT achieves", + "type": "text" + }, + { + "bbox": [ + 264, + 621, + 291, + 631 + ], + "score": 0.86, + "content": "8 4 . 4 \\%", + "type": "inline_equation" + }, + { + "bbox": [ + 292, + 621, + 478, + 634 + ], + "score": 1.0, + "content": "Top-1 accuracy with only 26M parameters and", + "type": "text" + }, + { + "bbox": [ + 478, + 621, + 505, + 631 + ], + "score": 0.84, + "content": "8 6 . 4 \\%", + "type": "inline_equation" + } + ], + "index": 25 + }, + { + "bbox": [ + 105, + 631, + 381, + 645 + ], + "spans": [ + { + "bbox": [ + 105, + 631, + 381, + 645 + ], + "score": 1.0, + "content": "Top-1 accuracy with 150M parameters on ImageNet-1K benchmark.", + "type": "text" + } + ], + "index": 26 + } + ], + "index": 24, + "bbox_fs": [ + 105, + 587, + 505, + 645 + ] + }, + { + "type": "text", + "bbox": [ + 107, + 648, + 505, + 714 + ], + "lines": [ + { + "bbox": [ + 105, + 648, + 505, + 660 + ], + "spans": [ + { + "bbox": [ + 105, + 648, + 505, + 660 + ], + "score": 1.0, + "content": "Despite the effectiveness, token labeling has a limitation of requiring a pretrained model as the", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 105, + 658, + 506, + 672 + ], + "spans": [ + { + "bbox": [ + 105, + 658, + 506, + 672 + ], + "score": 1.0, + "content": "machine annotator. Fortunately, the machine annotating procedure can be done in advance to", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 105, + 670, + 505, + 682 + ], + "spans": [ + { + "bbox": [ + 105, + 670, + 505, + 682 + ], + "score": 1.0, + "content": "avoid introducing extra computational cost in training. This makes our method quite different from", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 105, + 680, + 506, + 694 + ], + "spans": [ + { + "bbox": [ + 105, + 680, + 506, + 694 + ], + "score": 1.0, + "content": "knowledge distillation methods that rely on online teaching. For users with limited machine resources", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 105, + 691, + 506, + 704 + ], + "spans": [ + { + "bbox": [ + 105, + 691, + 506, + 704 + ], + "score": 1.0, + "content": "on hand, our token labeling provides a promising training technique to improve the performance of", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 105, + 703, + 189, + 715 + ], + "spans": [ + { + "bbox": [ + 105, + 703, + 189, + 715 + ], + "score": 1.0, + "content": "vision transformers.", + "type": "text" + } + ], + "index": 32 + } + ], + "index": 29.5, + "bbox_fs": [ + 105, + 648, + 506, + 715 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "text", + "bbox": [ + 105, + 69, + 507, + 728 + ], + "lines": [ + { + "bbox": [ + 106, + 70, + 165, + 86 + ], + "spans": [ + { + "bbox": [ + 106, + 70, + 165, + 86 + ], + "score": 1.0, + "content": "References", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 109, + 89, + 506, + 102 + ], + "spans": [ + { + "bbox": [ + 109, + 89, + 506, + 102 + ], + "score": 1.0, + "content": "[1] Irwan Bello. Lambdanetworks: Modeling long-range interactions without attention. arXiv preprint", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 125, + 100, + 220, + 111 + ], + "spans": [ + { + "bbox": [ + 125, + 100, + 220, + 111 + ], + "score": 1.0, + "content": "arXiv:2102.08602, 2021.", + "type": "text" + } + ], + "index": 2 + }, + { + "bbox": [ + 109, + 116, + 506, + 130 + ], + "spans": [ + { + "bbox": [ + 109, + 116, + 506, + 130 + ], + "score": 1.0, + "content": "[2] Lucas Beyer, Olivier J Hénaff, Alexander Kolesnikov, Xiaohua Zhai, and Aäron van den Oord. Are we", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 126, + 127, + 351, + 138 + ], + "spans": [ + { + "bbox": [ + 126, + 127, + 351, + 138 + ], + "score": 1.0, + "content": "done with imagenet? arXiv preprint arXiv:2006.07159, 2020.", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 107, + 142, + 506, + 159 + ], + "spans": [ + { + "bbox": [ + 107, + 142, + 506, + 159 + ], + "score": 1.0, + "content": "[3] Andrew Brock, Soham De, Samuel L Smith, and Karen Simonyan. High-performance large-scale image", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 124, + 154, + 401, + 167 + ], + "spans": [ + { + "bbox": [ + 124, + 154, + 401, + 167 + ], + "score": 1.0, + "content": "recognition without normalization. arXiv preprint arXiv:2102.06171, 2021.", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 109, + 171, + 506, + 185 + ], + "spans": [ + { + "bbox": [ + 109, + 171, + 506, + 185 + ], + "score": 1.0, + "content": "[4] Tom B Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 124, + 181, + 507, + 195 + ], + "spans": [ + { + "bbox": [ + 124, + 181, + 507, + 195 + ], + "score": 1.0, + "content": "Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners.", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 126, + 192, + 273, + 204 + ], + "spans": [ + { + "bbox": [ + 126, + 192, + 273, + 204 + ], + "score": 1.0, + "content": "arXiv preprint arXiv:2005.14165, 2020.", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 108, + 208, + 506, + 223 + ], + "spans": [ + { + "bbox": [ + 108, + 208, + 506, + 223 + ], + "score": 1.0, + "content": "[5] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 124, + 219, + 488, + 232 + ], + "spans": [ + { + "bbox": [ + 124, + 219, + 488, + 232 + ], + "score": 1.0, + "content": "Zagoruyko. End-to-end object detection with transformers. arXiv preprint arXiv:2005.12872, 2020.", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 110, + 237, + 505, + 249 + ], + "spans": [ + { + "bbox": [ + 110, + 237, + 505, + 249 + ], + "score": 1.0, + "content": "[6] Chun-Fu Chen, Quanfu Fan, and Rameswar Panda. Crossvit: Cross-attention multi-scale vision transformer", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 125, + 247, + 362, + 259 + ], + "spans": [ + { + "bbox": [ + 125, + 247, + 362, + 259 + ], + "score": 1.0, + "content": "for image classification. arXiv preprint arXiv:2103.14899, 2021.", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 109, + 264, + 506, + 278 + ], + "spans": [ + { + "bbox": [ + 109, + 264, + 506, + 278 + ], + "score": 1.0, + "content": "[7] Hanting Chen, Yunhe Wang, Tianyu Guo, Chang Xu, Yiping Deng, Zhenhua Liu, Siwei Ma, Chunjing", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 124, + 274, + 506, + 288 + ], + "spans": [ + { + "bbox": [ + 124, + 274, + 506, + 288 + ], + "score": 1.0, + "content": "Xu, Chao Xu, and Wen Gao. Pre-trained image processing transformer. arXiv preprint arXiv:2012.00364,", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 126, + 284, + 150, + 297 + ], + "spans": [ + { + "bbox": [ + 126, + 284, + 150, + 297 + ], + "score": 1.0, + "content": "2020.", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 110, + 302, + 505, + 314 + ], + "spans": [ + { + "bbox": [ + 110, + 302, + 505, + 314 + ], + "score": 1.0, + "content": "[8] Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. Encoder-decoder", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 124, + 311, + 506, + 325 + ], + "spans": [ + { + "bbox": [ + 124, + 311, + 506, + 325 + ], + "score": 1.0, + "content": "with atrous separable convolution for semantic image segmentation. In Proceedings of the European", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 124, + 322, + 354, + 334 + ], + "spans": [ + { + "bbox": [ + 124, + 322, + 354, + 334 + ], + "score": 1.0, + "content": "conference on computer vision (ECCV), pages 801–818, 2018.", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 109, + 339, + 507, + 352 + ], + "spans": [ + { + "bbox": [ + 109, + 339, + 507, + 352 + ], + "score": 1.0, + "content": "[9] Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, and Ilya Sutskever.", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 124, + 348, + 508, + 363 + ], + "spans": [ + { + "bbox": [ + 124, + 348, + 508, + 363 + ], + "score": 1.0, + "content": "Generative pretraining from pixels. In International Conference on Machine Learning, pages 1691–1703.", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 124, + 359, + 179, + 371 + ], + "spans": [ + { + "bbox": [ + 124, + 359, + 179, + 371 + ], + "score": 1.0, + "content": "PMLR, 2020.", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 106, + 377, + 506, + 388 + ], + "spans": [ + { + "bbox": [ + 106, + 377, + 506, + 388 + ], + "score": 1.0, + "content": "[10] Ekin D Cubuk, Barret Zoph, Jonathon Shlens, and Quoc V Le. Randaugment: Practical automated data", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 124, + 386, + 506, + 400 + ], + "spans": [ + { + "bbox": [ + 124, + 386, + 506, + 400 + ], + "score": 1.0, + "content": "augmentation with a reduced search space. In Proceedings of the IEEE/CVF Conference on Computer", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 126, + 397, + 365, + 409 + ], + "spans": [ + { + "bbox": [ + 126, + 397, + 365, + 409 + ], + "score": 1.0, + "content": "Vision and Pattern Recognition Workshops, pages 702–703, 2020.", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 105, + 414, + 506, + 428 + ], + "spans": [ + { + "bbox": [ + 105, + 414, + 506, + 428 + ], + "score": 1.0, + "content": "[11] Zhigang Dai, Bolun Cai, Yugeng Lin, and Junying Chen. Up-detr: Unsupervised pre-training for object", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 124, + 424, + 378, + 438 + ], + "spans": [ + { + "bbox": [ + 124, + 424, + 378, + 438 + ], + "score": 1.0, + "content": "detection with transformers. arXiv preprint arXiv:2011.09094, 2020.", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 106, + 442, + 506, + 453 + ], + "spans": [ + { + "bbox": [ + 106, + 442, + 506, + 453 + ], + "score": 1.0, + "content": "[12] Stéphane d’Ascoli, Hugo Touvron, Matthew Leavitt, Ari Morcos, Giulio Biroli, and Levent Sagun. Convit:", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 126, + 452, + 506, + 464 + ], + "spans": [ + { + "bbox": [ + 126, + 452, + 506, + 464 + ], + "score": 1.0, + "content": "Improving vision transformers with soft convolutional inductive biases. arXiv preprint arXiv:2103.10697,", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 125, + 461, + 152, + 474 + ], + "spans": [ + { + "bbox": [ + 125, + 461, + 152, + 474 + ], + "score": 1.0, + "content": "2021.", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 106, + 480, + 505, + 491 + ], + "spans": [ + { + "bbox": [ + 106, + 480, + 505, + 491 + ], + "score": 1.0, + "content": "[13] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 125, + 488, + 507, + 502 + ], + "spans": [ + { + "bbox": [ + 125, + 488, + 507, + 502 + ], + "score": 1.0, + "content": "image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255.", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 125, + 499, + 169, + 511 + ], + "spans": [ + { + "bbox": [ + 125, + 499, + 169, + 511 + ], + "score": 1.0, + "content": "Ieee, 2009.", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 105, + 516, + 507, + 530 + ], + "spans": [ + { + "bbox": [ + 105, + 516, + 507, + 530 + ], + "score": 1.0, + "content": "[14] Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirec-", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 125, + 526, + 447, + 540 + ], + "spans": [ + { + "bbox": [ + 125, + 526, + 447, + 540 + ], + "score": 1.0, + "content": "tional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018.", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 106, + 544, + 506, + 556 + ], + "spans": [ + { + "bbox": [ + 106, + 544, + 506, + 556 + ], + "score": 1.0, + "content": "[15] Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 125, + 554, + 506, + 567 + ], + "spans": [ + { + "bbox": [ + 125, + 554, + 506, + 567 + ], + "score": 1.0, + "content": "Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 125, + 564, + 486, + 577 + ], + "spans": [ + { + "bbox": [ + 125, + 564, + 486, + 577 + ], + "score": 1.0, + "content": "16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020.", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 106, + 582, + 506, + 594 + ], + "spans": [ + { + "bbox": [ + 106, + 582, + 506, + 594 + ], + "score": 1.0, + "content": "[16] Kai Han, An Xiao, Enhua Wu, Jianyuan Guo, Chunjing Xu, and Yunhe Wang. Transformer in transformer.", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 126, + 592, + 272, + 604 + ], + "spans": [ + { + "bbox": [ + 126, + 592, + 272, + 604 + ], + "score": 1.0, + "content": "arXiv preprint arXiv:2103.00112, 2021.", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 106, + 609, + 504, + 621 + ], + "spans": [ + { + "bbox": [ + 106, + 609, + 504, + 621 + ], + "score": 1.0, + "content": "[17] Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. Mask r-cnn. In Proceedings of the IEEE", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 125, + 619, + 381, + 632 + ], + "spans": [ + { + "bbox": [ + 125, + 619, + 381, + 632 + ], + "score": 1.0, + "content": "international conference on computer vision, pages 2961–2969, 2017.", + "type": "text" + } + ], + "index": 42 + }, + { + "bbox": [ + 104, + 636, + 507, + 649 + ], + "spans": [ + { + "bbox": [ + 104, + 636, + 507, + 649 + ], + "score": 1.0, + "content": "[18] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition.", + "type": "text" + } + ], + "index": 43 + }, + { + "bbox": [ + 124, + 646, + 506, + 660 + ], + "spans": [ + { + "bbox": [ + 124, + 646, + 506, + 660 + ], + "score": 1.0, + "content": "In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016.", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 104, + 663, + 506, + 677 + ], + "spans": [ + { + "bbox": [ + 104, + 663, + 506, + 677 + ], + "score": 1.0, + "content": "[19] Tong He, Zhi Zhang, Hang Zhang, Zhongyue Zhang, Junyuan Xie, and Mu Li. Bag of tricks for image", + "type": "text" + } + ], + "index": 45 + }, + { + "bbox": [ + 125, + 674, + 506, + 687 + ], + "spans": [ + { + "bbox": [ + 125, + 674, + 506, + 687 + ], + "score": 1.0, + "content": "classification with convolutional neural networks. In Proceedings of the IEEE/CVF Conference on", + "type": "text" + } + ], + "index": 46 + }, + { + "bbox": [ + 125, + 684, + 362, + 698 + ], + "spans": [ + { + "bbox": [ + 125, + 684, + 362, + 698 + ], + "score": 1.0, + "content": "Computer Vision and Pattern Recognition, pages 558–567, 2019.", + "type": "text" + } + ], + "index": 47 + }, + { + "bbox": [ + 105, + 700, + 507, + 714 + ], + "spans": [ + { + "bbox": [ + 105, + 700, + 507, + 714 + ], + "score": 1.0, + "content": "[20] Byeongho Heo, Sangdoo Yun, Dongyoon Han, Sanghyuk Chun, Junsuk Choe, and Seong Joon Oh.", + "type": "text" + } + ], + "index": 48 + }, + { + "bbox": [ + 125, + 711, + 469, + 725 + ], + "spans": [ + { + "bbox": [ + 125, + 711, + 469, + 725 + ], + "score": 1.0, + "content": "Rethinking spatial dimensions of vision transformers. arXiv preprint arXiv:2103.16302, 2021.", + "type": "text" + } + ], + "index": 49 + } + ], + "index": 24.5 + } + ], + "page_idx": 10, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 300, + 741, + 310, + 750 + ], + "lines": [ + { + "bbox": [ + 299, + 740, + 312, + 755 + ], + "spans": [ + { + "bbox": [ + 299, + 740, + 312, + 755 + ], + "score": 1.0, + "content": "", + "type": "text", + "height": 15, + "width": 13 + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "list", + "bbox": [ + 105, + 69, + 507, + 728 + ], + "lines": [ + { + "bbox": [ + 106, + 70, + 165, + 86 + ], + "spans": [ + { + "bbox": [ + 106, + 70, + 165, + 86 + ], + "score": 1.0, + "content": "References", + "type": "text" + } + ], + "index": 0, + "is_list_start_line": true, + "is_list_end_line": true + }, + { + "bbox": [ + 109, + 89, + 506, + 102 + ], + "spans": [ + { + "bbox": [ + 109, + 89, + 506, + 102 + ], + "score": 1.0, + "content": "[1] Irwan Bello. Lambdanetworks: Modeling long-range interactions without attention. arXiv preprint", + "type": "text" + } + ], + "index": 1, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 100, + 220, + 111 + ], + "spans": [ + { + "bbox": [ + 125, + 100, + 220, + 111 + ], + "score": 1.0, + "content": "arXiv:2102.08602, 2021.", + "type": "text" + } + ], + "index": 2, + "is_list_end_line": true + }, + { + "bbox": [ + 109, + 116, + 506, + 130 + ], + "spans": [ + { + "bbox": [ + 109, + 116, + 506, + 130 + ], + "score": 1.0, + "content": "[2] Lucas Beyer, Olivier J Hénaff, Alexander Kolesnikov, Xiaohua Zhai, and Aäron van den Oord. Are we", + "type": "text" + } + ], + "index": 3, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 127, + 351, + 138 + ], + "spans": [ + { + "bbox": [ + 126, + 127, + 351, + 138 + ], + "score": 1.0, + "content": "done with imagenet? arXiv preprint arXiv:2006.07159, 2020.", + "type": "text" + } + ], + "index": 4, + "is_list_end_line": true + }, + { + "bbox": [ + 107, + 142, + 506, + 159 + ], + "spans": [ + { + "bbox": [ + 107, + 142, + 506, + 159 + ], + "score": 1.0, + "content": "[3] Andrew Brock, Soham De, Samuel L Smith, and Karen Simonyan. High-performance large-scale image", + "type": "text" + } + ], + "index": 5, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 154, + 401, + 167 + ], + "spans": [ + { + "bbox": [ + 124, + 154, + 401, + 167 + ], + "score": 1.0, + "content": "recognition without normalization. arXiv preprint arXiv:2102.06171, 2021.", + "type": "text" + } + ], + "index": 6, + "is_list_end_line": true + }, + { + "bbox": [ + 109, + 171, + 506, + 185 + ], + "spans": [ + { + "bbox": [ + 109, + 171, + 506, + 185 + ], + "score": 1.0, + "content": "[4] Tom B Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind", + "type": "text" + } + ], + "index": 7, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 181, + 507, + 195 + ], + "spans": [ + { + "bbox": [ + 124, + 181, + 507, + 195 + ], + "score": 1.0, + "content": "Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners.", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 126, + 192, + 273, + 204 + ], + "spans": [ + { + "bbox": [ + 126, + 192, + 273, + 204 + ], + "score": 1.0, + "content": "arXiv preprint arXiv:2005.14165, 2020.", + "type": "text" + } + ], + "index": 9, + "is_list_end_line": true + }, + { + "bbox": [ + 108, + 208, + 506, + 223 + ], + "spans": [ + { + "bbox": [ + 108, + 208, + 506, + 223 + ], + "score": 1.0, + "content": "[5] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey", + "type": "text" + } + ], + "index": 10, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 219, + 488, + 232 + ], + "spans": [ + { + "bbox": [ + 124, + 219, + 488, + 232 + ], + "score": 1.0, + "content": "Zagoruyko. End-to-end object detection with transformers. arXiv preprint arXiv:2005.12872, 2020.", + "type": "text" + } + ], + "index": 11, + "is_list_end_line": true + }, + { + "bbox": [ + 110, + 237, + 505, + 249 + ], + "spans": [ + { + "bbox": [ + 110, + 237, + 505, + 249 + ], + "score": 1.0, + "content": "[6] Chun-Fu Chen, Quanfu Fan, and Rameswar Panda. Crossvit: Cross-attention multi-scale vision transformer", + "type": "text" + } + ], + "index": 12, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 247, + 362, + 259 + ], + "spans": [ + { + "bbox": [ + 125, + 247, + 362, + 259 + ], + "score": 1.0, + "content": "for image classification. arXiv preprint arXiv:2103.14899, 2021.", + "type": "text" + } + ], + "index": 13, + "is_list_end_line": true + }, + { + "bbox": [ + 109, + 264, + 506, + 278 + ], + "spans": [ + { + "bbox": [ + 109, + 264, + 506, + 278 + ], + "score": 1.0, + "content": "[7] Hanting Chen, Yunhe Wang, Tianyu Guo, Chang Xu, Yiping Deng, Zhenhua Liu, Siwei Ma, Chunjing", + "type": "text" + } + ], + "index": 14, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 274, + 506, + 288 + ], + "spans": [ + { + "bbox": [ + 124, + 274, + 506, + 288 + ], + "score": 1.0, + "content": "Xu, Chao Xu, and Wen Gao. Pre-trained image processing transformer. arXiv preprint arXiv:2012.00364,", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 126, + 284, + 150, + 297 + ], + "spans": [ + { + "bbox": [ + 126, + 284, + 150, + 297 + ], + "score": 1.0, + "content": "2020.", + "type": "text" + } + ], + "index": 16, + "is_list_end_line": true + }, + { + "bbox": [ + 110, + 302, + 505, + 314 + ], + "spans": [ + { + "bbox": [ + 110, + 302, + 505, + 314 + ], + "score": 1.0, + "content": "[8] Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. Encoder-decoder", + "type": "text" + } + ], + "index": 17, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 311, + 506, + 325 + ], + "spans": [ + { + "bbox": [ + 124, + 311, + 506, + 325 + ], + "score": 1.0, + "content": "with atrous separable convolution for semantic image segmentation. In Proceedings of the European", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 124, + 322, + 354, + 334 + ], + "spans": [ + { + "bbox": [ + 124, + 322, + 354, + 334 + ], + "score": 1.0, + "content": "conference on computer vision (ECCV), pages 801–818, 2018.", + "type": "text" + } + ], + "index": 19, + "is_list_end_line": true + }, + { + "bbox": [ + 109, + 339, + 507, + 352 + ], + "spans": [ + { + "bbox": [ + 109, + 339, + 507, + 352 + ], + "score": 1.0, + "content": "[9] Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, and Ilya Sutskever.", + "type": "text" + } + ], + "index": 20, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 348, + 508, + 363 + ], + "spans": [ + { + "bbox": [ + 124, + 348, + 508, + 363 + ], + "score": 1.0, + "content": "Generative pretraining from pixels. In International Conference on Machine Learning, pages 1691–1703.", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 124, + 359, + 179, + 371 + ], + "spans": [ + { + "bbox": [ + 124, + 359, + 179, + 371 + ], + "score": 1.0, + "content": "PMLR, 2020.", + "type": "text" + } + ], + "index": 22, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 377, + 506, + 388 + ], + "spans": [ + { + "bbox": [ + 106, + 377, + 506, + 388 + ], + "score": 1.0, + "content": "[10] Ekin D Cubuk, Barret Zoph, Jonathon Shlens, and Quoc V Le. Randaugment: Practical automated data", + "type": "text" + } + ], + "index": 23, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 386, + 506, + 400 + ], + "spans": [ + { + "bbox": [ + 124, + 386, + 506, + 400 + ], + "score": 1.0, + "content": "augmentation with a reduced search space. In Proceedings of the IEEE/CVF Conference on Computer", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 126, + 397, + 365, + 409 + ], + "spans": [ + { + "bbox": [ + 126, + 397, + 365, + 409 + ], + "score": 1.0, + "content": "Vision and Pattern Recognition Workshops, pages 702–703, 2020.", + "type": "text" + } + ], + "index": 25, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 414, + 506, + 428 + ], + "spans": [ + { + "bbox": [ + 105, + 414, + 506, + 428 + ], + "score": 1.0, + "content": "[11] Zhigang Dai, Bolun Cai, Yugeng Lin, and Junying Chen. Up-detr: Unsupervised pre-training for object", + "type": "text" + } + ], + "index": 26, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 424, + 378, + 438 + ], + "spans": [ + { + "bbox": [ + 124, + 424, + 378, + 438 + ], + "score": 1.0, + "content": "detection with transformers. arXiv preprint arXiv:2011.09094, 2020.", + "type": "text" + } + ], + "index": 27, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 442, + 506, + 453 + ], + "spans": [ + { + "bbox": [ + 106, + 442, + 506, + 453 + ], + "score": 1.0, + "content": "[12] Stéphane d’Ascoli, Hugo Touvron, Matthew Leavitt, Ari Morcos, Giulio Biroli, and Levent Sagun. Convit:", + "type": "text" + } + ], + "index": 28, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 452, + 506, + 464 + ], + "spans": [ + { + "bbox": [ + 126, + 452, + 506, + 464 + ], + "score": 1.0, + "content": "Improving vision transformers with soft convolutional inductive biases. arXiv preprint arXiv:2103.10697,", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 125, + 461, + 152, + 474 + ], + "spans": [ + { + "bbox": [ + 125, + 461, + 152, + 474 + ], + "score": 1.0, + "content": "2021.", + "type": "text" + } + ], + "index": 30, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 480, + 505, + 491 + ], + "spans": [ + { + "bbox": [ + 106, + 480, + 505, + 491 + ], + "score": 1.0, + "content": "[13] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical", + "type": "text" + } + ], + "index": 31, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 488, + 507, + 502 + ], + "spans": [ + { + "bbox": [ + 125, + 488, + 507, + 502 + ], + "score": 1.0, + "content": "image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255.", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 125, + 499, + 169, + 511 + ], + "spans": [ + { + "bbox": [ + 125, + 499, + 169, + 511 + ], + "score": 1.0, + "content": "Ieee, 2009.", + "type": "text" + } + ], + "index": 33, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 516, + 507, + 530 + ], + "spans": [ + { + "bbox": [ + 105, + 516, + 507, + 530 + ], + "score": 1.0, + "content": "[14] Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirec-", + "type": "text" + } + ], + "index": 34, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 526, + 447, + 540 + ], + "spans": [ + { + "bbox": [ + 125, + 526, + 447, + 540 + ], + "score": 1.0, + "content": "tional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018.", + "type": "text" + } + ], + "index": 35, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 544, + 506, + 556 + ], + "spans": [ + { + "bbox": [ + 106, + 544, + 506, + 556 + ], + "score": 1.0, + "content": "[15] Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas", + "type": "text" + } + ], + "index": 36, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 554, + 506, + 567 + ], + "spans": [ + { + "bbox": [ + 125, + 554, + 506, + 567 + ], + "score": 1.0, + "content": "Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 125, + 564, + 486, + 577 + ], + "spans": [ + { + "bbox": [ + 125, + 564, + 486, + 577 + ], + "score": 1.0, + "content": "16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020.", + "type": "text" + } + ], + "index": 38, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 582, + 506, + 594 + ], + "spans": [ + { + "bbox": [ + 106, + 582, + 506, + 594 + ], + "score": 1.0, + "content": "[16] Kai Han, An Xiao, Enhua Wu, Jianyuan Guo, Chunjing Xu, and Yunhe Wang. Transformer in transformer.", + "type": "text" + } + ], + "index": 39, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 592, + 272, + 604 + ], + "spans": [ + { + "bbox": [ + 126, + 592, + 272, + 604 + ], + "score": 1.0, + "content": "arXiv preprint arXiv:2103.00112, 2021.", + "type": "text" + } + ], + "index": 40, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 609, + 504, + 621 + ], + "spans": [ + { + "bbox": [ + 106, + 609, + 504, + 621 + ], + "score": 1.0, + "content": "[17] Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. Mask r-cnn. In Proceedings of the IEEE", + "type": "text" + } + ], + "index": 41, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 619, + 381, + 632 + ], + "spans": [ + { + "bbox": [ + 125, + 619, + 381, + 632 + ], + "score": 1.0, + "content": "international conference on computer vision, pages 2961–2969, 2017.", + "type": "text" + } + ], + "index": 42, + "is_list_end_line": true + }, + { + "bbox": [ + 104, + 636, + 507, + 649 + ], + "spans": [ + { + "bbox": [ + 104, + 636, + 507, + 649 + ], + "score": 1.0, + "content": "[18] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition.", + "type": "text" + } + ], + "index": 43, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 646, + 506, + 660 + ], + "spans": [ + { + "bbox": [ + 124, + 646, + 506, + 660 + ], + "score": 1.0, + "content": "In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016.", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 104, + 663, + 506, + 677 + ], + "spans": [ + { + "bbox": [ + 104, + 663, + 506, + 677 + ], + "score": 1.0, + "content": "[19] Tong He, Zhi Zhang, Hang Zhang, Zhongyue Zhang, Junyuan Xie, and Mu Li. Bag of tricks for image", + "type": "text" + } + ], + "index": 45, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 674, + 506, + 687 + ], + "spans": [ + { + "bbox": [ + 125, + 674, + 506, + 687 + ], + "score": 1.0, + "content": "classification with convolutional neural networks. In Proceedings of the IEEE/CVF Conference on", + "type": "text" + } + ], + "index": 46 + }, + { + "bbox": [ + 125, + 684, + 362, + 698 + ], + "spans": [ + { + "bbox": [ + 125, + 684, + 362, + 698 + ], + "score": 1.0, + "content": "Computer Vision and Pattern Recognition, pages 558–567, 2019.", + "type": "text" + } + ], + "index": 47, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 700, + 507, + 714 + ], + "spans": [ + { + "bbox": [ + 105, + 700, + 507, + 714 + ], + "score": 1.0, + "content": "[20] Byeongho Heo, Sangdoo Yun, Dongyoon Han, Sanghyuk Chun, Junsuk Choe, and Seong Joon Oh.", + "type": "text" + } + ], + "index": 48, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 711, + 469, + 725 + ], + "spans": [ + { + "bbox": [ + 125, + 711, + 469, + 725 + ], + "score": 1.0, + "content": "Rethinking spatial dimensions of vision transformers. arXiv preprint arXiv:2103.16302, 2021.", + "type": "text" + } + ], + "index": 49, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 74, + 505, + 86 + ], + "spans": [ + { + "bbox": [ + 106, + 74, + 505, + 86 + ], + "score": 1.0, + "content": "[21] Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distilling the knowledge in a neural network. arXiv", + "type": "text", + "cross_page": true + } + ], + "index": 0, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 83, + 251, + 96 + ], + "spans": [ + { + "bbox": [ + 124, + 83, + 251, + 96 + ], + "score": 1.0, + "content": "preprint arXiv:1503.02531, 2015.", + "type": "text", + "cross_page": true + } + ], + "index": 1, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 101, + 506, + 113 + ], + "spans": [ + { + "bbox": [ + 106, + 101, + 506, + 113 + ], + "score": 1.0, + "content": "[22] Qibin Hou, Li Zhang, Ming-Ming Cheng, and Jiashi Feng. Strip pooling: Rethinking spatial pooling for", + "type": "text", + "cross_page": true + } + ], + "index": 2, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 111, + 506, + 123 + ], + "spans": [ + { + "bbox": [ + 125, + 111, + 506, + 123 + ], + "score": 1.0, + "content": "scene parsing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,", + "type": "text", + "cross_page": true + } + ], + "index": 3 + }, + { + "bbox": [ + 124, + 120, + 217, + 133 + ], + "spans": [ + { + "bbox": [ + 124, + 120, + 217, + 133 + ], + "score": 1.0, + "content": "pages 4003–4012, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 4, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 138, + 506, + 151 + ], + "spans": [ + { + "bbox": [ + 105, + 138, + 506, + 151 + ], + "score": 1.0, + "content": "[23] Gao Huang, Yu Sun, Zhuang Liu, Daniel Sedra, and Kilian Q Weinberger. Deep networks with stochastic", + "type": "text", + "cross_page": true + } + ], + "index": 5, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 148, + 429, + 160 + ], + "spans": [ + { + "bbox": [ + 126, + 148, + 429, + 160 + ], + "score": 1.0, + "content": "depth. In European conference on computer vision, pages 646–661. Springer, 2016.", + "type": "text", + "cross_page": true + } + ], + "index": 6, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 165, + 506, + 177 + ], + "spans": [ + { + "bbox": [ + 106, + 165, + 506, + 177 + ], + "score": 1.0, + "content": "[24] Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis,", + "type": "text", + "cross_page": true + } + ], + "index": 7, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 174, + 506, + 189 + ], + "spans": [ + { + "bbox": [ + 124, + 174, + 506, + 189 + ], + "score": 1.0, + "content": "Luke Zettlemoyer, and Veselin Stoyanov. Roberta: A robustly optimized bert pretraining approach. arXiv", + "type": "text", + "cross_page": true + } + ], + "index": 8 + }, + { + "bbox": [ + 124, + 185, + 251, + 198 + ], + "spans": [ + { + "bbox": [ + 124, + 185, + 251, + 198 + ], + "score": 1.0, + "content": "preprint arXiv:1907.11692, 2019.", + "type": "text", + "cross_page": true + } + ], + "index": 9, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 204, + 505, + 213 + ], + "spans": [ + { + "bbox": [ + 106, + 204, + 505, + 213 + ], + "score": 1.0, + "content": "[25] Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin", + "type": "text", + "cross_page": true + } + ], + "index": 10, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 213, + 506, + 225 + ], + "spans": [ + { + "bbox": [ + 126, + 213, + 506, + 225 + ], + "score": 1.0, + "content": "transformer: Hierarchical vision transformer using shifted windows. arXiv preprint arXiv:2103.14030,", + "type": "text", + "cross_page": true + } + ], + "index": 11 + }, + { + "bbox": [ + 125, + 222, + 151, + 235 + ], + "spans": [ + { + "bbox": [ + 125, + 222, + 151, + 235 + ], + "score": 1.0, + "content": "2021.", + "type": "text", + "cross_page": true + } + ], + "index": 12, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 240, + 506, + 252 + ], + "spans": [ + { + "bbox": [ + 106, + 240, + 506, + 252 + ], + "score": 1.0, + "content": "[26] Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmenta-", + "type": "text", + "cross_page": true + } + ], + "index": 13, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 250, + 506, + 262 + ], + "spans": [ + { + "bbox": [ + 125, + 250, + 506, + 262 + ], + "score": 1.0, + "content": "tion. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3431–3440,", + "type": "text", + "cross_page": true + } + ], + "index": 14 + }, + { + "bbox": [ + 124, + 258, + 152, + 273 + ], + "spans": [ + { + "bbox": [ + 124, + 258, + 152, + 273 + ], + "score": 1.0, + "content": "2015.", + "type": "text", + "cross_page": true + } + ], + "index": 15, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 277, + 506, + 289 + ], + "spans": [ + { + "bbox": [ + 106, + 277, + 506, + 289 + ], + "score": 1.0, + "content": "[27] Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint", + "type": "text", + "cross_page": true + } + ], + "index": 16, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 287, + 219, + 299 + ], + "spans": [ + { + "bbox": [ + 126, + 287, + 219, + 299 + ], + "score": 1.0, + "content": "arXiv:1711.05101, 2017.", + "type": "text", + "cross_page": true + } + ], + "index": 17, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 304, + 506, + 318 + ], + "spans": [ + { + "bbox": [ + 105, + 304, + 506, + 318 + ], + "score": 1.0, + "content": "[28] Niki Parmar, Ashish Vaswani, Jakob Uszkoreit, Łukasz Kaiser, Noam Shazeer, Alexander Ku, and Dustin", + "type": "text", + "cross_page": true + } + ], + "index": 18, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 315, + 366, + 327 + ], + "spans": [ + { + "bbox": [ + 126, + 315, + 366, + 327 + ], + "score": 1.0, + "content": "Tran. Image transformer. arXiv preprint arXiv:1802.05751, 2018.", + "type": "text", + "cross_page": true + } + ], + "index": 19, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 332, + 506, + 344 + ], + "spans": [ + { + "bbox": [ + 106, + 332, + 506, + 344 + ], + "score": 1.0, + "content": "[29] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen,", + "type": "text", + "cross_page": true + } + ], + "index": 20, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 340, + 506, + 355 + ], + "spans": [ + { + "bbox": [ + 124, + 340, + 506, + 355 + ], + "score": 1.0, + "content": "Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep", + "type": "text", + "cross_page": true + } + ], + "index": 21 + }, + { + "bbox": [ + 125, + 351, + 475, + 365 + ], + "spans": [ + { + "bbox": [ + 125, + 351, + 475, + 365 + ], + "score": 1.0, + "content": "learning library. In Advances in neural information processing systems, pages 8026–8037, 2019.", + "type": "text", + "cross_page": true + } + ], + "index": 22, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 367, + 506, + 382 + ], + "spans": [ + { + "bbox": [ + 105, + 367, + 506, + 382 + ], + "score": 1.0, + "content": "[30] Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, and Piotr Dollár. Designing network", + "type": "text", + "cross_page": true + } + ], + "index": 23, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 378, + 506, + 392 + ], + "spans": [ + { + "bbox": [ + 125, + 378, + 506, + 392 + ], + "score": 1.0, + "content": "design spaces. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,", + "type": "text", + "cross_page": true + } + ], + "index": 24 + }, + { + "bbox": [ + 124, + 389, + 227, + 400 + ], + "spans": [ + { + "bbox": [ + 124, + 389, + 227, + 400 + ], + "score": 1.0, + "content": "pages 10428–10436, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 25, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 406, + 506, + 418 + ], + "spans": [ + { + "bbox": [ + 106, + 406, + 506, + 418 + ], + "score": 1.0, + "content": "[31] Aravind Srinivas, Tsung-Yi Lin, Niki Parmar, Jonathon Shlens, Pieter Abbeel, and Ashish Vaswani.", + "type": "text", + "cross_page": true + } + ], + "index": 26, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 416, + 444, + 428 + ], + "spans": [ + { + "bbox": [ + 126, + 416, + 444, + 428 + ], + "score": 1.0, + "content": "Bottleneck transformers for visual recognition. arXiv preprint arXiv:2101.11605, 2021.", + "type": "text", + "cross_page": true + } + ], + "index": 27, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 433, + 507, + 447 + ], + "spans": [ + { + "bbox": [ + 105, + 433, + 507, + 447 + ], + "score": 1.0, + "content": "[32] Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout:", + "type": "text", + "cross_page": true + } + ], + "index": 28, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 443, + 506, + 455 + ], + "spans": [ + { + "bbox": [ + 125, + 443, + 506, + 455 + ], + "score": 1.0, + "content": "a simple way to prevent neural networks from overfitting. The journal of machine learning research,", + "type": "text", + "cross_page": true + } + ], + "index": 29 + }, + { + "bbox": [ + 126, + 453, + 216, + 465 + ], + "spans": [ + { + "bbox": [ + 126, + 453, + 216, + 465 + ], + "score": 1.0, + "content": "15(1):1929–1958, 2014.", + "type": "text", + "cross_page": true + } + ], + "index": 30, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 471, + 506, + 483 + ], + "spans": [ + { + "bbox": [ + 106, + 471, + 506, + 483 + ], + "score": 1.0, + "content": "[33] Zhiqing Sun, Shengcao Cao, Yiming Yang, and Kris Kitani. Rethinking transformer-based set prediction", + "type": "text", + "cross_page": true + } + ], + "index": 31, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 480, + 348, + 492 + ], + "spans": [ + { + "bbox": [ + 126, + 480, + 348, + 492 + ], + "score": 1.0, + "content": "for object detection. arXiv preprint arXiv:2011.10881, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 32, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 498, + 506, + 510 + ], + "spans": [ + { + "bbox": [ + 106, + 498, + 506, + 510 + ], + "score": 1.0, + "content": "[34] Mingxing Tan and Quoc V Le. Efficientnet: Rethinking model scaling for convolutional neural networks.", + "type": "text", + "cross_page": true + } + ], + "index": 33, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 507, + 273, + 520 + ], + "spans": [ + { + "bbox": [ + 124, + 507, + 273, + 520 + ], + "score": 1.0, + "content": "arXiv preprint arXiv:1905.11946, 2019.", + "type": "text", + "cross_page": true + } + ], + "index": 34, + "is_list_end_line": true + }, + { + "bbox": [ + 104, + 524, + 506, + 538 + ], + "spans": [ + { + "bbox": [ + 104, + 524, + 506, + 538 + ], + "score": 1.0, + "content": "[35] Ilya Tolstikhin, Neil Houlsby, Alexander Kolesnikov, Lucas Beyer, Xiaohua Zhai, Thomas Unterthiner,", + "type": "text", + "cross_page": true + } + ], + "index": 35, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 536, + 505, + 547 + ], + "spans": [ + { + "bbox": [ + 125, + 536, + 505, + 547 + ], + "score": 1.0, + "content": "Jessica Yung, Daniel Keysers, Jakob Uszkoreit, Mario Lucic, et al. Mlp-mixer: An all-mlp architecture for", + "type": "text", + "cross_page": true + } + ], + "index": 36 + }, + { + "bbox": [ + 124, + 544, + 300, + 558 + ], + "spans": [ + { + "bbox": [ + 124, + 544, + 300, + 558 + ], + "score": 1.0, + "content": "vision. arXiv preprint arXiv:2105.01601, 2021.", + "type": "text", + "cross_page": true + } + ], + "index": 37, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 561, + 506, + 574 + ], + "spans": [ + { + "bbox": [ + 105, + 561, + 506, + 574 + ], + "score": 1.0, + "content": "[36] Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, and Hervé", + "type": "text", + "cross_page": true + } + ], + "index": 38, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 573, + 505, + 585 + ], + "spans": [ + { + "bbox": [ + 125, + 573, + 505, + 585 + ], + "score": 1.0, + "content": "Jégou. Training data-efficient image transformers & distillation through attention. arXiv preprint", + "type": "text", + "cross_page": true + } + ], + "index": 39 + }, + { + "bbox": [ + 125, + 583, + 219, + 594 + ], + "spans": [ + { + "bbox": [ + 125, + 583, + 219, + 594 + ], + "score": 1.0, + "content": "arXiv:2012.12877, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 40, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 599, + 506, + 613 + ], + "spans": [ + { + "bbox": [ + 105, + 599, + 506, + 613 + ], + "score": 1.0, + "content": "[37] Hugo Touvron, Matthieu Cord, Alexandre Sablayrolles, Gabriel Synnaeve, and Hervé Jégou. Going deeper", + "type": "text", + "cross_page": true + } + ], + "index": 41, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 609, + 366, + 622 + ], + "spans": [ + { + "bbox": [ + 124, + 609, + 366, + 622 + ], + "score": 1.0, + "content": "with image transformers. arXiv preprint arXiv:2103.17239, 2021.", + "type": "text", + "cross_page": true + } + ], + "index": 42, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 626, + 506, + 640 + ], + "spans": [ + { + "bbox": [ + 105, + 626, + 506, + 640 + ], + "score": 1.0, + "content": "[38] Hugo Touvron, Andrea Vedaldi, Matthijs Douze, and Hervé Jégou. Fixing the train-test resolution", + "type": "text", + "cross_page": true + } + ], + "index": 43, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 636, + 320, + 649 + ], + "spans": [ + { + "bbox": [ + 124, + 636, + 320, + 649 + ], + "score": 1.0, + "content": "discrepancy. arXiv preprint arXiv:1906.06423, 2019.", + "type": "text", + "cross_page": true + } + ], + "index": 44, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 653, + 506, + 667 + ], + "spans": [ + { + "bbox": [ + 105, + 653, + 506, + 667 + ], + "score": 1.0, + "content": "[39] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz", + "type": "text", + "cross_page": true + } + ], + "index": 45, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 663, + 507, + 677 + ], + "spans": [ + { + "bbox": [ + 124, + 663, + 507, + 677 + ], + "score": 1.0, + "content": "Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems,", + "type": "text", + "cross_page": true + } + ], + "index": 46 + }, + { + "bbox": [ + 126, + 675, + 206, + 686 + ], + "spans": [ + { + "bbox": [ + 126, + 675, + 206, + 686 + ], + "score": 1.0, + "content": "30:5998–6008, 2017.", + "type": "text", + "cross_page": true + } + ], + "index": 47, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 691, + 506, + 705 + ], + "spans": [ + { + "bbox": [ + 105, + 691, + 506, + 705 + ], + "score": 1.0, + "content": "[40] Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and", + "type": "text", + "cross_page": true + } + ], + "index": 48, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 702, + 506, + 713 + ], + "spans": [ + { + "bbox": [ + 125, + 702, + 506, + 713 + ], + "score": 1.0, + "content": "Ling Shao. Pyramid vision transformer: A versatile backbone for dense prediction without convolutions.", + "type": "text", + "cross_page": true + } + ], + "index": 49 + }, + { + "bbox": [ + 124, + 711, + 273, + 724 + ], + "spans": [ + { + "bbox": [ + 124, + 711, + 273, + 724 + ], + "score": 1.0, + "content": "arXiv preprint arXiv:2102.12122, 2021.", + "type": "text", + "cross_page": true + } + ], + "index": 50, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 74, + 506, + 86 + ], + "spans": [ + { + "bbox": [ + 105, + 74, + 506, + 86 + ], + "score": 1.0, + "content": "[41] Yuqing Wang, Zhaoliang Xu, Xinlong Wang, Chunhua Shen, Baoshan Cheng, Hao Shen, and Huaxia Xia.", + "type": "text", + "cross_page": true + } + ], + "index": 0, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 83, + 488, + 95 + ], + "spans": [ + { + "bbox": [ + 126, + 83, + 488, + 95 + ], + "score": 1.0, + "content": "End-to-end video instance segmentation with transformers. arXiv preprint arXiv:2011.14503, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 1, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 98, + 506, + 113 + ], + "spans": [ + { + "bbox": [ + 105, + 99, + 199, + 111 + ], + "score": 1.0, + "content": "[42] Ross Wightman.", + "type": "text", + "cross_page": true + }, + { + "bbox": [ + 218, + 98, + 325, + 113 + ], + "score": 1.0, + "content": "Pytorch image models.", + "type": "text", + "cross_page": true + }, + { + "bbox": [ + 345, + 100, + 506, + 112 + ], + "score": 1.0, + "content": "https://github.com/rwightman/", + "type": "text", + "cross_page": true + } + ], + "index": 2, + "is_list_start_line": true + }, + { + "bbox": [ + 123, + 109, + 264, + 122 + ], + "spans": [ + { + "bbox": [ + 123, + 109, + 264, + 122 + ], + "score": 1.0, + "content": "pytorch-image-models, 2019.", + "type": "text", + "cross_page": true + } + ], + "index": 3, + "is_list_end_line": true + }, + { + "bbox": [ + 104, + 124, + 508, + 138 + ], + "spans": [ + { + "bbox": [ + 104, + 124, + 508, + 138 + ], + "score": 1.0, + "content": "[43] Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, and Lei Zhang. Cvt:", + "type": "text", + "cross_page": true + } + ], + "index": 4, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 136, + 450, + 147 + ], + "spans": [ + { + "bbox": [ + 126, + 136, + 450, + 147 + ], + "score": 1.0, + "content": "Introducing convolutions to vision transformers. arXiv preprint arXiv:2103.15808, 2021.", + "type": "text", + "cross_page": true + } + ], + "index": 5, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 150, + 506, + 165 + ], + "spans": [ + { + "bbox": [ + 105, + 150, + 506, + 165 + ], + "score": 1.0, + "content": "[44] Tete Xiao, Yingcheng Liu, Bolei Zhou, Yuning Jiang, and Jian Sun. Unified perceptual parsing for scene", + "type": "text", + "cross_page": true + } + ], + "index": 6, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 162, + 506, + 173 + ], + "spans": [ + { + "bbox": [ + 126, + 162, + 506, + 173 + ], + "score": 1.0, + "content": "understanding. In Proceedings of the European Conference on Computer Vision (ECCV), pages 418–434,", + "type": "text", + "cross_page": true + } + ], + "index": 7 + }, + { + "bbox": [ + 125, + 171, + 151, + 183 + ], + "spans": [ + { + "bbox": [ + 125, + 171, + 151, + 183 + ], + "score": 1.0, + "content": "2018.", + "type": "text", + "cross_page": true + } + ], + "index": 8, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 187, + 506, + 199 + ], + "spans": [ + { + "bbox": [ + 106, + 187, + 506, + 199 + ], + "score": 1.0, + "content": "[45] Fuzhi Yang, Huan Yang, Jianlong Fu, Hongtao Lu, and Baining Guo. Learning texture transformer network", + "type": "text", + "cross_page": true + } + ], + "index": 9, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 196, + 506, + 211 + ], + "spans": [ + { + "bbox": [ + 125, + 196, + 506, + 211 + ], + "score": 1.0, + "content": "for image super-resolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern", + "type": "text", + "cross_page": true + } + ], + "index": 10 + }, + { + "bbox": [ + 123, + 207, + 266, + 219 + ], + "spans": [ + { + "bbox": [ + 123, + 207, + 266, + 219 + ], + "score": 1.0, + "content": "Recognition, pages 5791–5800, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 11, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 221, + 506, + 237 + ], + "spans": [ + { + "bbox": [ + 105, + 221, + 506, + 237 + ], + "score": 1.0, + "content": "[46] Li Yuan, Yunpeng Chen, Tao Wang, Weihao Yu, Yujun Shi, Francis EH Tay, Jiashi Feng, and Shuicheng", + "type": "text", + "cross_page": true + } + ], + "index": 12, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 234, + 506, + 246 + ], + "spans": [ + { + "bbox": [ + 125, + 234, + 506, + 246 + ], + "score": 1.0, + "content": "Yan. Tokens-to-token vit: Training vision transformers from scratch on imagenet. arXiv preprint", + "type": "text", + "cross_page": true + } + ], + "index": 13 + }, + { + "bbox": [ + 124, + 243, + 219, + 254 + ], + "spans": [ + { + "bbox": [ + 124, + 243, + 219, + 254 + ], + "score": 1.0, + "content": "arXiv:2101.11986, 2021.", + "type": "text", + "cross_page": true + } + ], + "index": 14, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 259, + 506, + 271 + ], + "spans": [ + { + "bbox": [ + 106, + 259, + 506, + 271 + ], + "score": 1.0, + "content": "[47] Li Yuan, Francis EH Tay, Guilin Li, Tao Wang, and Jiashi Feng. Revisiting knowledge distillation via label", + "type": "text", + "cross_page": true + } + ], + "index": 15, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 268, + 506, + 282 + ], + "spans": [ + { + "bbox": [ + 124, + 268, + 506, + 282 + ], + "score": 1.0, + "content": "smoothing regularization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern", + "type": "text", + "cross_page": true + } + ], + "index": 16 + }, + { + "bbox": [ + 125, + 279, + 266, + 293 + ], + "spans": [ + { + "bbox": [ + 125, + 279, + 266, + 293 + ], + "score": 1.0, + "content": "Recognition, pages 3903–3911, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 17, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 295, + 506, + 307 + ], + "spans": [ + { + "bbox": [ + 106, + 295, + 506, + 307 + ], + "score": 1.0, + "content": "[48] Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, and Youngjoon Yoo.", + "type": "text", + "cross_page": true + } + ], + "index": 18, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 304, + 504, + 316 + ], + "spans": [ + { + "bbox": [ + 126, + 304, + 504, + 316 + ], + "score": 1.0, + "content": "Cutmix: Regularization strategy to train strong classifiers with localizable features. In Proceedings of the", + "type": "text", + "cross_page": true + } + ], + "index": 19 + }, + { + "bbox": [ + 125, + 314, + 427, + 328 + ], + "spans": [ + { + "bbox": [ + 125, + 314, + 427, + 328 + ], + "score": 1.0, + "content": "IEEE/CVF International Conference on Computer Vision, pages 6023–6032, 2019.", + "type": "text", + "cross_page": true + } + ], + "index": 20, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 330, + 507, + 344 + ], + "spans": [ + { + "bbox": [ + 105, + 330, + 507, + 344 + ], + "score": 1.0, + "content": "[49] Sangdoo Yun, Seong Joon Oh, Byeongho Heo, Dongyoon Han, Junsuk Choe, and Sanghyuk Chun.", + "type": "text", + "cross_page": true + } + ], + "index": 21, + "is_list_start_line": true + }, + { + "bbox": [ + 123, + 339, + 506, + 354 + ], + "spans": [ + { + "bbox": [ + 123, + 339, + 506, + 354 + ], + "score": 1.0, + "content": "Re-labeling imagenet: from single to multi-labels, from global to localized labels. arXiv preprint", + "type": "text", + "cross_page": true + } + ], + "index": 22 + }, + { + "bbox": [ + 124, + 351, + 219, + 362 + ], + "spans": [ + { + "bbox": [ + 124, + 351, + 219, + 362 + ], + "score": 1.0, + "content": "arXiv:2101.05022, 2021.", + "type": "text", + "cross_page": true + } + ], + "index": 23, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 367, + 506, + 379 + ], + "spans": [ + { + "bbox": [ + 106, + 367, + 506, + 379 + ], + "score": 1.0, + "content": "[50] Yanhong Zeng, Jianlong Fu, and Hongyang Chao. Learning joint spatial-temporal transformations for", + "type": "text", + "cross_page": true + } + ], + "index": 24, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 377, + 472, + 389 + ], + "spans": [ + { + "bbox": [ + 126, + 377, + 472, + 389 + ], + "score": 1.0, + "content": "video inpainting. In European Conference on Computer Vision, pages 528–543. Springer, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 25, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 393, + 505, + 405 + ], + "spans": [ + { + "bbox": [ + 105, + 393, + 505, + 405 + ], + "score": 1.0, + "content": "[51] Hang Zhang, Chongruo Wu, Zhongyue Zhang, Yi Zhu, Zhi Zhang, Haibin Lin, Yue Sun, Tong He, Jonas", + "type": "text", + "cross_page": true + } + ], + "index": 26, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 401, + 506, + 416 + ], + "spans": [ + { + "bbox": [ + 124, + 401, + 506, + 416 + ], + "score": 1.0, + "content": "Muller, R. Manmatha, Mu Li, and Alexander Smola. Resnest: Split-attention networks. arXiv preprint", + "type": "text", + "cross_page": true + } + ], + "index": 27 + }, + { + "bbox": [ + 125, + 412, + 220, + 424 + ], + "spans": [ + { + "bbox": [ + 125, + 412, + 220, + 424 + ], + "score": 1.0, + "content": "arXiv:2004.08955, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 28, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 429, + 506, + 441 + ], + "spans": [ + { + "bbox": [ + 106, + 429, + 506, + 441 + ], + "score": 1.0, + "content": "[52] Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and David Lopez-Paz. mixup: Beyond empirical risk", + "type": "text", + "cross_page": true + } + ], + "index": 29, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 438, + 326, + 450 + ], + "spans": [ + { + "bbox": [ + 125, + 438, + 326, + 450 + ], + "score": 1.0, + "content": "minimization. arXiv preprint arXiv:1710.09412, 2017.", + "type": "text", + "cross_page": true + } + ], + "index": 30, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 455, + 506, + 466 + ], + "spans": [ + { + "bbox": [ + 106, + 455, + 506, + 466 + ], + "score": 1.0, + "content": "[53] Hengshuang Zhao, Li Jiang, Jiaya Jia, Philip Torr, and Vladlen Koltun. Point transformer. arXiv preprint", + "type": "text", + "cross_page": true + } + ], + "index": 31, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 464, + 221, + 476 + ], + "spans": [ + { + "bbox": [ + 125, + 464, + 221, + 476 + ], + "score": 1.0, + "content": "arXiv:2012.09164, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 32, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 479, + 506, + 494 + ], + "spans": [ + { + "bbox": [ + 105, + 479, + 506, + 494 + ], + "score": 1.0, + "content": "[54] Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, and Jiaya Jia. Pyramid scene parsing", + "type": "text", + "cross_page": true + } + ], + "index": 33, + "is_list_start_line": true + }, + { + "bbox": [ + 124, + 489, + 506, + 504 + ], + "spans": [ + { + "bbox": [ + 124, + 489, + 506, + 504 + ], + "score": 1.0, + "content": "network. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages", + "type": "text", + "cross_page": true + } + ], + "index": 34 + }, + { + "bbox": [ + 126, + 500, + 196, + 512 + ], + "spans": [ + { + "bbox": [ + 126, + 500, + 196, + 512 + ], + "score": 1.0, + "content": "2881–2890, 2017.", + "type": "text", + "cross_page": true + } + ], + "index": 35, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 516, + 505, + 528 + ], + "spans": [ + { + "bbox": [ + 106, + 516, + 505, + 528 + ], + "score": 1.0, + "content": "[55] Minghang Zheng, Peng Gao, Xiaogang Wang, Hongsheng Li, and Hao Dong. End-to-end object detection", + "type": "text", + "cross_page": true + } + ], + "index": 36, + "is_list_start_line": true + }, + { + "bbox": [ + 123, + 525, + 408, + 539 + ], + "spans": [ + { + "bbox": [ + 123, + 525, + 408, + 539 + ], + "score": 1.0, + "content": "with adaptive clustering transformer. arXiv preprint arXiv:2011.09315, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 37, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 540, + 506, + 554 + ], + "spans": [ + { + "bbox": [ + 105, + 540, + 506, + 554 + ], + "score": 1.0, + "content": "[56] Sixiao Zheng, Jiachen Lu, Hengshuang Zhao, Xiatian Zhu, Zekun Luo, Yabiao Wang, Yanwei Fu, Jianfeng", + "type": "text", + "cross_page": true + } + ], + "index": 38, + "is_list_start_line": true + }, + { + "bbox": [ + 123, + 551, + 506, + 565 + ], + "spans": [ + { + "bbox": [ + 123, + 551, + 506, + 565 + ], + "score": 1.0, + "content": "Feng, Tao Xiang, Philip HS Torr, et al. Rethinking semantic segmentation from a sequence-to-sequence", + "type": "text", + "cross_page": true + } + ], + "index": 39 + }, + { + "bbox": [ + 123, + 562, + 386, + 574 + ], + "spans": [ + { + "bbox": [ + 123, + 562, + 386, + 574 + ], + "score": 1.0, + "content": "perspective with transformers. arXiv preprint arXiv:2012.15840, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 40, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 578, + 506, + 590 + ], + "spans": [ + { + "bbox": [ + 106, + 578, + 506, + 590 + ], + "score": 1.0, + "content": "[57] Zhun Zhong, Liang Zheng, Guoliang Kang, Shaozi Li, and Yi Yang. Random erasing data augmentation.", + "type": "text", + "cross_page": true + } + ], + "index": 41, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 588, + 504, + 600 + ], + "spans": [ + { + "bbox": [ + 126, + 588, + 504, + 600 + ], + "score": 1.0, + "content": "In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 13001–13008, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 42 + }, + { + "bbox": [ + 106, + 604, + 506, + 616 + ], + "spans": [ + { + "bbox": [ + 106, + 604, + 506, + 616 + ], + "score": 1.0, + "content": "[58] Bolei Zhou, Hang Zhao, Xavier Puig, Tete Xiao, Sanja Fidler, Adela Barriuso, and Antonio Torralba.", + "type": "text", + "cross_page": true + } + ], + "index": 43, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 614, + 506, + 626 + ], + "spans": [ + { + "bbox": [ + 125, + 614, + 506, + 626 + ], + "score": 1.0, + "content": "Semantic understanding of scenes through the ade20k dataset. International Journal of Computer Vision,", + "type": "text", + "cross_page": true + } + ], + "index": 44 + }, + { + "bbox": [ + 125, + 623, + 212, + 635 + ], + "spans": [ + { + "bbox": [ + 125, + 623, + 212, + 635 + ], + "score": 1.0, + "content": "127(3):302–321, 2019.", + "type": "text", + "cross_page": true + } + ], + "index": 45, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 640, + 506, + 652 + ], + "spans": [ + { + "bbox": [ + 106, + 640, + 506, + 652 + ], + "score": 1.0, + "content": "[59] Daquan Zhou, Bingyi Kang, Xiaojie Jin, Linjie Yang, Xiaochen Lian, Qibin Hou, and Jiashi Feng. Deepvit:", + "type": "text", + "cross_page": true + } + ], + "index": 46, + "is_list_start_line": true + }, + { + "bbox": [ + 125, + 649, + 403, + 663 + ], + "spans": [ + { + "bbox": [ + 125, + 649, + 403, + 663 + ], + "score": 1.0, + "content": "Towards deeper vision transformer. arXiv preprint arXiv:2103.11886, 2021.", + "type": "text", + "cross_page": true + } + ], + "index": 47, + "is_list_end_line": true + }, + { + "bbox": [ + 106, + 666, + 506, + 678 + ], + "spans": [ + { + "bbox": [ + 106, + 666, + 506, + 678 + ], + "score": 1.0, + "content": "[60] Luowei Zhou, Yingbo Zhou, Jason J Corso, Richard Socher, and Caiming Xiong. End-to-end dense video", + "type": "text", + "cross_page": true + } + ], + "index": 48, + "is_list_start_line": true + }, + { + "bbox": [ + 126, + 676, + 506, + 687 + ], + "spans": [ + { + "bbox": [ + 126, + 676, + 506, + 687 + ], + "score": 1.0, + "content": "captioning with masked transformer. In Proceedings of the IEEE Conference on Computer Vision and", + "type": "text", + "cross_page": true + } + ], + "index": 49 + }, + { + "bbox": [ + 124, + 684, + 293, + 699 + ], + "spans": [ + { + "bbox": [ + 124, + 684, + 293, + 699 + ], + "score": 1.0, + "content": "Pattern Recognition, pages 8739–8748, 2018.", + "type": "text", + "cross_page": true + } + ], + "index": 50, + "is_list_end_line": true + }, + { + "bbox": [ + 105, + 701, + 506, + 715 + ], + "spans": [ + { + "bbox": [ + 105, + 701, + 506, + 715 + ], + "score": 1.0, + "content": "[61] Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, and Jifeng Dai. Deformable detr: Deformable", + "type": "text", + "cross_page": true + } + ], + "index": 51, + "is_list_start_line": true + }, + { + "bbox": [ + 123, + 711, + 437, + 724 + ], + "spans": [ + { + "bbox": [ + 123, + 711, + 437, + 724 + ], + "score": 1.0, + "content": "transformers for end-to-end object detection. arXiv preprint arXiv:2010.04159, 2020.", + "type": "text", + "cross_page": true + } + ], + "index": 52, + "is_list_end_line": true + } + ], + "index": 24.5, + "bbox_fs": [ + 104, + 70, + 508, + 725 + ] + } + ] + }, + { + "preproc_blocks": [ + { + "type": "text", + "bbox": [ + 105, + 58, + 507, + 728 + ], + "lines": [ + { + "bbox": [ + 106, + 74, + 505, + 86 + ], + "spans": [ + { + "bbox": [ + 106, + 74, + 505, + 86 + ], + "score": 1.0, + "content": "[21] Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distilling the knowledge in a neural network. arXiv", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 124, + 83, + 251, + 96 + ], + "spans": [ + { + "bbox": [ + 124, + 83, + 251, + 96 + ], + "score": 1.0, + "content": "preprint arXiv:1503.02531, 2015.", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 106, + 101, + 506, + 113 + ], + "spans": [ + { + "bbox": [ + 106, + 101, + 506, + 113 + ], + "score": 1.0, + "content": "[22] Qibin Hou, Li Zhang, Ming-Ming Cheng, and Jiashi Feng. Strip pooling: Rethinking spatial pooling for", + "type": "text" + } + ], + "index": 2 + }, + { + "bbox": [ + 125, + 111, + 506, + 123 + ], + "spans": [ + { + "bbox": [ + 125, + 111, + 506, + 123 + ], + "score": 1.0, + "content": "scene parsing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 124, + 120, + 217, + 133 + ], + "spans": [ + { + "bbox": [ + 124, + 120, + 217, + 133 + ], + "score": 1.0, + "content": "pages 4003–4012, 2020.", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 105, + 138, + 506, + 151 + ], + "spans": [ + { + "bbox": [ + 105, + 138, + 506, + 151 + ], + "score": 1.0, + "content": "[23] Gao Huang, Yu Sun, Zhuang Liu, Daniel Sedra, and Kilian Q Weinberger. Deep networks with stochastic", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 126, + 148, + 429, + 160 + ], + "spans": [ + { + "bbox": [ + 126, + 148, + 429, + 160 + ], + "score": 1.0, + "content": "depth. In European conference on computer vision, pages 646–661. Springer, 2016.", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 106, + 165, + 506, + 177 + ], + "spans": [ + { + "bbox": [ + 106, + 165, + 506, + 177 + ], + "score": 1.0, + "content": "[24] Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis,", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 124, + 174, + 506, + 189 + ], + "spans": [ + { + "bbox": [ + 124, + 174, + 506, + 189 + ], + "score": 1.0, + "content": "Luke Zettlemoyer, and Veselin Stoyanov. Roberta: A robustly optimized bert pretraining approach. arXiv", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 124, + 185, + 251, + 198 + ], + "spans": [ + { + "bbox": [ + 124, + 185, + 251, + 198 + ], + "score": 1.0, + "content": "preprint arXiv:1907.11692, 2019.", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 106, + 204, + 505, + 213 + ], + "spans": [ + { + "bbox": [ + 106, + 204, + 505, + 213 + ], + "score": 1.0, + "content": "[25] Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 126, + 213, + 506, + 225 + ], + "spans": [ + { + "bbox": [ + 126, + 213, + 506, + 225 + ], + "score": 1.0, + "content": "transformer: Hierarchical vision transformer using shifted windows. arXiv preprint arXiv:2103.14030,", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 125, + 222, + 151, + 235 + ], + "spans": [ + { + "bbox": [ + 125, + 222, + 151, + 235 + ], + "score": 1.0, + "content": "2021.", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 106, + 240, + 506, + 252 + ], + "spans": [ + { + "bbox": [ + 106, + 240, + 506, + 252 + ], + "score": 1.0, + "content": "[26] Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmenta-", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 125, + 250, + 506, + 262 + ], + "spans": [ + { + "bbox": [ + 125, + 250, + 506, + 262 + ], + "score": 1.0, + "content": "tion. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3431–3440,", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 124, + 258, + 152, + 273 + ], + "spans": [ + { + "bbox": [ + 124, + 258, + 152, + 273 + ], + "score": 1.0, + "content": "2015.", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 106, + 277, + 506, + 289 + ], + "spans": [ + { + "bbox": [ + 106, + 277, + 506, + 289 + ], + "score": 1.0, + "content": "[27] Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 126, + 287, + 219, + 299 + ], + "spans": [ + { + "bbox": [ + 126, + 287, + 219, + 299 + ], + "score": 1.0, + "content": "arXiv:1711.05101, 2017.", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 105, + 304, + 506, + 318 + ], + "spans": [ + { + "bbox": [ + 105, + 304, + 506, + 318 + ], + "score": 1.0, + "content": "[28] Niki Parmar, Ashish Vaswani, Jakob Uszkoreit, Łukasz Kaiser, Noam Shazeer, Alexander Ku, and Dustin", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 126, + 315, + 366, + 327 + ], + "spans": [ + { + "bbox": [ + 126, + 315, + 366, + 327 + ], + "score": 1.0, + "content": "Tran. Image transformer. arXiv preprint arXiv:1802.05751, 2018.", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 106, + 332, + 506, + 344 + ], + "spans": [ + { + "bbox": [ + 106, + 332, + 506, + 344 + ], + "score": 1.0, + "content": "[29] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen,", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 124, + 340, + 506, + 355 + ], + "spans": [ + { + "bbox": [ + 124, + 340, + 506, + 355 + ], + "score": 1.0, + "content": "Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 125, + 351, + 475, + 365 + ], + "spans": [ + { + "bbox": [ + 125, + 351, + 475, + 365 + ], + "score": 1.0, + "content": "learning library. In Advances in neural information processing systems, pages 8026–8037, 2019.", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 105, + 367, + 506, + 382 + ], + "spans": [ + { + "bbox": [ + 105, + 367, + 506, + 382 + ], + "score": 1.0, + "content": "[30] Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, and Piotr Dollár. Designing network", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 125, + 378, + 506, + 392 + ], + "spans": [ + { + "bbox": [ + 125, + 378, + 506, + 392 + ], + "score": 1.0, + "content": "design spaces. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 124, + 389, + 227, + 400 + ], + "spans": [ + { + "bbox": [ + 124, + 389, + 227, + 400 + ], + "score": 1.0, + "content": "pages 10428–10436, 2020.", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 106, + 406, + 506, + 418 + ], + "spans": [ + { + "bbox": [ + 106, + 406, + 506, + 418 + ], + "score": 1.0, + "content": "[31] Aravind Srinivas, Tsung-Yi Lin, Niki Parmar, Jonathon Shlens, Pieter Abbeel, and Ashish Vaswani.", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 126, + 416, + 444, + 428 + ], + "spans": [ + { + "bbox": [ + 126, + 416, + 444, + 428 + ], + "score": 1.0, + "content": "Bottleneck transformers for visual recognition. arXiv preprint arXiv:2101.11605, 2021.", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 105, + 433, + 507, + 447 + ], + "spans": [ + { + "bbox": [ + 105, + 433, + 507, + 447 + ], + "score": 1.0, + "content": "[32] Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout:", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 125, + 443, + 506, + 455 + ], + "spans": [ + { + "bbox": [ + 125, + 443, + 506, + 455 + ], + "score": 1.0, + "content": "a simple way to prevent neural networks from overfitting. The journal of machine learning research,", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 126, + 453, + 216, + 465 + ], + "spans": [ + { + "bbox": [ + 126, + 453, + 216, + 465 + ], + "score": 1.0, + "content": "15(1):1929–1958, 2014.", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 106, + 471, + 506, + 483 + ], + "spans": [ + { + "bbox": [ + 106, + 471, + 506, + 483 + ], + "score": 1.0, + "content": "[33] Zhiqing Sun, Shengcao Cao, Yiming Yang, and Kris Kitani. Rethinking transformer-based set prediction", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 126, + 480, + 348, + 492 + ], + "spans": [ + { + "bbox": [ + 126, + 480, + 348, + 492 + ], + "score": 1.0, + "content": "for object detection. arXiv preprint arXiv:2011.10881, 2020.", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 106, + 498, + 506, + 510 + ], + "spans": [ + { + "bbox": [ + 106, + 498, + 506, + 510 + ], + "score": 1.0, + "content": "[34] Mingxing Tan and Quoc V Le. Efficientnet: Rethinking model scaling for convolutional neural networks.", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 124, + 507, + 273, + 520 + ], + "spans": [ + { + "bbox": [ + 124, + 507, + 273, + 520 + ], + "score": 1.0, + "content": "arXiv preprint arXiv:1905.11946, 2019.", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 104, + 524, + 506, + 538 + ], + "spans": [ + { + "bbox": [ + 104, + 524, + 506, + 538 + ], + "score": 1.0, + "content": "[35] Ilya Tolstikhin, Neil Houlsby, Alexander Kolesnikov, Lucas Beyer, Xiaohua Zhai, Thomas Unterthiner,", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 125, + 536, + 505, + 547 + ], + "spans": [ + { + "bbox": [ + 125, + 536, + 505, + 547 + ], + "score": 1.0, + "content": "Jessica Yung, Daniel Keysers, Jakob Uszkoreit, Mario Lucic, et al. Mlp-mixer: An all-mlp architecture for", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 124, + 544, + 300, + 558 + ], + "spans": [ + { + "bbox": [ + 124, + 544, + 300, + 558 + ], + "score": 1.0, + "content": "vision. arXiv preprint arXiv:2105.01601, 2021.", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 105, + 561, + 506, + 574 + ], + "spans": [ + { + "bbox": [ + 105, + 561, + 506, + 574 + ], + "score": 1.0, + "content": "[36] Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, and Hervé", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 125, + 573, + 505, + 585 + ], + "spans": [ + { + "bbox": [ + 125, + 573, + 505, + 585 + ], + "score": 1.0, + "content": "Jégou. Training data-efficient image transformers & distillation through attention. arXiv preprint", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 125, + 583, + 219, + 594 + ], + "spans": [ + { + "bbox": [ + 125, + 583, + 219, + 594 + ], + "score": 1.0, + "content": "arXiv:2012.12877, 2020.", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 105, + 599, + 506, + 613 + ], + "spans": [ + { + "bbox": [ + 105, + 599, + 506, + 613 + ], + "score": 1.0, + "content": "[37] Hugo Touvron, Matthieu Cord, Alexandre Sablayrolles, Gabriel Synnaeve, and Hervé Jégou. Going deeper", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 124, + 609, + 366, + 622 + ], + "spans": [ + { + "bbox": [ + 124, + 609, + 366, + 622 + ], + "score": 1.0, + "content": "with image transformers. arXiv preprint arXiv:2103.17239, 2021.", + "type": "text" + } + ], + "index": 42 + }, + { + "bbox": [ + 105, + 626, + 506, + 640 + ], + "spans": [ + { + "bbox": [ + 105, + 626, + 506, + 640 + ], + "score": 1.0, + "content": "[38] Hugo Touvron, Andrea Vedaldi, Matthijs Douze, and Hervé Jégou. Fixing the train-test resolution", + "type": "text" + } + ], + "index": 43 + }, + { + "bbox": [ + 124, + 636, + 320, + 649 + ], + "spans": [ + { + "bbox": [ + 124, + 636, + 320, + 649 + ], + "score": 1.0, + "content": "discrepancy. arXiv preprint arXiv:1906.06423, 2019.", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 105, + 653, + 506, + 667 + ], + "spans": [ + { + "bbox": [ + 105, + 653, + 506, + 667 + ], + "score": 1.0, + "content": "[39] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz", + "type": "text" + } + ], + "index": 45 + }, + { + "bbox": [ + 124, + 663, + 507, + 677 + ], + "spans": [ + { + "bbox": [ + 124, + 663, + 507, + 677 + ], + "score": 1.0, + "content": "Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems,", + "type": "text" + } + ], + "index": 46 + }, + { + "bbox": [ + 126, + 675, + 206, + 686 + ], + "spans": [ + { + "bbox": [ + 126, + 675, + 206, + 686 + ], + "score": 1.0, + "content": "30:5998–6008, 2017.", + "type": "text" + } + ], + "index": 47 + }, + { + "bbox": [ + 105, + 691, + 506, + 705 + ], + "spans": [ + { + "bbox": [ + 105, + 691, + 506, + 705 + ], + "score": 1.0, + "content": "[40] Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and", + "type": "text" + } + ], + "index": 48 + }, + { + "bbox": [ + 125, + 702, + 506, + 713 + ], + "spans": [ + { + "bbox": [ + 125, + 702, + 506, + 713 + ], + "score": 1.0, + "content": "Ling Shao. Pyramid vision transformer: A versatile backbone for dense prediction without convolutions.", + "type": "text" + } + ], + "index": 49 + }, + { + "bbox": [ + 124, + 711, + 273, + 724 + ], + "spans": [ + { + "bbox": [ + 124, + 711, + 273, + 724 + ], + "score": 1.0, + "content": "arXiv preprint arXiv:2102.12122, 2021.", + "type": "text" + } + ], + "index": 50 + } + ], + "index": 25 + } + ], + "page_idx": 11, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 300, + 741, + 311, + 750 + ], + "lines": [ + { + "bbox": [ + 299, + 740, + 313, + 754 + ], + "spans": [ + { + "bbox": [ + 299, + 740, + 313, + 754 + ], + "score": 1.0, + "content": "12", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "list", + "bbox": [ + 105, + 58, + 507, + 728 + ], + "lines": [], + "index": 25, + "bbox_fs": [ + 104, + 74, + 507, + 724 + ], + "lines_deleted": true + } + ] + }, + { + "preproc_blocks": [ + { + "type": "text", + "bbox": [ + 105, + 52, + 507, + 730 + ], + "lines": [ + { + "bbox": [ + 105, + 74, + 506, + 86 + ], + "spans": [ + { + "bbox": [ + 105, + 74, + 506, + 86 + ], + "score": 1.0, + "content": "[41] Yuqing Wang, Zhaoliang Xu, Xinlong Wang, Chunhua Shen, Baoshan Cheng, Hao Shen, and Huaxia Xia.", + "type": "text" + } + ], + "index": 0 + }, + { + "bbox": [ + 126, + 83, + 488, + 95 + ], + "spans": [ + { + "bbox": [ + 126, + 83, + 488, + 95 + ], + "score": 1.0, + "content": "End-to-end video instance segmentation with transformers. arXiv preprint arXiv:2011.14503, 2020.", + "type": "text" + } + ], + "index": 1 + }, + { + "bbox": [ + 105, + 98, + 506, + 113 + ], + "spans": [ + { + "bbox": [ + 105, + 99, + 199, + 111 + ], + "score": 1.0, + "content": "[42] Ross Wightman.", + "type": "text" + }, + { + "bbox": [ + 218, + 98, + 325, + 113 + ], + "score": 1.0, + "content": "Pytorch image models.", + "type": "text" + }, + { + "bbox": [ + 345, + 100, + 506, + 112 + ], + "score": 1.0, + "content": "https://github.com/rwightman/", + "type": "text" + } + ], + "index": 2 + }, + { + "bbox": [ + 123, + 109, + 264, + 122 + ], + "spans": [ + { + "bbox": [ + 123, + 109, + 264, + 122 + ], + "score": 1.0, + "content": "pytorch-image-models, 2019.", + "type": "text" + } + ], + "index": 3 + }, + { + "bbox": [ + 104, + 124, + 508, + 138 + ], + "spans": [ + { + "bbox": [ + 104, + 124, + 508, + 138 + ], + "score": 1.0, + "content": "[43] Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, and Lei Zhang. Cvt:", + "type": "text" + } + ], + "index": 4 + }, + { + "bbox": [ + 126, + 136, + 450, + 147 + ], + "spans": [ + { + "bbox": [ + 126, + 136, + 450, + 147 + ], + "score": 1.0, + "content": "Introducing convolutions to vision transformers. arXiv preprint arXiv:2103.15808, 2021.", + "type": "text" + } + ], + "index": 5 + }, + { + "bbox": [ + 105, + 150, + 506, + 165 + ], + "spans": [ + { + "bbox": [ + 105, + 150, + 506, + 165 + ], + "score": 1.0, + "content": "[44] Tete Xiao, Yingcheng Liu, Bolei Zhou, Yuning Jiang, and Jian Sun. Unified perceptual parsing for scene", + "type": "text" + } + ], + "index": 6 + }, + { + "bbox": [ + 126, + 162, + 506, + 173 + ], + "spans": [ + { + "bbox": [ + 126, + 162, + 506, + 173 + ], + "score": 1.0, + "content": "understanding. In Proceedings of the European Conference on Computer Vision (ECCV), pages 418–434,", + "type": "text" + } + ], + "index": 7 + }, + { + "bbox": [ + 125, + 171, + 151, + 183 + ], + "spans": [ + { + "bbox": [ + 125, + 171, + 151, + 183 + ], + "score": 1.0, + "content": "2018.", + "type": "text" + } + ], + "index": 8 + }, + { + "bbox": [ + 106, + 187, + 506, + 199 + ], + "spans": [ + { + "bbox": [ + 106, + 187, + 506, + 199 + ], + "score": 1.0, + "content": "[45] Fuzhi Yang, Huan Yang, Jianlong Fu, Hongtao Lu, and Baining Guo. Learning texture transformer network", + "type": "text" + } + ], + "index": 9 + }, + { + "bbox": [ + 125, + 196, + 506, + 211 + ], + "spans": [ + { + "bbox": [ + 125, + 196, + 506, + 211 + ], + "score": 1.0, + "content": "for image super-resolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern", + "type": "text" + } + ], + "index": 10 + }, + { + "bbox": [ + 123, + 207, + 266, + 219 + ], + "spans": [ + { + "bbox": [ + 123, + 207, + 266, + 219 + ], + "score": 1.0, + "content": "Recognition, pages 5791–5800, 2020.", + "type": "text" + } + ], + "index": 11 + }, + { + "bbox": [ + 105, + 221, + 506, + 237 + ], + "spans": [ + { + "bbox": [ + 105, + 221, + 506, + 237 + ], + "score": 1.0, + "content": "[46] Li Yuan, Yunpeng Chen, Tao Wang, Weihao Yu, Yujun Shi, Francis EH Tay, Jiashi Feng, and Shuicheng", + "type": "text" + } + ], + "index": 12 + }, + { + "bbox": [ + 125, + 234, + 506, + 246 + ], + "spans": [ + { + "bbox": [ + 125, + 234, + 506, + 246 + ], + "score": 1.0, + "content": "Yan. Tokens-to-token vit: Training vision transformers from scratch on imagenet. arXiv preprint", + "type": "text" + } + ], + "index": 13 + }, + { + "bbox": [ + 124, + 243, + 219, + 254 + ], + "spans": [ + { + "bbox": [ + 124, + 243, + 219, + 254 + ], + "score": 1.0, + "content": "arXiv:2101.11986, 2021.", + "type": "text" + } + ], + "index": 14 + }, + { + "bbox": [ + 106, + 259, + 506, + 271 + ], + "spans": [ + { + "bbox": [ + 106, + 259, + 506, + 271 + ], + "score": 1.0, + "content": "[47] Li Yuan, Francis EH Tay, Guilin Li, Tao Wang, and Jiashi Feng. Revisiting knowledge distillation via label", + "type": "text" + } + ], + "index": 15 + }, + { + "bbox": [ + 124, + 268, + 506, + 282 + ], + "spans": [ + { + "bbox": [ + 124, + 268, + 506, + 282 + ], + "score": 1.0, + "content": "smoothing regularization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern", + "type": "text" + } + ], + "index": 16 + }, + { + "bbox": [ + 125, + 279, + 266, + 293 + ], + "spans": [ + { + "bbox": [ + 125, + 279, + 266, + 293 + ], + "score": 1.0, + "content": "Recognition, pages 3903–3911, 2020.", + "type": "text" + } + ], + "index": 17 + }, + { + "bbox": [ + 106, + 295, + 506, + 307 + ], + "spans": [ + { + "bbox": [ + 106, + 295, + 506, + 307 + ], + "score": 1.0, + "content": "[48] Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, and Youngjoon Yoo.", + "type": "text" + } + ], + "index": 18 + }, + { + "bbox": [ + 126, + 304, + 504, + 316 + ], + "spans": [ + { + "bbox": [ + 126, + 304, + 504, + 316 + ], + "score": 1.0, + "content": "Cutmix: Regularization strategy to train strong classifiers with localizable features. In Proceedings of the", + "type": "text" + } + ], + "index": 19 + }, + { + "bbox": [ + 125, + 314, + 427, + 328 + ], + "spans": [ + { + "bbox": [ + 125, + 314, + 427, + 328 + ], + "score": 1.0, + "content": "IEEE/CVF International Conference on Computer Vision, pages 6023–6032, 2019.", + "type": "text" + } + ], + "index": 20 + }, + { + "bbox": [ + 105, + 330, + 507, + 344 + ], + "spans": [ + { + "bbox": [ + 105, + 330, + 507, + 344 + ], + "score": 1.0, + "content": "[49] Sangdoo Yun, Seong Joon Oh, Byeongho Heo, Dongyoon Han, Junsuk Choe, and Sanghyuk Chun.", + "type": "text" + } + ], + "index": 21 + }, + { + "bbox": [ + 123, + 339, + 506, + 354 + ], + "spans": [ + { + "bbox": [ + 123, + 339, + 506, + 354 + ], + "score": 1.0, + "content": "Re-labeling imagenet: from single to multi-labels, from global to localized labels. arXiv preprint", + "type": "text" + } + ], + "index": 22 + }, + { + "bbox": [ + 124, + 351, + 219, + 362 + ], + "spans": [ + { + "bbox": [ + 124, + 351, + 219, + 362 + ], + "score": 1.0, + "content": "arXiv:2101.05022, 2021.", + "type": "text" + } + ], + "index": 23 + }, + { + "bbox": [ + 106, + 367, + 506, + 379 + ], + "spans": [ + { + "bbox": [ + 106, + 367, + 506, + 379 + ], + "score": 1.0, + "content": "[50] Yanhong Zeng, Jianlong Fu, and Hongyang Chao. Learning joint spatial-temporal transformations for", + "type": "text" + } + ], + "index": 24 + }, + { + "bbox": [ + 126, + 377, + 472, + 389 + ], + "spans": [ + { + "bbox": [ + 126, + 377, + 472, + 389 + ], + "score": 1.0, + "content": "video inpainting. In European Conference on Computer Vision, pages 528–543. Springer, 2020.", + "type": "text" + } + ], + "index": 25 + }, + { + "bbox": [ + 105, + 393, + 505, + 405 + ], + "spans": [ + { + "bbox": [ + 105, + 393, + 505, + 405 + ], + "score": 1.0, + "content": "[51] Hang Zhang, Chongruo Wu, Zhongyue Zhang, Yi Zhu, Zhi Zhang, Haibin Lin, Yue Sun, Tong He, Jonas", + "type": "text" + } + ], + "index": 26 + }, + { + "bbox": [ + 124, + 401, + 506, + 416 + ], + "spans": [ + { + "bbox": [ + 124, + 401, + 506, + 416 + ], + "score": 1.0, + "content": "Muller, R. Manmatha, Mu Li, and Alexander Smola. Resnest: Split-attention networks. arXiv preprint", + "type": "text" + } + ], + "index": 27 + }, + { + "bbox": [ + 125, + 412, + 220, + 424 + ], + "spans": [ + { + "bbox": [ + 125, + 412, + 220, + 424 + ], + "score": 1.0, + "content": "arXiv:2004.08955, 2020.", + "type": "text" + } + ], + "index": 28 + }, + { + "bbox": [ + 106, + 429, + 506, + 441 + ], + "spans": [ + { + "bbox": [ + 106, + 429, + 506, + 441 + ], + "score": 1.0, + "content": "[52] Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and David Lopez-Paz. mixup: Beyond empirical risk", + "type": "text" + } + ], + "index": 29 + }, + { + "bbox": [ + 125, + 438, + 326, + 450 + ], + "spans": [ + { + "bbox": [ + 125, + 438, + 326, + 450 + ], + "score": 1.0, + "content": "minimization. arXiv preprint arXiv:1710.09412, 2017.", + "type": "text" + } + ], + "index": 30 + }, + { + "bbox": [ + 106, + 455, + 506, + 466 + ], + "spans": [ + { + "bbox": [ + 106, + 455, + 506, + 466 + ], + "score": 1.0, + "content": "[53] Hengshuang Zhao, Li Jiang, Jiaya Jia, Philip Torr, and Vladlen Koltun. Point transformer. arXiv preprint", + "type": "text" + } + ], + "index": 31 + }, + { + "bbox": [ + 125, + 464, + 221, + 476 + ], + "spans": [ + { + "bbox": [ + 125, + 464, + 221, + 476 + ], + "score": 1.0, + "content": "arXiv:2012.09164, 2020.", + "type": "text" + } + ], + "index": 32 + }, + { + "bbox": [ + 105, + 479, + 506, + 494 + ], + "spans": [ + { + "bbox": [ + 105, + 479, + 506, + 494 + ], + "score": 1.0, + "content": "[54] Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, and Jiaya Jia. Pyramid scene parsing", + "type": "text" + } + ], + "index": 33 + }, + { + "bbox": [ + 124, + 489, + 506, + 504 + ], + "spans": [ + { + "bbox": [ + 124, + 489, + 506, + 504 + ], + "score": 1.0, + "content": "network. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages", + "type": "text" + } + ], + "index": 34 + }, + { + "bbox": [ + 126, + 500, + 196, + 512 + ], + "spans": [ + { + "bbox": [ + 126, + 500, + 196, + 512 + ], + "score": 1.0, + "content": "2881–2890, 2017.", + "type": "text" + } + ], + "index": 35 + }, + { + "bbox": [ + 106, + 516, + 505, + 528 + ], + "spans": [ + { + "bbox": [ + 106, + 516, + 505, + 528 + ], + "score": 1.0, + "content": "[55] Minghang Zheng, Peng Gao, Xiaogang Wang, Hongsheng Li, and Hao Dong. End-to-end object detection", + "type": "text" + } + ], + "index": 36 + }, + { + "bbox": [ + 123, + 525, + 408, + 539 + ], + "spans": [ + { + "bbox": [ + 123, + 525, + 408, + 539 + ], + "score": 1.0, + "content": "with adaptive clustering transformer. arXiv preprint arXiv:2011.09315, 2020.", + "type": "text" + } + ], + "index": 37 + }, + { + "bbox": [ + 105, + 540, + 506, + 554 + ], + "spans": [ + { + "bbox": [ + 105, + 540, + 506, + 554 + ], + "score": 1.0, + "content": "[56] Sixiao Zheng, Jiachen Lu, Hengshuang Zhao, Xiatian Zhu, Zekun Luo, Yabiao Wang, Yanwei Fu, Jianfeng", + "type": "text" + } + ], + "index": 38 + }, + { + "bbox": [ + 123, + 551, + 506, + 565 + ], + "spans": [ + { + "bbox": [ + 123, + 551, + 506, + 565 + ], + "score": 1.0, + "content": "Feng, Tao Xiang, Philip HS Torr, et al. Rethinking semantic segmentation from a sequence-to-sequence", + "type": "text" + } + ], + "index": 39 + }, + { + "bbox": [ + 123, + 562, + 386, + 574 + ], + "spans": [ + { + "bbox": [ + 123, + 562, + 386, + 574 + ], + "score": 1.0, + "content": "perspective with transformers. arXiv preprint arXiv:2012.15840, 2020.", + "type": "text" + } + ], + "index": 40 + }, + { + "bbox": [ + 106, + 578, + 506, + 590 + ], + "spans": [ + { + "bbox": [ + 106, + 578, + 506, + 590 + ], + "score": 1.0, + "content": "[57] Zhun Zhong, Liang Zheng, Guoliang Kang, Shaozi Li, and Yi Yang. Random erasing data augmentation.", + "type": "text" + } + ], + "index": 41 + }, + { + "bbox": [ + 126, + 588, + 504, + 600 + ], + "spans": [ + { + "bbox": [ + 126, + 588, + 504, + 600 + ], + "score": 1.0, + "content": "In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 13001–13008, 2020.", + "type": "text" + } + ], + "index": 42 + }, + { + "bbox": [ + 106, + 604, + 506, + 616 + ], + "spans": [ + { + "bbox": [ + 106, + 604, + 506, + 616 + ], + "score": 1.0, + "content": "[58] Bolei Zhou, Hang Zhao, Xavier Puig, Tete Xiao, Sanja Fidler, Adela Barriuso, and Antonio Torralba.", + "type": "text" + } + ], + "index": 43 + }, + { + "bbox": [ + 125, + 614, + 506, + 626 + ], + "spans": [ + { + "bbox": [ + 125, + 614, + 506, + 626 + ], + "score": 1.0, + "content": "Semantic understanding of scenes through the ade20k dataset. International Journal of Computer Vision,", + "type": "text" + } + ], + "index": 44 + }, + { + "bbox": [ + 125, + 623, + 212, + 635 + ], + "spans": [ + { + "bbox": [ + 125, + 623, + 212, + 635 + ], + "score": 1.0, + "content": "127(3):302–321, 2019.", + "type": "text" + } + ], + "index": 45 + }, + { + "bbox": [ + 106, + 640, + 506, + 652 + ], + "spans": [ + { + "bbox": [ + 106, + 640, + 506, + 652 + ], + "score": 1.0, + "content": "[59] Daquan Zhou, Bingyi Kang, Xiaojie Jin, Linjie Yang, Xiaochen Lian, Qibin Hou, and Jiashi Feng. Deepvit:", + "type": "text" + } + ], + "index": 46 + }, + { + "bbox": [ + 125, + 649, + 403, + 663 + ], + "spans": [ + { + "bbox": [ + 125, + 649, + 403, + 663 + ], + "score": 1.0, + "content": "Towards deeper vision transformer. arXiv preprint arXiv:2103.11886, 2021.", + "type": "text" + } + ], + "index": 47 + }, + { + "bbox": [ + 106, + 666, + 506, + 678 + ], + "spans": [ + { + "bbox": [ + 106, + 666, + 506, + 678 + ], + "score": 1.0, + "content": "[60] Luowei Zhou, Yingbo Zhou, Jason J Corso, Richard Socher, and Caiming Xiong. End-to-end dense video", + "type": "text" + } + ], + "index": 48 + }, + { + "bbox": [ + 126, + 676, + 506, + 687 + ], + "spans": [ + { + "bbox": [ + 126, + 676, + 506, + 687 + ], + "score": 1.0, + "content": "captioning with masked transformer. In Proceedings of the IEEE Conference on Computer Vision and", + "type": "text" + } + ], + "index": 49 + }, + { + "bbox": [ + 124, + 684, + 293, + 699 + ], + "spans": [ + { + "bbox": [ + 124, + 684, + 293, + 699 + ], + "score": 1.0, + "content": "Pattern Recognition, pages 8739–8748, 2018.", + "type": "text" + } + ], + "index": 50 + }, + { + "bbox": [ + 105, + 701, + 506, + 715 + ], + "spans": [ + { + "bbox": [ + 105, + 701, + 506, + 715 + ], + "score": 1.0, + "content": "[61] Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, and Jifeng Dai. Deformable detr: Deformable", + "type": "text" + } + ], + "index": 51 + }, + { + "bbox": [ + 123, + 711, + 437, + 724 + ], + "spans": [ + { + "bbox": [ + 123, + 711, + 437, + 724 + ], + "score": 1.0, + "content": "transformers for end-to-end object detection. arXiv preprint arXiv:2010.04159, 2020.", + "type": "text" + } + ], + "index": 52 + } + ], + "index": 26 + } + ], + "page_idx": 12, + "page_size": [ + 612, + 792 + ], + "discarded_blocks": [ + { + "type": "discarded", + "bbox": [ + 300, + 741, + 311, + 750 + ], + "lines": [ + { + "bbox": [ + 299, + 740, + 312, + 754 + ], + "spans": [ + { + "bbox": [ + 299, + 740, + 312, + 754 + ], + "score": 1.0, + "content": "13", + "type": "text" + } + ] + } + ] + } + ], + "para_blocks": [ + { + "type": "list", + "bbox": [ + 105, + 52, + 507, + 730 + ], + "lines": [], + "index": 26, + "bbox_fs": [ + 104, + 74, + 508, + 724 + ], + "lines_deleted": true + } + ] + } + ], + "_backend": "pipeline", + "_version_name": "2.2.2" +} \ No newline at end of file diff --git a/parse/train/2vubO341F_E/2vubO341F_E_model.json b/parse/train/2vubO341F_E/2vubO341F_E_model.json new file mode 100644 index 0000000000000000000000000000000000000000..97b60aecbc51675fda78a948ff92add3790072a0 --- /dev/null +++ b/parse/train/2vubO341F_E/2vubO341F_E_model.json @@ -0,0 +1,14382 @@ +[ + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 397, + 966, + 1304, + 966, + 1304, + 1484, + 397, + 1484 + ], + "score": 0.984 + }, + { + "category_id": 1, + "poly": [ + 298, + 1609, + 1404, + 1609, + 1404, + 1793, + 298, + 1793 + ], + "score": 0.973 + }, + { + "category_id": 1, + "poly": [ + 300, + 1807, + 1404, + 1807, + 1404, + 1900, + 300, + 1900 + ], + "score": 0.967 + }, + { + "category_id": 0, + "poly": [ + 357, + 270, + 1345, + 270, + 1345, + 381, + 357, + 381 + ], + "score": 0.957 + }, + { + "category_id": 2, + "poly": [ + 329, + 1920, + 1133, + 1920, + 1133, + 1981, + 329, + 1981 + ], + "score": 0.937 + }, + { + "category_id": 1, + "poly": [ + 356, + 760, + 1344, + 760, + 1344, + 818, + 356, + 818 + ], + "score": 0.922 + }, + { + "category_id": 0, + "poly": [ + 299, + 1539, + 530, + 1539, + 530, + 1578, + 299, + 1578 + ], + "score": 0.909 + }, + { + "category_id": 0, + "poly": [ + 787, + 895, + 913, + 895, + 913, + 932, + 787, + 932 + ], + "score": 0.891 + }, + { + "category_id": 2, + "poly": [ + 297, + 2033, + 1071, + 2033, + 1071, + 2062, + 297, + 2062 + ], + "score": 0.89 + }, + { + "category_id": 1, + "poly": [ + 515, + 645, + 1183, + 645, + 1183, + 680, + 515, + 680 + ], + "score": 0.884 + }, + { + "category_id": 1, + "poly": [ + 451, + 565, + 1249, + 565, + 1249, + 601, + 451, + 601 + ], + "score": 0.828 + }, + { + "category_id": 1, + "poly": [ + 340, + 485, + 1352, + 485, + 1352, + 524, + 340, + 524 + ], + "score": 0.793 + }, + { + "category_id": 0, + "poly": [ + 645, + 723, + 1054, + 723, + 1054, + 756, + 645, + 756 + ], + "score": 0.509 + }, + { + "category_id": 1, + "poly": [ + 645, + 723, + 1054, + 723, + 1054, + 756, + 645, + 756 + ], + "score": 0.382 + }, + { + "category_id": 13, + "poly": [ + 964, + 1300, + 1041, + 1300, + 1041, + 1329, + 964, + 1329 + ], + "score": 0.88, + "latex": "8 6 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 643, + 1361, + 696, + 1361, + 696, + 1390, + 643, + 1390 + ], + "score": 0.87, + "latex": "86 \\%" + }, + { + "category_id": 13, + "poly": [ + 1005, + 1270, + 1082, + 1270, + 1082, + 1299, + 1005, + 1299 + ], + "score": 0.86, + "latex": "8 4 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 437, + 1361, + 534, + 1361, + 534, + 1391, + 437, + 1391 + ], + "score": 0.76, + "latex": "( 2 5 0 \\mathbf { M } + )" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 264.0, + 1349.0, + 264.0, + 1349.0, + 333.0, + 349.0, + 333.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 565.0, + 325.0, + 1137.0, + 325.0, + 1137.0, + 384.0, + 565.0, + 384.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 329.0, + 1916.0, + 801.0, + 1916.0, + 801.0, + 1951.0, + 329.0, + 1951.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 329.0, + 1946.0, + 1135.0, + 1946.0, + 1135.0, + 1984.0, + 329.0, + 1984.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1536.0, + 536.0, + 1536.0, + 536.0, + 1584.0, + 292.0, + 1584.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 783.0, + 895.0, + 918.0, + 895.0, + 918.0, + 934.0, + 783.0, + 934.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 2028.0, + 1073.0, + 2028.0, + 1073.0, + 2066.0, + 294.0, + 2066.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 639.0, + 716.0, + 1058.0, + 716.0, + 1058.0, + 767.0, + 639.0, + 767.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 964.0, + 1306.0, + 964.0, + 1306.0, + 1004.0, + 392.0, + 1004.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 993.0, + 1306.0, + 993.0, + 1306.0, + 1034.0, + 392.0, + 1034.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 395.0, + 1029.0, + 1305.0, + 1029.0, + 1305.0, + 1062.0, + 395.0, + 1062.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 1059.0, + 1308.0, + 1059.0, + 1308.0, + 1092.0, + 394.0, + 1092.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 1089.0, + 1306.0, + 1089.0, + 1306.0, + 1123.0, + 392.0, + 1123.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 1119.0, + 1305.0, + 1119.0, + 1305.0, + 1154.0, + 393.0, + 1154.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 1148.0, + 1308.0, + 1148.0, + 1308.0, + 1184.0, + 393.0, + 1184.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 1179.0, + 1306.0, + 1179.0, + 1306.0, + 1214.0, + 392.0, + 1214.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 1211.0, + 1304.0, + 1211.0, + 1304.0, + 1242.0, + 394.0, + 1242.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 1238.0, + 1306.0, + 1238.0, + 1306.0, + 1278.0, + 392.0, + 1278.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 1268.0, + 1004.0, + 1268.0, + 1004.0, + 1305.0, + 393.0, + 1305.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1083.0, + 1268.0, + 1305.0, + 1268.0, + 1305.0, + 1305.0, + 1083.0, + 1305.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 1299.0, + 963.0, + 1299.0, + 963.0, + 1336.0, + 393.0, + 1336.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1042.0, + 1299.0, + 1306.0, + 1299.0, + 1306.0, + 1336.0, + 1042.0, + 1336.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 1331.0, + 1308.0, + 1331.0, + 1308.0, + 1364.0, + 394.0, + 1364.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 1360.0, + 436.0, + 1360.0, + 436.0, + 1396.0, + 393.0, + 1396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 535.0, + 1360.0, + 642.0, + 1360.0, + 642.0, + 1396.0, + 535.0, + 1396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 697.0, + 1360.0, + 1305.0, + 1360.0, + 1305.0, + 1396.0, + 697.0, + 1396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 395.0, + 1391.0, + 1305.0, + 1391.0, + 1305.0, + 1425.0, + 395.0, + 1425.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 1419.0, + 1306.0, + 1419.0, + 1306.0, + 1459.0, + 393.0, + 1459.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 1449.0, + 1275.0, + 1449.0, + 1275.0, + 1489.0, + 392.0, + 1489.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1605.0, + 1406.0, + 1605.0, + 1406.0, + 1650.0, + 292.0, + 1650.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1640.0, + 1405.0, + 1640.0, + 1405.0, + 1676.0, + 294.0, + 1676.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1672.0, + 1405.0, + 1672.0, + 1405.0, + 1707.0, + 294.0, + 1707.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1700.0, + 1404.0, + 1700.0, + 1404.0, + 1739.0, + 293.0, + 1739.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1730.0, + 1405.0, + 1730.0, + 1405.0, + 1767.0, + 293.0, + 1767.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1763.0, + 538.0, + 1763.0, + 538.0, + 1796.0, + 293.0, + 1796.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1806.0, + 1409.0, + 1806.0, + 1409.0, + 1844.0, + 295.0, + 1844.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1838.0, + 1408.0, + 1838.0, + 1408.0, + 1872.0, + 296.0, + 1872.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1868.0, + 1405.0, + 1868.0, + 1405.0, + 1905.0, + 294.0, + 1905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 357.0, + 758.0, + 1348.0, + 758.0, + 1348.0, + 789.0, + 357.0, + 789.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 366.0, + 788.0, + 1334.0, + 788.0, + 1334.0, + 822.0, + 366.0, + 822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 512.0, + 639.0, + 1187.0, + 639.0, + 1187.0, + 692.0, + 512.0, + 692.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 445.0, + 561.0, + 599.0, + 561.0, + 599.0, + 606.0, + 445.0, + 606.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 759.0, + 563.0, + 933.0, + 563.0, + 933.0, + 607.0, + 759.0, + 607.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1093.0, + 563.0, + 1249.0, + 563.0, + 1249.0, + 607.0, + 1093.0, + 607.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 340.0, + 483.0, + 541.0, + 483.0, + 541.0, + 532.0, + 340.0, + 532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 573.0, + 484.0, + 764.0, + 484.0, + 764.0, + 525.0, + 573.0, + 525.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 793.0, + 484.0, + 921.0, + 484.0, + 921.0, + 527.0, + 793.0, + 527.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 966.0, + 483.0, + 1161.0, + 483.0, + 1161.0, + 532.0, + 966.0, + 532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1209.0, + 482.0, + 1356.0, + 482.0, + 1356.0, + 528.0, + 1209.0, + 528.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 639.0, + 716.0, + 1058.0, + 716.0, + 1058.0, + 767.0, + 639.0, + 767.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 0, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1793, + 1403, + 1793, + 1403, + 2008, + 298, + 2008 + ], + "score": 0.982 + }, + { + "category_id": 1, + "poly": [ + 298, + 1239, + 1404, + 1239, + 1404, + 1453, + 298, + 1453 + ], + "score": 0.982 + }, + { + "category_id": 1, + "poly": [ + 298, + 1565, + 1404, + 1565, + 1404, + 1781, + 298, + 1781 + ], + "score": 0.981 + }, + { + "category_id": 1, + "poly": [ + 298, + 951, + 1404, + 951, + 1404, + 1226, + 298, + 1226 + ], + "score": 0.98 + }, + { + "category_id": 1, + "poly": [ + 299, + 844, + 1403, + 844, + 1403, + 938, + 299, + 938 + ], + "score": 0.971 + }, + { + "category_id": 3, + "poly": [ + 298, + 193, + 1403, + 193, + 1403, + 585, + 298, + 585 + ], + "score": 0.971 + }, + { + "category_id": 4, + "poly": [ + 296, + 600, + 1407, + 600, + 1407, + 783, + 296, + 783 + ], + "score": 0.964 + }, + { + "category_id": 0, + "poly": [ + 299, + 1495, + 548, + 1495, + 548, + 1534, + 299, + 1534 + ], + "score": 0.914 + }, + { + "category_id": 2, + "poly": [ + 841, + 2062, + 858, + 2062, + 858, + 2084, + 841, + 2084 + ], + "score": 0.715 + }, + { + "category_id": 2, + "poly": [ + 841, + 2061, + 859, + 2061, + 859, + 2084, + 841, + 2084 + ], + "score": 0.17 + }, + { + "category_id": 13, + "poly": [ + 1282, + 723, + 1403, + 723, + 1403, + 752, + 1282, + 752 + ], + "score": 0.9, + "latex": "2 2 4 \\times 2 2 4" + }, + { + "category_id": 13, + "poly": [ + 567, + 1301, + 643, + 1301, + 643, + 1330, + 567, + 1330 + ], + "score": 0.87, + "latex": "8 5 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 804, + 1361, + 878, + 1361, + 878, + 1390, + 804, + 1390 + ], + "score": 0.87, + "latex": "8 6 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 1225, + 984, + 1253, + 984, + 1253, + 1010, + 1225, + 1010 + ], + "score": 0.82, + "latex": "K" + }, + { + "category_id": 13, + "poly": [ + 471, + 1043, + 501, + 1043, + 501, + 1070, + 471, + 1070 + ], + "score": 0.81, + "latex": "K" + }, + { + "category_id": 13, + "poly": [ + 538, + 754, + 564, + 754, + 564, + 780, + 538, + 780 + ], + "score": 0.8, + "latex": "@" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 192.0, + 361.0, + 192.0, + 361.0, + 226.0, + 320.0, + 226.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 700.0, + 190.0, + 825.0, + 190.0, + 825.0, + 224.0, + 700.0, + 224.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1274.0, + 199.0, + 1344.0, + 199.0, + 1344.0, + 230.0, + 1274.0, + 230.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 278.0, + 235.0, + 553.0, + 235.0, + 553.0, + 470.0, + 278.0, + 470.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 572.0, + 253.0, + 694.0, + 253.0, + 694.0, + 307.0, + 572.0, + 307.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 700.0, + 221.0, + 789.0, + 221.0, + 789.0, + 354.0, + 700.0, + 354.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 813.0, + 220.0, + 1139.0, + 220.0, + 1139.0, + 365.0, + 813.0, + 365.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1272.0, + 225.0, + 1370.0, + 225.0, + 1370.0, + 361.0, + 1272.0, + 361.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 372.0, + 355.0, + 372.0, + 355.0, + 403.0, + 320.0, + 403.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 431.0, + 381.0, + 498.0, + 381.0, + 498.0, + 412.0, + 431.0, + 412.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 502.0, + 392.0, + 525.0, + 392.0, + 525.0, + 412.0, + 502.0, + 412.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 646.0, + 351.0, + 779.0, + 351.0, + 779.0, + 384.0, + 646.0, + 384.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 798.0, + 351.0, + 881.0, + 351.0, + 881.0, + 421.0, + 798.0, + 421.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1079.0, + 356.0, + 1110.0, + 356.0, + 1110.0, + 386.0, + 1079.0, + 386.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1270.0, + 353.0, + 1364.0, + 353.0, + 1364.0, + 415.0, + 1270.0, + 415.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 446.0, + 400.0, + 576.0, + 400.0, + 576.0, + 449.0, + 446.0, + 449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1272.0, + 407.0, + 1348.0, + 407.0, + 1348.0, + 438.0, + 1272.0, + 438.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1274.0, + 434.0, + 1405.0, + 434.0, + 1405.0, + 463.0, + 1274.0, + 463.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 315.0, + 459.0, + 358.0, + 459.0, + 358.0, + 535.0, + 315.0, + 535.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1075.0, + 474.0, + 1113.0, + 474.0, + 1113.0, + 518.0, + 1075.0, + 518.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1272.0, + 458.0, + 1364.0, + 458.0, + 1364.0, + 546.0, + 1272.0, + 546.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 346.0, + 530.0, + 381.0, + 530.0, + 381.0, + 562.0, + 346.0, + 562.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 568.0, + 530.0, + 602.0, + 530.0, + 602.0, + 561.0, + 568.0, + 561.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 786.0, + 527.0, + 827.0, + 527.0, + 827.0, + 562.0, + 786.0, + 562.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1010.0, + 530.0, + 1044.0, + 530.0, + 1044.0, + 561.0, + 1010.0, + 561.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1226.0, + 530.0, + 1331.0, + 530.0, + 1331.0, + 568.0, + 1226.0, + 568.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 736.0, + 550.0, + 877.0, + 550.0, + 877.0, + 592.0, + 736.0, + 592.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 598.0, + 1409.0, + 598.0, + 1409.0, + 638.0, + 293.0, + 638.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 632.0, + 1409.0, + 632.0, + 1409.0, + 667.0, + 294.0, + 667.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 661.0, + 1407.0, + 661.0, + 1407.0, + 697.0, + 294.0, + 697.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 692.0, + 1405.0, + 692.0, + 1405.0, + 726.0, + 293.0, + 726.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 722.0, + 1281.0, + 722.0, + 1281.0, + 757.0, + 293.0, + 757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1404.0, + 722.0, + 1407.0, + 722.0, + 1407.0, + 757.0, + 1404.0, + 757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 752.0, + 537.0, + 752.0, + 537.0, + 788.0, + 295.0, + 788.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 565.0, + 752.0, + 577.0, + 752.0, + 577.0, + 788.0, + 565.0, + 788.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1493.0, + 552.0, + 1493.0, + 552.0, + 1538.0, + 292.0, + 1538.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2059.0, + 862.0, + 2059.0, + 862.0, + 2093.0, + 839.0, + 2093.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2058.0, + 862.0, + 2058.0, + 862.0, + 2093.0, + 838.0, + 2093.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1792.0, + 1404.0, + 1792.0, + 1404.0, + 1828.0, + 294.0, + 1828.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1824.0, + 1404.0, + 1824.0, + 1404.0, + 1859.0, + 296.0, + 1859.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1853.0, + 1407.0, + 1853.0, + 1407.0, + 1891.0, + 292.0, + 1891.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1885.0, + 1405.0, + 1885.0, + 1405.0, + 1919.0, + 292.0, + 1919.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1912.0, + 1405.0, + 1912.0, + 1405.0, + 1952.0, + 292.0, + 1952.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1946.0, + 1404.0, + 1946.0, + 1404.0, + 1980.0, + 294.0, + 1980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1974.0, + 879.0, + 1974.0, + 879.0, + 2011.0, + 295.0, + 2011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1239.0, + 1406.0, + 1239.0, + 1406.0, + 1274.0, + 294.0, + 1274.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1270.0, + 1405.0, + 1270.0, + 1405.0, + 1303.0, + 293.0, + 1303.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1298.0, + 566.0, + 1298.0, + 566.0, + 1335.0, + 293.0, + 1335.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 644.0, + 1298.0, + 1406.0, + 1298.0, + 1406.0, + 1335.0, + 644.0, + 1335.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1326.0, + 1404.0, + 1326.0, + 1404.0, + 1368.0, + 293.0, + 1368.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1362.0, + 803.0, + 1362.0, + 803.0, + 1393.0, + 296.0, + 1393.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 879.0, + 1362.0, + 1402.0, + 1362.0, + 1402.0, + 1393.0, + 879.0, + 1393.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1390.0, + 1409.0, + 1390.0, + 1409.0, + 1427.0, + 292.0, + 1427.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1422.0, + 651.0, + 1422.0, + 651.0, + 1456.0, + 294.0, + 1456.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1566.0, + 1405.0, + 1566.0, + 1405.0, + 1600.0, + 294.0, + 1600.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1595.0, + 1407.0, + 1595.0, + 1407.0, + 1632.0, + 293.0, + 1632.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1626.0, + 1405.0, + 1626.0, + 1405.0, + 1661.0, + 294.0, + 1661.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1658.0, + 1405.0, + 1658.0, + 1405.0, + 1693.0, + 294.0, + 1693.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1688.0, + 1408.0, + 1688.0, + 1408.0, + 1722.0, + 294.0, + 1722.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1717.0, + 1406.0, + 1717.0, + 1406.0, + 1752.0, + 293.0, + 1752.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1750.0, + 784.0, + 1750.0, + 784.0, + 1784.0, + 296.0, + 1784.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 953.0, + 1405.0, + 953.0, + 1405.0, + 986.0, + 296.0, + 986.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 982.0, + 1224.0, + 982.0, + 1224.0, + 1014.0, + 296.0, + 1014.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1254.0, + 982.0, + 1404.0, + 982.0, + 1404.0, + 1014.0, + 1254.0, + 1014.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1013.0, + 1406.0, + 1013.0, + 1406.0, + 1048.0, + 293.0, + 1048.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1042.0, + 470.0, + 1042.0, + 470.0, + 1078.0, + 294.0, + 1078.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 502.0, + 1042.0, + 1405.0, + 1042.0, + 1405.0, + 1078.0, + 502.0, + 1078.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1074.0, + 1406.0, + 1074.0, + 1406.0, + 1111.0, + 293.0, + 1111.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1104.0, + 1406.0, + 1104.0, + 1406.0, + 1141.0, + 294.0, + 1141.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1134.0, + 1406.0, + 1134.0, + 1406.0, + 1168.0, + 293.0, + 1168.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1161.0, + 1406.0, + 1161.0, + 1406.0, + 1199.0, + 293.0, + 1199.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1194.0, + 704.0, + 1194.0, + 704.0, + 1230.0, + 294.0, + 1230.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 846.0, + 1403.0, + 846.0, + 1403.0, + 880.0, + 297.0, + 880.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 876.0, + 1405.0, + 876.0, + 1405.0, + 910.0, + 295.0, + 910.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 903.0, + 1141.0, + 903.0, + 1141.0, + 942.0, + 293.0, + 942.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 1, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1735, + 1402, + 1735, + 1402, + 1893, + 298, + 1893 + ], + "score": 0.978 + }, + { + "category_id": 1, + "poly": [ + 298, + 1164, + 1405, + 1164, + 1405, + 1318, + 298, + 1318 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 298, + 996, + 1404, + 996, + 1404, + 1151, + 298, + 1151 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 297, + 1598, + 1403, + 1598, + 1403, + 1722, + 297, + 1722 + ], + "score": 0.977 + }, + { + "category_id": 4, + "poly": [ + 295, + 719, + 1405, + 719, + 1405, + 934, + 295, + 934 + ], + "score": 0.973 + }, + { + "category_id": 3, + "poly": [ + 356, + 198, + 1351, + 198, + 1351, + 701, + 356, + 701 + ], + "score": 0.972 + }, + { + "category_id": 1, + "poly": [ + 298, + 1437, + 1402, + 1437, + 1402, + 1501, + 298, + 1501 + ], + "score": 0.956 + }, + { + "category_id": 8, + "poly": [ + 723, + 1912, + 976, + 1912, + 976, + 1955, + 723, + 1955 + ], + "score": 0.933 + }, + { + "category_id": 1, + "poly": [ + 297, + 1973, + 1110, + 1973, + 1110, + 2009, + 297, + 2009 + ], + "score": 0.926 + }, + { + "category_id": 0, + "poly": [ + 300, + 1540, + 719, + 1540, + 719, + 1573, + 300, + 1573 + ], + "score": 0.92 + }, + { + "category_id": 0, + "poly": [ + 300, + 1364, + 692, + 1364, + 692, + 1404, + 300, + 1404 + ], + "score": 0.919 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1919, + 1400, + 1919, + 1400, + 1949, + 1366, + 1949 + ], + "score": 0.873 + }, + { + "category_id": 2, + "poly": [ + 841, + 2061, + 858, + 2061, + 858, + 2085, + 841, + 2085 + ], + "score": 0.654 + }, + { + "category_id": 2, + "poly": [ + 841, + 2061, + 859, + 2061, + 859, + 2085, + 841, + 2085 + ], + "score": 0.455 + }, + { + "category_id": 13, + "poly": [ + 371, + 1975, + 448, + 1975, + 448, + 2009, + 371, + 2009 + ], + "score": 0.93, + "latex": "H ( \\cdot , \\cdot )" + }, + { + "category_id": 13, + "poly": [ + 1002, + 1796, + 1216, + 1796, + 1216, + 1830, + 1002, + 1830 + ], + "score": 0.93, + "latex": "[ \\bar { X } ^ { c l s } , X ^ { 1 } , . . . , X ^ { N } ]" + }, + { + "category_id": 14, + "poly": [ + 723, + 1912, + 976, + 1912, + 976, + 1953, + 723, + 1953 + ], + "score": 0.92, + "latex": "L _ { c l s } = H ( { X } ^ { c l s } , { y } ^ { c l s } ) ," + }, + { + "category_id": 13, + "poly": [ + 784, + 1831, + 916, + 1831, + 916, + 1862, + 784, + 1862 + ], + "score": 0.92, + "latex": "X ^ { 1 } , . . . , X ^ { N }" + }, + { + "category_id": 13, + "poly": [ + 870, + 1977, + 913, + 1977, + 913, + 2008, + 870, + 2008 + ], + "score": 0.92, + "latex": "y ^ { c l s }" + }, + { + "category_id": 13, + "poly": [ + 676, + 1832, + 731, + 1832, + 731, + 1857, + 676, + 1857 + ], + "score": 0.9, + "latex": "X ^ { c l s }" + }, + { + "category_id": 13, + "poly": [ + 895, + 1866, + 912, + 1866, + 912, + 1888, + 895, + 1888 + ], + "score": 0.85, + "latex": "I" + }, + { + "category_id": 13, + "poly": [ + 1304, + 1798, + 1332, + 1798, + 1332, + 1824, + 1304, + 1824 + ], + "score": 0.82, + "latex": "N" + }, + { + "category_id": 13, + "poly": [ + 407, + 1798, + 426, + 1798, + 426, + 1825, + 407, + 1825 + ], + "score": 0.56, + "latex": "I" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 719.0, + 1403.0, + 719.0, + 1403.0, + 754.0, + 295.0, + 754.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 749.0, + 1403.0, + 749.0, + 1403.0, + 783.0, + 295.0, + 783.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 778.0, + 1407.0, + 778.0, + 1407.0, + 818.0, + 292.0, + 818.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 813.0, + 1403.0, + 813.0, + 1403.0, + 844.0, + 295.0, + 844.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 839.0, + 1406.0, + 839.0, + 1406.0, + 877.0, + 293.0, + 877.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 871.0, + 1405.0, + 871.0, + 1405.0, + 905.0, + 295.0, + 905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 899.0, + 1216.0, + 899.0, + 1216.0, + 936.0, + 294.0, + 936.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 907.0, + 203.0, + 1043.0, + 203.0, + 1043.0, + 240.0, + 907.0, + 240.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 567.0, + 332.0, + 692.0, + 332.0, + 692.0, + 359.0, + 567.0, + 359.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 732.0, + 369.0, + 823.0, + 369.0, + 823.0, + 396.0, + 732.0, + 396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 881.0, + 374.0, + 933.0, + 374.0, + 933.0, + 393.0, + 881.0, + 393.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 372.0, + 427.0, + 518.0, + 427.0, + 518.0, + 460.0, + 372.0, + 460.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 861.0, + 418.0, + 1171.0, + 418.0, + 1171.0, + 447.0, + 861.0, + 447.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 567.0, + 515.0, + 687.0, + 515.0, + 687.0, + 549.0, + 567.0, + 549.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 815.0, + 582.0, + 1133.0, + 582.0, + 1133.0, + 612.0, + 815.0, + 612.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1243.0, + 579.0, + 1352.0, + 579.0, + 1352.0, + 613.0, + 1243.0, + 613.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 390.0, + 668.0, + 498.0, + 668.0, + 498.0, + 701.0, + 390.0, + 701.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 762.0, + 640.0, + 1247.0, + 640.0, + 1247.0, + 707.0, + 762.0, + 707.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1275.0, + 249.0, + 1326.0, + 249.0, + 1326.0, + 277.0, + 1275.0, + 277.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1539.0, + 724.0, + 1539.0, + 724.0, + 1578.0, + 294.0, + 1578.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1361.0, + 694.0, + 1361.0, + 694.0, + 1409.0, + 291.0, + 1409.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2058.0, + 860.0, + 2058.0, + 860.0, + 2089.0, + 839.0, + 2089.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2058.0, + 862.0, + 2058.0, + 862.0, + 2091.0, + 838.0, + 2091.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1735.0, + 1404.0, + 1735.0, + 1404.0, + 1772.0, + 294.0, + 1772.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1760.0, + 1406.0, + 1760.0, + 1406.0, + 1805.0, + 293.0, + 1805.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1791.0, + 406.0, + 1791.0, + 406.0, + 1831.0, + 292.0, + 1831.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 427.0, + 1791.0, + 1001.0, + 1791.0, + 1001.0, + 1831.0, + 427.0, + 1831.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1217.0, + 1791.0, + 1303.0, + 1791.0, + 1303.0, + 1831.0, + 1217.0, + 1831.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1333.0, + 1791.0, + 1407.0, + 1791.0, + 1407.0, + 1831.0, + 1333.0, + 1831.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1822.0, + 675.0, + 1822.0, + 675.0, + 1868.0, + 292.0, + 1868.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 732.0, + 1822.0, + 783.0, + 1822.0, + 783.0, + 1868.0, + 732.0, + 1868.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 917.0, + 1822.0, + 1408.0, + 1822.0, + 1408.0, + 1868.0, + 917.0, + 1868.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1861.0, + 894.0, + 1861.0, + 894.0, + 1895.0, + 296.0, + 1895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 913.0, + 1861.0, + 1111.0, + 1861.0, + 1111.0, + 1895.0, + 913.0, + 1895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1164.0, + 1406.0, + 1164.0, + 1406.0, + 1197.0, + 297.0, + 1197.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1196.0, + 1405.0, + 1196.0, + 1405.0, + 1228.0, + 294.0, + 1228.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1226.0, + 1405.0, + 1226.0, + 1405.0, + 1260.0, + 296.0, + 1260.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1257.0, + 1405.0, + 1257.0, + 1405.0, + 1290.0, + 296.0, + 1290.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1289.0, + 713.0, + 1289.0, + 713.0, + 1318.0, + 297.0, + 1318.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 998.0, + 1404.0, + 998.0, + 1404.0, + 1031.0, + 297.0, + 1031.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1030.0, + 1404.0, + 1030.0, + 1404.0, + 1063.0, + 296.0, + 1063.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1058.0, + 1405.0, + 1058.0, + 1405.0, + 1095.0, + 294.0, + 1095.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1091.0, + 1404.0, + 1091.0, + 1404.0, + 1124.0, + 294.0, + 1124.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1118.0, + 398.0, + 1118.0, + 398.0, + 1152.0, + 294.0, + 1152.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1600.0, + 1405.0, + 1600.0, + 1405.0, + 1633.0, + 293.0, + 1633.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1630.0, + 1405.0, + 1630.0, + 1405.0, + 1662.0, + 293.0, + 1662.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1661.0, + 1405.0, + 1661.0, + 1405.0, + 1694.0, + 295.0, + 1694.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1689.0, + 1124.0, + 1689.0, + 1124.0, + 1725.0, + 293.0, + 1725.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1437.0, + 1404.0, + 1437.0, + 1404.0, + 1473.0, + 294.0, + 1473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1468.0, + 837.0, + 1468.0, + 837.0, + 1505.0, + 294.0, + 1505.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1970.0, + 370.0, + 1970.0, + 370.0, + 2013.0, + 294.0, + 2013.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 449.0, + 1970.0, + 869.0, + 1970.0, + 869.0, + 2013.0, + 449.0, + 2013.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 914.0, + 1970.0, + 1114.0, + 1970.0, + 1114.0, + 2013.0, + 914.0, + 2013.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 2, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1641, + 1405, + 1641, + 1405, + 2007, + 297, + 2007 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 297, + 807, + 1404, + 807, + 1404, + 1023, + 297, + 1023 + ], + "score": 0.982 + }, + { + "category_id": 1, + "poly": [ + 298, + 1035, + 1404, + 1035, + 1404, + 1189, + 298, + 1189 + ], + "score": 0.979 + }, + { + "category_id": 1, + "poly": [ + 298, + 671, + 1403, + 671, + 1403, + 794, + 298, + 794 + ], + "score": 0.978 + }, + { + "category_id": 4, + "poly": [ + 861, + 195, + 1365, + 195, + 1365, + 531, + 861, + 531 + ], + "score": 0.967 + }, + { + "category_id": 8, + "poly": [ + 709, + 1209, + 989, + 1209, + 989, + 1299, + 709, + 1299 + ], + "score": 0.958 + }, + { + "category_id": 1, + "poly": [ + 298, + 1566, + 1404, + 1566, + 1404, + 1627, + 298, + 1627 + ], + "score": 0.952 + }, + { + "category_id": 3, + "poly": [ + 338, + 200, + 835, + 200, + 835, + 535, + 338, + 535 + ], + "score": 0.949 + }, + { + "category_id": 8, + "poly": [ + 653, + 1445, + 1121, + 1445, + 1121, + 1534, + 653, + 1534 + ], + "score": 0.941 + }, + { + "category_id": 0, + "poly": [ + 298, + 608, + 549, + 608, + 549, + 642, + 298, + 642 + ], + "score": 0.917 + }, + { + "category_id": 1, + "poly": [ + 294, + 1330, + 1302, + 1330, + 1302, + 1364, + 294, + 1364 + ], + "score": 0.903 + }, + { + "category_id": 8, + "poly": [ + 579, + 1396, + 959, + 1396, + 959, + 1436, + 579, + 1436 + ], + "score": 0.896 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1238, + 1400, + 1238, + 1400, + 1268, + 1366, + 1268 + ], + "score": 0.886 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1402, + 1400, + 1402, + 1400, + 1433, + 1366, + 1433 + ], + "score": 0.885 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1474, + 1400, + 1474, + 1400, + 1503, + 1366, + 1503 + ], + "score": 0.877 + }, + { + "category_id": 2, + "poly": [ + 841, + 2062, + 858, + 2062, + 858, + 2084, + 841, + 2084 + ], + "score": 0.764 + }, + { + "category_id": 14, + "poly": [ + 706, + 1204, + 993, + 1204, + 993, + 1300, + 706, + 1300 + ], + "score": 0.95, + "latex": "L _ { t l } = \\frac { 1 } { N } \\sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) ." + }, + { + "category_id": 13, + "poly": [ + 1159, + 959, + 1286, + 959, + 1286, + 994, + 1159, + 994 + ], + "score": 0.94, + "latex": "[ y ^ { 1 } , . . . , y ^ { N } ]" + }, + { + "category_id": 13, + "poly": [ + 908, + 1126, + 1035, + 1126, + 1035, + 1161, + 908, + 1161 + ], + "score": 0.94, + "latex": "[ y ^ { 1 } , . . . , y ^ { N } ]" + }, + { + "category_id": 14, + "poly": [ + 575, + 1392, + 1120, + 1392, + 1120, + 1537, + 575, + 1537 + ], + "score": 0.94, + "latex": "\\begin{array} { l } { { { \\cal L } _ { t o t a l } = H ( X ^ { c l s } , y ^ { c l s } ) + \\beta \\cdot L _ { t l } , } } \\\\ { { { } } } \\\\ { { = H ( X ^ { c l s } , y ^ { c l s } ) + \\beta \\cdot \\displaystyle \\frac { 1 } { N } \\sum _ { i = 1 } ^ { N } H ( X ^ { i } , y ^ { i } ) , } } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 444, + 1126, + 577, + 1126, + 577, + 1159, + 444, + 1159 + ], + "score": 0.93, + "latex": "X ^ { 1 } , . . . , X ^ { N }" + }, + { + "category_id": 13, + "poly": [ + 1313, + 928, + 1359, + 928, + 1359, + 962, + 1313, + 962 + ], + "score": 0.92, + "latex": "y ^ { c l s }" + }, + { + "category_id": 13, + "poly": [ + 368, + 962, + 456, + 962, + 456, + 989, + 368, + 989 + ], + "score": 0.88, + "latex": "K \\times N" + }, + { + "category_id": 13, + "poly": [ + 372, + 1568, + 392, + 1568, + 392, + 1598, + 372, + 1598 + ], + "score": 0.85, + "latex": "\\beta" + }, + { + "category_id": 13, + "poly": [ + 424, + 1334, + 452, + 1334, + 452, + 1359, + 424, + 1359 + ], + "score": 0.84, + "latex": "H" + }, + { + "category_id": 13, + "poly": [ + 662, + 962, + 690, + 962, + 690, + 988, + 662, + 988 + ], + "score": 0.84, + "latex": "K" + }, + { + "category_id": 13, + "poly": [ + 1373, + 962, + 1401, + 962, + 1401, + 988, + 1373, + 988 + ], + "score": 0.82, + "latex": "N" + }, + { + "category_id": 13, + "poly": [ + 1065, + 931, + 1093, + 931, + 1093, + 959, + 1065, + 959 + ], + "score": 0.82, + "latex": "K" + }, + { + "category_id": 15, + "poly": [ + 861.0, + 196.0, + 1363.0, + 196.0, + 1363.0, + 230.0, + 861.0, + 230.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 861.0, + 227.0, + 1367.0, + 227.0, + 1367.0, + 261.0, + 861.0, + 261.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 861.0, + 257.0, + 1364.0, + 257.0, + 1364.0, + 291.0, + 861.0, + 291.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 861.0, + 289.0, + 1364.0, + 289.0, + 1364.0, + 319.0, + 861.0, + 319.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 860.0, + 317.0, + 1365.0, + 317.0, + 1365.0, + 350.0, + 860.0, + 350.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 859.0, + 347.0, + 1367.0, + 347.0, + 1367.0, + 380.0, + 859.0, + 380.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 860.0, + 380.0, + 1366.0, + 380.0, + 1366.0, + 410.0, + 860.0, + 410.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 860.0, + 408.0, + 1366.0, + 408.0, + 1366.0, + 441.0, + 860.0, + 441.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 861.0, + 439.0, + 1365.0, + 439.0, + 1365.0, + 472.0, + 861.0, + 472.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 859.0, + 467.0, + 1369.0, + 467.0, + 1369.0, + 503.0, + 859.0, + 503.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 859.0, + 498.0, + 1365.0, + 498.0, + 1365.0, + 531.0, + 859.0, + 531.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 442.0, + 393.0, + 454.0, + 393.0, + 454.0, + 405.0, + 442.0, + 405.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 601.0, + 552.0, + 601.0, + 552.0, + 649.0, + 292.0, + 649.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2059.0, + 862.0, + 2059.0, + 862.0, + 2091.0, + 838.0, + 2091.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1640.0, + 1405.0, + 1640.0, + 1405.0, + 1680.0, + 291.0, + 1680.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1670.0, + 1406.0, + 1670.0, + 1406.0, + 1710.0, + 292.0, + 1710.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1700.0, + 1408.0, + 1700.0, + 1408.0, + 1740.0, + 291.0, + 1740.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1734.0, + 1405.0, + 1734.0, + 1405.0, + 1768.0, + 295.0, + 1768.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1762.0, + 1406.0, + 1762.0, + 1406.0, + 1799.0, + 292.0, + 1799.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1792.0, + 1406.0, + 1792.0, + 1406.0, + 1829.0, + 292.0, + 1829.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1823.0, + 1405.0, + 1823.0, + 1405.0, + 1859.0, + 292.0, + 1859.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1854.0, + 1405.0, + 1854.0, + 1405.0, + 1888.0, + 295.0, + 1888.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1885.0, + 1405.0, + 1885.0, + 1405.0, + 1919.0, + 295.0, + 1919.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1914.0, + 1403.0, + 1914.0, + 1403.0, + 1948.0, + 294.0, + 1948.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1947.0, + 1405.0, + 1947.0, + 1405.0, + 1981.0, + 295.0, + 1981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1979.0, + 561.0, + 1979.0, + 561.0, + 2008.0, + 295.0, + 2008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 808.0, + 1404.0, + 808.0, + 1404.0, + 842.0, + 295.0, + 842.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 836.0, + 1404.0, + 836.0, + 1404.0, + 877.0, + 294.0, + 877.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 869.0, + 1406.0, + 869.0, + 1406.0, + 904.0, + 295.0, + 904.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 901.0, + 1405.0, + 901.0, + 1405.0, + 932.0, + 296.0, + 932.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 929.0, + 1064.0, + 929.0, + 1064.0, + 963.0, + 294.0, + 963.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1094.0, + 929.0, + 1312.0, + 929.0, + 1312.0, + 963.0, + 1094.0, + 963.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1360.0, + 929.0, + 1405.0, + 929.0, + 1405.0, + 963.0, + 1360.0, + 963.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 957.0, + 367.0, + 957.0, + 367.0, + 998.0, + 291.0, + 998.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 457.0, + 957.0, + 661.0, + 957.0, + 661.0, + 998.0, + 457.0, + 998.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 691.0, + 957.0, + 1158.0, + 957.0, + 1158.0, + 998.0, + 691.0, + 998.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1287.0, + 957.0, + 1372.0, + 957.0, + 1372.0, + 998.0, + 1287.0, + 998.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1402.0, + 957.0, + 1406.0, + 957.0, + 1406.0, + 998.0, + 1402.0, + 998.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 989.0, + 756.0, + 989.0, + 756.0, + 1025.0, + 292.0, + 1025.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1034.0, + 1405.0, + 1034.0, + 1405.0, + 1072.0, + 295.0, + 1072.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1067.0, + 1405.0, + 1067.0, + 1405.0, + 1100.0, + 296.0, + 1100.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1097.0, + 1405.0, + 1097.0, + 1405.0, + 1131.0, + 296.0, + 1131.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1121.0, + 443.0, + 1121.0, + 443.0, + 1166.0, + 292.0, + 1166.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 578.0, + 1121.0, + 907.0, + 1121.0, + 907.0, + 1166.0, + 578.0, + 1166.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1036.0, + 1121.0, + 1408.0, + 1121.0, + 1408.0, + 1166.0, + 1036.0, + 1166.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1157.0, + 449.0, + 1157.0, + 449.0, + 1191.0, + 296.0, + 1191.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 670.0, + 1404.0, + 670.0, + 1404.0, + 708.0, + 292.0, + 708.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 699.0, + 1404.0, + 699.0, + 1404.0, + 740.0, + 292.0, + 740.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 734.0, + 1404.0, + 734.0, + 1404.0, + 770.0, + 294.0, + 770.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 764.0, + 677.0, + 764.0, + 677.0, + 796.0, + 293.0, + 796.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1561.0, + 371.0, + 1561.0, + 371.0, + 1603.0, + 292.0, + 1603.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 1561.0, + 1405.0, + 1561.0, + 1405.0, + 1603.0, + 393.0, + 1603.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 289.0, + 1591.0, + 351.0, + 1591.0, + 351.0, + 1632.0, + 289.0, + 1632.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1330.0, + 423.0, + 1330.0, + 423.0, + 1366.0, + 295.0, + 1366.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 453.0, + 1330.0, + 1301.0, + 1330.0, + 1301.0, + 1366.0, + 453.0, + 1366.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 3, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1671, + 1404, + 1671, + 1404, + 2007, + 297, + 2007 + ], + "score": 0.984 + }, + { + "category_id": 1, + "poly": [ + 297, + 261, + 1405, + 261, + 1405, + 568, + 297, + 568 + ], + "score": 0.984 + }, + { + "category_id": 1, + "poly": [ + 298, + 1471, + 1404, + 1471, + 1404, + 1658, + 298, + 1658 + ], + "score": 0.98 + }, + { + "category_id": 1, + "poly": [ + 297, + 1274, + 1404, + 1274, + 1404, + 1460, + 297, + 1460 + ], + "score": 0.979 + }, + { + "category_id": 1, + "poly": [ + 298, + 579, + 1403, + 579, + 1403, + 704, + 298, + 704 + ], + "score": 0.976 + }, + { + "category_id": 1, + "poly": [ + 297, + 791, + 1400, + 791, + 1400, + 855, + 297, + 855 + ], + "score": 0.952 + }, + { + "category_id": 8, + "poly": [ + 677, + 993, + 1021, + 993, + 1021, + 1038, + 677, + 1038 + ], + "score": 0.944 + }, + { + "category_id": 8, + "poly": [ + 670, + 880, + 1030, + 880, + 1030, + 919, + 670, + 919 + ], + "score": 0.937 + }, + { + "category_id": 8, + "poly": [ + 670, + 729, + 1030, + 729, + 1030, + 770, + 670, + 770 + ], + "score": 0.932 + }, + { + "category_id": 1, + "poly": [ + 300, + 942, + 806, + 942, + 806, + 974, + 300, + 974 + ], + "score": 0.931 + }, + { + "category_id": 1, + "poly": [ + 296, + 1057, + 873, + 1057, + 873, + 1090, + 296, + 1090 + ], + "score": 0.929 + }, + { + "category_id": 0, + "poly": [ + 299, + 1215, + 578, + 1215, + 578, + 1249, + 299, + 1249 + ], + "score": 0.925 + }, + { + "category_id": 0, + "poly": [ + 298, + 1141, + 533, + 1141, + 533, + 1179, + 298, + 1179 + ], + "score": 0.922 + }, + { + "category_id": 0, + "poly": [ + 300, + 202, + 734, + 202, + 734, + 235, + 300, + 235 + ], + "score": 0.917 + }, + { + "category_id": 9, + "poly": [ + 1366, + 735, + 1400, + 735, + 1400, + 765, + 1366, + 765 + ], + "score": 0.884 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1001, + 1400, + 1001, + 1400, + 1032, + 1366, + 1032 + ], + "score": 0.882 + }, + { + "category_id": 9, + "poly": [ + 1366, + 886, + 1400, + 886, + 1400, + 915, + 1366, + 915 + ], + "score": 0.878 + }, + { + "category_id": 2, + "poly": [ + 841, + 2061, + 858, + 2061, + 858, + 2085, + 841, + 2085 + ], + "score": 0.705 + }, + { + "category_id": 2, + "poly": [ + 841, + 2061, + 859, + 2061, + 859, + 2085, + 841, + 2085 + ], + "score": 0.144 + }, + { + "category_id": 13, + "poly": [ + 752, + 640, + 936, + 640, + 936, + 675, + 752, + 675 + ], + "score": 0.93, + "latex": "T _ { 1 } = [ t _ { 1 } ^ { 1 } , . . . , t _ { 1 } ^ { N } ]" + }, + { + "category_id": 13, + "poly": [ + 407, + 608, + 600, + 608, + 600, + 644, + 407, + 644 + ], + "score": 0.93, + "latex": "Y _ { 2 } = [ y _ { 2 } ^ { 1 } , . . . , y _ { 2 } ^ { N } ]" + }, + { + "category_id": 13, + "poly": [ + 394, + 1502, + 659, + 1502, + 659, + 1542, + 394, + 1542 + ], + "score": 0.93, + "latex": "\\begin{array} { r } { l r = 1 0 ^ { - 3 } \\times \\frac { b a t c h \\_ s i z e } { 6 4 0 } } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 987, + 640, + 1171, + 640, + 1171, + 675, + 987, + 675 + ], + "score": 0.93, + "latex": "\\bar { T _ { 2 } } = [ t _ { 2 } ^ { 1 } , . . . , t _ { 2 } ^ { N } ]" + }, + { + "category_id": 13, + "poly": [ + 743, + 583, + 806, + 583, + 806, + 612, + 743, + 612 + ], + "score": 0.93, + "latex": "I _ { 1 } , I _ { 2 }" + }, + { + "category_id": 13, + "poly": [ + 325, + 1883, + 530, + 1883, + 530, + 1913, + 325, + 1913 + ], + "score": 0.93, + "latex": "L \\in \\mathbb { R } ^ { 1 8 \\times 1 8 \\times 1 0 0 0 }" + }, + { + "category_id": 13, + "poly": [ + 1207, + 577, + 1400, + 577, + 1400, + 614, + 1207, + 614 + ], + "score": 0.93, + "latex": "Y _ { 1 } = [ y _ { 1 } ^ { 1 } , . . . , y _ { 1 } ^ { N } ]" + }, + { + "category_id": 14, + "poly": [ + 677, + 989, + 1021, + 989, + 1021, + 1036, + 677, + 1036 + ], + "score": 0.92, + "latex": "y ^ { \\hat { c } l s } = \\bar { M } y _ { 1 } ^ { c l s } + ( 1 - \\bar { M } ) y _ { 2 } ^ { c l s } ," + }, + { + "category_id": 14, + "poly": [ + 668, + 875, + 1032, + 875, + 1032, + 920, + 668, + 920 + ], + "score": 0.92, + "latex": "\\hat { Y } = Y _ { 1 } \\odot M + Y _ { 2 } \\odot ( 1 - M ) ." + }, + { + "category_id": 13, + "poly": [ + 713, + 1503, + 822, + 1503, + 822, + 1535, + 713, + 1535 + ], + "score": 0.91, + "latex": "5 \\times 1 0 ^ { - 2 }" + }, + { + "category_id": 14, + "poly": [ + 669, + 726, + 1028, + 726, + 1028, + 770, + 669, + 770 + ], + "score": 0.91, + "latex": "\\hat { T } = T _ { 1 } \\odot M + T _ { 2 } \\odot ( 1 - M ) ," + }, + { + "category_id": 13, + "poly": [ + 446, + 1855, + 565, + 1855, + 565, + 1883, + 446, + 1883 + ], + "score": 0.89, + "latex": "5 7 6 \\times 5 7 6" + }, + { + "category_id": 13, + "poly": [ + 953, + 1671, + 1029, + 1671, + 1029, + 1703, + 953, + 1703 + ], + "score": 0.87, + "latex": "8 6 . 3 \\%" + }, + { + "category_id": 13, + "poly": [ + 371, + 1057, + 403, + 1057, + 403, + 1086, + 371, + 1086 + ], + "score": 0.87, + "latex": "\\bar { M }" + }, + { + "category_id": 13, + "poly": [ + 832, + 1060, + 864, + 1060, + 864, + 1086, + 832, + 1086 + ], + "score": 0.86, + "latex": "M" + }, + { + "category_id": 13, + "poly": [ + 1135, + 823, + 1167, + 823, + 1167, + 850, + 1135, + 850 + ], + "score": 0.83, + "latex": "M" + }, + { + "category_id": 13, + "poly": [ + 1056, + 675, + 1088, + 675, + 1088, + 699, + 1056, + 699 + ], + "score": 0.83, + "latex": "M" + }, + { + "category_id": 13, + "poly": [ + 373, + 795, + 398, + 795, + 398, + 821, + 373, + 821 + ], + "score": 0.83, + "latex": "\\odot" + }, + { + "category_id": 13, + "poly": [ + 1252, + 793, + 1285, + 793, + 1285, + 819, + 1252, + 819 + ], + "score": 0.82, + "latex": "M" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1208.0, + 583.0, + 1208.0, + 583.0, + 1260.0, + 292.0, + 1260.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 289.0, + 1136.0, + 537.0, + 1136.0, + 537.0, + 1187.0, + 289.0, + 1187.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 198.0, + 737.0, + 198.0, + 737.0, + 242.0, + 291.0, + 242.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2058.0, + 861.0, + 2058.0, + 861.0, + 2093.0, + 839.0, + 2093.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2058.0, + 862.0, + 2058.0, + 862.0, + 2092.0, + 838.0, + 2092.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1671.0, + 952.0, + 1671.0, + 952.0, + 1709.0, + 294.0, + 1709.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1030.0, + 1671.0, + 1405.0, + 1671.0, + 1405.0, + 1709.0, + 1030.0, + 1709.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1703.0, + 1405.0, + 1703.0, + 1405.0, + 1737.0, + 294.0, + 1737.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1732.0, + 1406.0, + 1732.0, + 1406.0, + 1771.0, + 294.0, + 1771.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1763.0, + 1405.0, + 1763.0, + 1405.0, + 1798.0, + 294.0, + 1798.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1791.0, + 1406.0, + 1791.0, + 1406.0, + 1832.0, + 291.0, + 1832.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1824.0, + 1405.0, + 1824.0, + 1405.0, + 1859.0, + 295.0, + 1859.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1850.0, + 445.0, + 1850.0, + 445.0, + 1892.0, + 291.0, + 1892.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 566.0, + 1850.0, + 1405.0, + 1850.0, + 1405.0, + 1892.0, + 566.0, + 1892.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 289.0, + 1875.0, + 324.0, + 1875.0, + 324.0, + 1926.0, + 289.0, + 1926.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 531.0, + 1875.0, + 1408.0, + 1875.0, + 1408.0, + 1926.0, + 531.0, + 1926.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1913.0, + 1406.0, + 1913.0, + 1406.0, + 1952.0, + 292.0, + 1952.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1944.0, + 1406.0, + 1944.0, + 1406.0, + 1981.0, + 292.0, + 1981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1974.0, + 1335.0, + 1974.0, + 1335.0, + 2014.0, + 292.0, + 2014.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 262.0, + 1408.0, + 262.0, + 1408.0, + 297.0, + 295.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 294.0, + 1405.0, + 294.0, + 1405.0, + 325.0, + 296.0, + 325.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 320.0, + 1407.0, + 320.0, + 1407.0, + 359.0, + 292.0, + 359.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 352.0, + 1407.0, + 352.0, + 1407.0, + 389.0, + 291.0, + 389.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 383.0, + 1408.0, + 383.0, + 1408.0, + 420.0, + 292.0, + 420.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 412.0, + 1403.0, + 412.0, + 1403.0, + 448.0, + 295.0, + 448.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 444.0, + 1406.0, + 444.0, + 1406.0, + 479.0, + 294.0, + 479.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 476.0, + 1403.0, + 476.0, + 1403.0, + 507.0, + 295.0, + 507.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 505.0, + 1403.0, + 505.0, + 1403.0, + 540.0, + 294.0, + 540.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 535.0, + 544.0, + 535.0, + 544.0, + 572.0, + 294.0, + 572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1468.0, + 1405.0, + 1468.0, + 1405.0, + 1511.0, + 292.0, + 1511.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 287.0, + 1490.0, + 393.0, + 1490.0, + 393.0, + 1549.0, + 287.0, + 1549.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 660.0, + 1490.0, + 712.0, + 1490.0, + 712.0, + 1549.0, + 660.0, + 1549.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 823.0, + 1490.0, + 1411.0, + 1490.0, + 1411.0, + 1549.0, + 823.0, + 1549.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1536.0, + 1405.0, + 1536.0, + 1405.0, + 1572.0, + 294.0, + 1572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1567.0, + 1404.0, + 1567.0, + 1404.0, + 1603.0, + 293.0, + 1603.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1591.0, + 1406.0, + 1591.0, + 1406.0, + 1636.0, + 292.0, + 1636.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1629.0, + 803.0, + 1629.0, + 803.0, + 1661.0, + 296.0, + 1661.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1272.0, + 1405.0, + 1272.0, + 1405.0, + 1311.0, + 294.0, + 1311.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1307.0, + 1406.0, + 1307.0, + 1406.0, + 1339.0, + 296.0, + 1339.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1338.0, + 1405.0, + 1338.0, + 1405.0, + 1370.0, + 296.0, + 1370.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1364.0, + 1406.0, + 1364.0, + 1406.0, + 1403.0, + 294.0, + 1403.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1394.0, + 1405.0, + 1394.0, + 1405.0, + 1433.0, + 292.0, + 1433.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1426.0, + 1240.0, + 1426.0, + 1240.0, + 1464.0, + 294.0, + 1464.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 575.0, + 742.0, + 575.0, + 742.0, + 618.0, + 291.0, + 618.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 807.0, + 575.0, + 1206.0, + 575.0, + 1206.0, + 618.0, + 807.0, + 618.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1401.0, + 575.0, + 1407.0, + 575.0, + 1407.0, + 618.0, + 1401.0, + 618.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 609.0, + 406.0, + 609.0, + 406.0, + 646.0, + 291.0, + 646.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 601.0, + 609.0, + 1406.0, + 609.0, + 1406.0, + 646.0, + 601.0, + 646.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 636.0, + 751.0, + 636.0, + 751.0, + 680.0, + 292.0, + 680.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 937.0, + 636.0, + 986.0, + 636.0, + 986.0, + 680.0, + 937.0, + 680.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1172.0, + 636.0, + 1410.0, + 636.0, + 1410.0, + 680.0, + 1172.0, + 680.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 671.0, + 1055.0, + 671.0, + 1055.0, + 706.0, + 292.0, + 706.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1089.0, + 671.0, + 1221.0, + 671.0, + 1221.0, + 706.0, + 1089.0, + 706.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 792.0, + 372.0, + 792.0, + 372.0, + 824.0, + 297.0, + 824.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 399.0, + 792.0, + 1251.0, + 792.0, + 1251.0, + 824.0, + 399.0, + 824.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1286.0, + 792.0, + 1402.0, + 792.0, + 1402.0, + 824.0, + 1286.0, + 824.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 822.0, + 1134.0, + 822.0, + 1134.0, + 856.0, + 293.0, + 856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1168.0, + 822.0, + 1179.0, + 822.0, + 1179.0, + 856.0, + 1168.0, + 856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 939.0, + 809.0, + 939.0, + 809.0, + 977.0, + 295.0, + 977.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1057.0, + 370.0, + 1057.0, + 370.0, + 1092.0, + 296.0, + 1092.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 404.0, + 1057.0, + 831.0, + 1057.0, + 831.0, + 1092.0, + 404.0, + 1092.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 865.0, + 1057.0, + 875.0, + 1057.0, + 875.0, + 1092.0, + 865.0, + 1092.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 4, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 596, + 1405, + 596, + 1405, + 872, + 297, + 872 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 297, + 1762, + 1405, + 1762, + 1405, + 2008, + 297, + 2008 + ], + "score": 0.981 + }, + { + "category_id": 1, + "poly": [ + 297, + 1474, + 1405, + 1474, + 1405, + 1750, + 297, + 1750 + ], + "score": 0.979 + }, + { + "category_id": 5, + "poly": [ + 297, + 307, + 1403, + 307, + 1403, + 488, + 297, + 488 + ], + "score": 0.977, + "html": "
NameDepthEmbed dim.MLP Ratio#Heads#ParamsThroughput (im/s)Test sizeTop-1 Acc. (%)
LV-ViT-T122403.048.5M2032.622479.1
LV-ViT-S163843.0626M1018.222483.3
LV-ViT-M205123.0856M668.922484.1
LV-ViT-L247683.012150M204.828885.3
" + }, + { + "category_id": 5, + "poly": [ + 296, + 1234, + 803, + 1234, + 803, + 1440, + 296, + 1440 + ], + "score": 0.976, + "html": "
Aug. MethodSupervisionTop-1 Acc.
MixTokenToken labeling83.3
MixTokenReLabel83.0
CutMixReLabel82.8
MixtokenImageNet Label82.5
CutMixImageNet Label82.4
" + }, + { + "category_id": 5, + "poly": [ + 824, + 1233, + 1404, + 1233, + 1404, + 1440, + 824, + 1440 + ], + "score": 0.976, + "html": "
MixTokenMixUpCutOutRandAugTop-1 Acc.
83.3
81.3
>x>xx>>>>>>>83.1
X√83.0
广XXX82.8
" + }, + { + "category_id": 1, + "poly": [ + 297, + 884, + 1405, + 884, + 1405, + 1069, + 297, + 1069 + ], + "score": 0.972 + }, + { + "category_id": 6, + "poly": [ + 296, + 216, + 1404, + 216, + 1404, + 306, + 296, + 306 + ], + "score": 0.957 + }, + { + "category_id": 6, + "poly": [ + 823, + 1111, + 1404, + 1111, + 1404, + 1233, + 823, + 1233 + ], + "score": 0.955 + }, + { + "category_id": 6, + "poly": [ + 295, + 1110, + 800, + 1110, + 800, + 1234, + 295, + 1234 + ], + "score": 0.949 + }, + { + "category_id": 0, + "poly": [ + 298, + 541, + 574, + 541, + 574, + 574, + 298, + 574 + ], + "score": 0.918 + }, + { + "category_id": 2, + "poly": [ + 840, + 2062, + 859, + 2062, + 859, + 2085, + 840, + 2085 + ], + "score": 0.785 + }, + { + "category_id": 13, + "poly": [ + 500, + 1914, + 582, + 1914, + 582, + 1945, + 500, + 1945 + ], + "score": 0.91, + "latex": "- 0 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 420, + 1885, + 474, + 1885, + 474, + 1914, + 420, + 1914 + ], + "score": 0.89, + "latex": "2 0 \\%" + }, + { + "category_id": 13, + "poly": [ + 399, + 1037, + 474, + 1037, + 474, + 1067, + 399, + 1067 + ], + "score": 0.89, + "latex": "8 3 . 3 \\%" + }, + { + "category_id": 13, + "poly": [ + 636, + 1687, + 698, + 1687, + 698, + 1717, + 636, + 1717 + ], + "score": 0.89, + "latex": "\\dot { 0 } . 3 \\%" + }, + { + "category_id": 13, + "poly": [ + 1016, + 1975, + 1084, + 1975, + 1084, + 2005, + 1016, + 2005 + ], + "score": 0.89, + "latex": "\\mathrm { \\bar { 1 0 0 \\% } }" + }, + { + "category_id": 13, + "poly": [ + 1301, + 1536, + 1376, + 1536, + 1376, + 1566, + 1301, + 1566 + ], + "score": 0.88, + "latex": "8 3 . 1 \\%" + }, + { + "category_id": 13, + "poly": [ + 873, + 1717, + 934, + 1717, + 934, + 1748, + 873, + 1748 + ], + "score": 0.88, + "latex": "0 . 5 \\%" + }, + { + "category_id": 13, + "poly": [ + 732, + 277, + 852, + 277, + 852, + 306, + 732, + 306 + ], + "score": 0.88, + "latex": "2 8 8 \\times 2 8 8" + }, + { + "category_id": 13, + "poly": [ + 1311, + 1884, + 1403, + 1884, + 1403, + 1916, + 1311, + 1916 + ], + "score": 0.87, + "latex": "( - 0 . 5 \\%" + }, + { + "category_id": 13, + "poly": [ + 297, + 1006, + 360, + 1006, + 360, + 1037, + 297, + 1037 + ], + "score": 0.87, + "latex": "0 . { \\bar { 2 } } \\%" + }, + { + "category_id": 13, + "poly": [ + 429, + 749, + 503, + 749, + 503, + 779, + 429, + 779 + ], + "score": 0.87, + "latex": "7 9 . 1 \\%" + }, + { + "category_id": 13, + "poly": [ + 328, + 1597, + 404, + 1597, + 404, + 1626, + 328, + 1626 + ], + "score": 0.87, + "latex": "8 3 . 3 \\%" + }, + { + "category_id": 13, + "poly": [ + 1340, + 945, + 1403, + 945, + 1403, + 977, + 1340, + 977 + ], + "score": 0.87, + "latex": "0 . 1 \\%" + }, + { + "category_id": 13, + "poly": [ + 296, + 277, + 418, + 277, + 418, + 306, + 296, + 306 + ], + "score": 0.8, + "latex": "2 2 4 \\times 2 2 4" + }, + { + "category_id": 13, + "poly": [ + 887, + 719, + 953, + 719, + 953, + 749, + 887, + 749 + ], + "score": 0.38, + "latex": "8 . 5 { \\bf M }" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 212.0, + 1405.0, + 212.0, + 1405.0, + 252.0, + 291.0, + 252.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 246.0, + 1405.0, + 246.0, + 1405.0, + 280.0, + 294.0, + 280.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 419.0, + 276.0, + 731.0, + 276.0, + 731.0, + 310.0, + 419.0, + 310.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 853.0, + 276.0, + 863.0, + 276.0, + 863.0, + 310.0, + 853.0, + 310.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 821.0, + 1107.0, + 1407.0, + 1107.0, + 1407.0, + 1147.0, + 821.0, + 1147.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 823.0, + 1141.0, + 1404.0, + 1141.0, + 1404.0, + 1175.0, + 823.0, + 1175.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 825.0, + 1173.0, + 1404.0, + 1173.0, + 1404.0, + 1204.0, + 825.0, + 1204.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 824.0, + 1203.0, + 1339.0, + 1203.0, + 1339.0, + 1236.0, + 824.0, + 1236.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1110.0, + 798.0, + 1110.0, + 798.0, + 1143.0, + 295.0, + 1143.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1142.0, + 797.0, + 1142.0, + 797.0, + 1173.0, + 297.0, + 1173.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1172.0, + 798.0, + 1172.0, + 798.0, + 1203.0, + 297.0, + 1203.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1201.0, + 768.0, + 1201.0, + 768.0, + 1236.0, + 295.0, + 1236.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 539.0, + 577.0, + 539.0, + 577.0, + 578.0, + 294.0, + 578.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 840.0, + 2060.0, + 862.0, + 2060.0, + 862.0, + 2091.0, + 840.0, + 2091.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 594.0, + 1406.0, + 594.0, + 1406.0, + 633.0, + 294.0, + 633.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 628.0, + 1403.0, + 628.0, + 1403.0, + 661.0, + 296.0, + 661.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 658.0, + 1406.0, + 658.0, + 1406.0, + 695.0, + 294.0, + 695.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 684.0, + 1405.0, + 684.0, + 1405.0, + 726.0, + 292.0, + 726.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 717.0, + 886.0, + 717.0, + 886.0, + 758.0, + 292.0, + 758.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 954.0, + 717.0, + 1407.0, + 717.0, + 1407.0, + 758.0, + 954.0, + 758.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 750.0, + 428.0, + 750.0, + 428.0, + 784.0, + 294.0, + 784.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 504.0, + 750.0, + 1407.0, + 750.0, + 1407.0, + 784.0, + 504.0, + 784.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 778.0, + 1406.0, + 778.0, + 1406.0, + 814.0, + 294.0, + 814.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 808.0, + 1407.0, + 808.0, + 1407.0, + 845.0, + 295.0, + 845.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 840.0, + 885.0, + 840.0, + 885.0, + 875.0, + 294.0, + 875.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1760.0, + 1403.0, + 1760.0, + 1403.0, + 1799.0, + 294.0, + 1799.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1794.0, + 1405.0, + 1794.0, + 1405.0, + 1829.0, + 292.0, + 1829.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1824.0, + 1405.0, + 1824.0, + 1405.0, + 1857.0, + 295.0, + 1857.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1855.0, + 1407.0, + 1855.0, + 1407.0, + 1889.0, + 295.0, + 1889.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1885.0, + 419.0, + 1885.0, + 419.0, + 1919.0, + 295.0, + 1919.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 475.0, + 1885.0, + 1310.0, + 1885.0, + 1310.0, + 1919.0, + 475.0, + 1919.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1913.0, + 499.0, + 1913.0, + 499.0, + 1949.0, + 294.0, + 1949.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 583.0, + 1913.0, + 1405.0, + 1913.0, + 1405.0, + 1949.0, + 583.0, + 1949.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1944.0, + 1405.0, + 1944.0, + 1405.0, + 1981.0, + 292.0, + 1981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1976.0, + 1015.0, + 1976.0, + 1015.0, + 2010.0, + 295.0, + 2010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1085.0, + 1976.0, + 1408.0, + 1976.0, + 1408.0, + 2010.0, + 1085.0, + 2010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1472.0, + 1406.0, + 1472.0, + 1406.0, + 1510.0, + 294.0, + 1510.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1502.0, + 1406.0, + 1502.0, + 1406.0, + 1540.0, + 292.0, + 1540.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1534.0, + 1300.0, + 1534.0, + 1300.0, + 1572.0, + 294.0, + 1572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1377.0, + 1534.0, + 1406.0, + 1534.0, + 1406.0, + 1572.0, + 1377.0, + 1572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1564.0, + 1406.0, + 1564.0, + 1406.0, + 1602.0, + 294.0, + 1602.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1594.0, + 327.0, + 1594.0, + 327.0, + 1633.0, + 292.0, + 1633.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 405.0, + 1594.0, + 1405.0, + 1594.0, + 1405.0, + 1633.0, + 405.0, + 1633.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1628.0, + 1405.0, + 1628.0, + 1405.0, + 1660.0, + 295.0, + 1660.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1656.0, + 1407.0, + 1656.0, + 1407.0, + 1693.0, + 294.0, + 1693.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1686.0, + 635.0, + 1686.0, + 635.0, + 1723.0, + 295.0, + 1723.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 699.0, + 1686.0, + 1406.0, + 1686.0, + 1406.0, + 1723.0, + 699.0, + 1723.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1718.0, + 872.0, + 1718.0, + 872.0, + 1752.0, + 294.0, + 1752.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 935.0, + 1718.0, + 944.0, + 1718.0, + 944.0, + 1752.0, + 935.0, + 1752.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 886.0, + 1406.0, + 886.0, + 1406.0, + 918.0, + 295.0, + 918.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 915.0, + 1407.0, + 915.0, + 1407.0, + 948.0, + 294.0, + 948.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 945.0, + 1339.0, + 945.0, + 1339.0, + 980.0, + 292.0, + 980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1404.0, + 945.0, + 1407.0, + 945.0, + 1407.0, + 980.0, + 1404.0, + 980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 974.0, + 1407.0, + 974.0, + 1407.0, + 1011.0, + 292.0, + 1011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1003.0, + 296.0, + 1003.0, + 296.0, + 1043.0, + 292.0, + 1043.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 361.0, + 1003.0, + 1407.0, + 1003.0, + 1407.0, + 1043.0, + 361.0, + 1043.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1038.0, + 398.0, + 1038.0, + 398.0, + 1067.0, + 292.0, + 1067.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 475.0, + 1038.0, + 485.0, + 1038.0, + 485.0, + 1067.0, + 475.0, + 1067.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 5, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1201, + 1405, + 1201, + 1405, + 1627, + 297, + 1627 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 298, + 1642, + 1405, + 1642, + 1405, + 2007, + 298, + 2007 + ], + "score": 0.982 + }, + { + "category_id": 5, + "poly": [ + 357, + 950, + 1343, + 950, + 1343, + 1144, + 357, + 1144 + ], + "score": 0.979, + "html": "
MethodOnline KDOnline TLTLReLabelVanilla
Tokens Utilized2AllAll11
DeiT-S Acc. (%)81.281.881.080.479.9
LV-ViT-S Acc. (%)83.083.583.382.882.4
Training Time (8× V100)63 hrs63 hrs45 hrs45 hrs41 hrs
" + }, + { + "category_id": 4, + "poly": [ + 296, + 620, + 1407, + 620, + 1407, + 802, + 296, + 802 + ], + "score": 0.962 + }, + { + "category_id": 6, + "poly": [ + 300, + 856, + 1403, + 856, + 1403, + 949, + 300, + 949 + ], + "score": 0.95 + }, + { + "category_id": 3, + "poly": [ + 305, + 229, + 1347, + 229, + 1347, + 606, + 305, + 606 + ], + "score": 0.916 + }, + { + "category_id": 2, + "poly": [ + 841, + 2061, + 858, + 2061, + 858, + 2084, + 841, + 2084 + ], + "score": 0.717 + }, + { + "category_id": 2, + "poly": [ + 841, + 2060, + 859, + 2060, + 859, + 2084, + 841, + 2084 + ], + "score": 0.111 + }, + { + "category_id": 13, + "poly": [ + 679, + 1884, + 754, + 1884, + 754, + 1914, + 679, + 1914 + ], + "score": 0.89, + "latex": "8 6 . 3 \\%" + }, + { + "category_id": 13, + "poly": [ + 792, + 1764, + 866, + 1764, + 866, + 1793, + 792, + 1793 + ], + "score": 0.89, + "latex": "8 1 . 6 \\%" + }, + { + "category_id": 13, + "poly": [ + 640, + 1945, + 715, + 1945, + 715, + 1975, + 640, + 1975 + ], + "score": 0.88, + "latex": "8 8 . 2 \\%" + }, + { + "category_id": 13, + "poly": [ + 737, + 1975, + 812, + 1975, + 812, + 2005, + 737, + 2005 + ], + "score": 0.88, + "latex": "8 \\mathrm { { 3 . 5 \\% } }" + }, + { + "category_id": 13, + "poly": [ + 297, + 1914, + 373, + 1914, + 373, + 1945, + 297, + 1945 + ], + "score": 0.87, + "latex": "8 3 . 3 \\%" + }, + { + "category_id": 13, + "poly": [ + 1018, + 1567, + 1079, + 1567, + 1079, + 1596, + 1018, + 1596 + ], + "score": 0.86, + "latex": "0 . 9 \\%" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 617.0, + 1407.0, + 617.0, + 1407.0, + 658.0, + 291.0, + 658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 648.0, + 1407.0, + 648.0, + 1407.0, + 687.0, + 293.0, + 687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 680.0, + 1407.0, + 680.0, + 1407.0, + 715.0, + 294.0, + 715.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 712.0, + 1404.0, + 712.0, + 1404.0, + 744.0, + 295.0, + 744.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 740.0, + 1406.0, + 740.0, + 1406.0, + 776.0, + 291.0, + 776.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 773.0, + 501.0, + 773.0, + 501.0, + 802.0, + 294.0, + 802.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 855.0, + 1405.0, + 855.0, + 1405.0, + 893.0, + 294.0, + 893.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 888.0, + 1405.0, + 888.0, + 1405.0, + 922.0, + 295.0, + 922.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 919.0, + 773.0, + 919.0, + 773.0, + 953.0, + 296.0, + 953.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 235.0, + 373.0, + 235.0, + 373.0, + 263.0, + 326.0, + 263.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 869.0, + 232.0, + 918.0, + 232.0, + 918.0, + 262.0, + 869.0, + 262.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 870.0, + 291.0, + 916.0, + 291.0, + 916.0, + 319.0, + 870.0, + 319.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 849.0, + 293.0, + 883.0, + 293.0, + 883.0, + 505.0, + 849.0, + 505.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 304.0, + 296.0, + 338.0, + 296.0, + 338.0, + 509.0, + 304.0, + 509.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 325.0, + 297.0, + 372.0, + 297.0, + 372.0, + 323.0, + 325.0, + 323.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 870.0, + 348.0, + 915.0, + 348.0, + 915.0, + 374.0, + 870.0, + 374.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 328.0, + 359.0, + 372.0, + 359.0, + 372.0, + 385.0, + 328.0, + 385.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 729.0, + 381.0, + 795.0, + 381.0, + 795.0, + 406.0, + 729.0, + 406.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 953.0, + 381.0, + 970.0, + 381.0, + 970.0, + 398.0, + 953.0, + 398.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1183.0, + 390.0, + 1195.0, + 390.0, + 1195.0, + 400.0, + 1183.0, + 400.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1208.0, + 379.0, + 1333.0, + 379.0, + 1333.0, + 406.0, + 1208.0, + 406.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 562.0, + 406.0, + 625.0, + 406.0, + 625.0, + 431.0, + 562.0, + 431.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 662.0, + 406.0, + 726.0, + 406.0, + 726.0, + 431.0, + 662.0, + 431.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 871.0, + 405.0, + 916.0, + 405.0, + 916.0, + 431.0, + 871.0, + 431.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1182.0, + 410.0, + 1194.0, + 410.0, + 1194.0, + 421.0, + 1182.0, + 421.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1207.0, + 400.0, + 1333.0, + 400.0, + 1333.0, + 427.0, + 1207.0, + 427.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 328.0, + 422.0, + 372.0, + 422.0, + 372.0, + 449.0, + 328.0, + 449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1180.0, + 429.0, + 1196.0, + 429.0, + 1196.0, + 444.0, + 1180.0, + 444.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1207.0, + 420.0, + 1333.0, + 420.0, + 1333.0, + 447.0, + 1207.0, + 447.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 461.0, + 443.0, + 527.0, + 443.0, + 527.0, + 468.0, + 461.0, + 468.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1180.0, + 449.0, + 1195.0, + 449.0, + 1195.0, + 465.0, + 1180.0, + 465.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1207.0, + 441.0, + 1333.0, + 441.0, + 1333.0, + 468.0, + 1207.0, + 468.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 871.0, + 461.0, + 918.0, + 461.0, + 918.0, + 490.0, + 871.0, + 490.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1180.0, + 470.0, + 1195.0, + 470.0, + 1195.0, + 486.0, + 1180.0, + 486.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1207.0, + 462.0, + 1332.0, + 462.0, + 1332.0, + 490.0, + 1207.0, + 490.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 483.0, + 372.0, + 483.0, + 372.0, + 509.0, + 326.0, + 509.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 395.0, + 493.0, + 461.0, + 493.0, + 461.0, + 518.0, + 395.0, + 518.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 694.0, + 481.0, + 758.0, + 481.0, + 758.0, + 511.0, + 694.0, + 511.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1182.0, + 493.0, + 1194.0, + 493.0, + 1194.0, + 504.0, + 1182.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1207.0, + 483.0, + 1333.0, + 483.0, + 1333.0, + 511.0, + 1207.0, + 511.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 677.0, + 534.0, + 692.0, + 534.0, + 692.0, + 549.0, + 677.0, + 549.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 701.0, + 503.0, + 786.0, + 503.0, + 786.0, + 553.0, + 701.0, + 553.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 869.0, + 516.0, + 918.0, + 516.0, + 918.0, + 545.0, + 869.0, + 545.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 974.0, + 516.0, + 1108.0, + 516.0, + 1108.0, + 543.0, + 974.0, + 543.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1180.0, + 514.0, + 1196.0, + 514.0, + 1196.0, + 547.0, + 1180.0, + 547.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1207.0, + 506.0, + 1320.0, + 506.0, + 1320.0, + 550.0, + 1207.0, + 550.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 545.0, + 373.0, + 545.0, + 373.0, + 574.0, + 326.0, + 574.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 398.0, + 563.0, + 422.0, + 563.0, + 422.0, + 589.0, + 398.0, + 589.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 460.0, + 563.0, + 493.0, + 563.0, + 493.0, + 590.0, + 460.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 527.0, + 564.0, + 558.0, + 564.0, + 558.0, + 588.0, + 527.0, + 588.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 595.0, + 564.0, + 624.0, + 564.0, + 624.0, + 588.0, + 595.0, + 588.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 661.0, + 562.0, + 694.0, + 562.0, + 694.0, + 590.0, + 661.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 724.0, + 562.0, + 764.0, + 562.0, + 764.0, + 589.0, + 724.0, + 589.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 903.0, + 562.0, + 935.0, + 562.0, + 935.0, + 590.0, + 903.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 972.0, + 562.0, + 1006.0, + 562.0, + 1006.0, + 590.0, + 972.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1043.0, + 562.0, + 1076.0, + 562.0, + 1076.0, + 590.0, + 1043.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1113.0, + 563.0, + 1146.0, + 563.0, + 1146.0, + 590.0, + 1113.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1183.0, + 563.0, + 1213.0, + 563.0, + 1213.0, + 588.0, + 1183.0, + 588.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1252.0, + 562.0, + 1285.0, + 562.0, + 1285.0, + 590.0, + 1252.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1321.0, + 562.0, + 1353.0, + 562.0, + 1353.0, + 590.0, + 1321.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 482.0, + 581.0, + 690.0, + 581.0, + 690.0, + 609.0, + 482.0, + 609.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1023.0, + 581.0, + 1234.0, + 581.0, + 1234.0, + 609.0, + 1023.0, + 609.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 722.25, + 279.0, + 796.25, + 279.0, + 796.25, + 311.5, + 722.25, + 311.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 662.0, + 291.0, + 737.0, + 291.0, + 737.0, + 321.0, + 662.0, + 321.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 561.0, + 301.5, + 631.0, + 301.5, + 631.0, + 335.5, + 561.0, + 335.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 460.0, + 327.5, + 531.0, + 327.5, + 531.0, + 361.0, + 460.0, + 361.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 426.5, + 465.0, + 426.5, + 465.0, + 460.5, + 393.0, + 460.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2059.0, + 860.0, + 2059.0, + 860.0, + 2091.0, + 839.0, + 2091.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2059.0, + 860.0, + 2059.0, + 860.0, + 2091.0, + 839.0, + 2091.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1199.0, + 1405.0, + 1199.0, + 1405.0, + 1238.0, + 294.0, + 1238.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1232.0, + 1405.0, + 1232.0, + 1405.0, + 1267.0, + 294.0, + 1267.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1264.0, + 1405.0, + 1264.0, + 1405.0, + 1299.0, + 295.0, + 1299.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1291.0, + 1406.0, + 1291.0, + 1406.0, + 1329.0, + 294.0, + 1329.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1322.0, + 1406.0, + 1322.0, + 1406.0, + 1360.0, + 292.0, + 1360.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1351.0, + 1405.0, + 1351.0, + 1405.0, + 1393.0, + 294.0, + 1393.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1384.0, + 1406.0, + 1384.0, + 1406.0, + 1419.0, + 295.0, + 1419.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1413.0, + 1405.0, + 1413.0, + 1405.0, + 1451.0, + 294.0, + 1451.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1443.0, + 1407.0, + 1443.0, + 1407.0, + 1480.0, + 294.0, + 1480.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1475.0, + 1407.0, + 1475.0, + 1407.0, + 1511.0, + 294.0, + 1511.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1504.0, + 1407.0, + 1504.0, + 1407.0, + 1542.0, + 292.0, + 1542.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1537.0, + 1405.0, + 1537.0, + 1405.0, + 1572.0, + 295.0, + 1572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1568.0, + 1017.0, + 1568.0, + 1017.0, + 1600.0, + 296.0, + 1600.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1080.0, + 1568.0, + 1403.0, + 1568.0, + 1403.0, + 1600.0, + 1080.0, + 1600.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1597.0, + 614.0, + 1597.0, + 614.0, + 1632.0, + 295.0, + 1632.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1643.0, + 1403.0, + 1643.0, + 1403.0, + 1677.0, + 296.0, + 1677.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1674.0, + 1405.0, + 1674.0, + 1405.0, + 1705.0, + 296.0, + 1705.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1701.0, + 1405.0, + 1701.0, + 1405.0, + 1737.0, + 293.0, + 1737.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1732.0, + 1406.0, + 1732.0, + 1406.0, + 1768.0, + 292.0, + 1768.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1764.0, + 791.0, + 1764.0, + 791.0, + 1798.0, + 295.0, + 1798.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 867.0, + 1764.0, + 1405.0, + 1764.0, + 1405.0, + 1798.0, + 867.0, + 1798.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1795.0, + 1405.0, + 1795.0, + 1405.0, + 1827.0, + 293.0, + 1827.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1822.0, + 1406.0, + 1822.0, + 1406.0, + 1860.0, + 293.0, + 1860.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1852.0, + 1408.0, + 1852.0, + 1408.0, + 1891.0, + 292.0, + 1891.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1884.0, + 678.0, + 1884.0, + 678.0, + 1918.0, + 295.0, + 1918.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 755.0, + 1884.0, + 1406.0, + 1884.0, + 1406.0, + 1918.0, + 755.0, + 1918.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 374.0, + 1914.0, + 1406.0, + 1914.0, + 1406.0, + 1948.0, + 374.0, + 1948.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1944.0, + 639.0, + 1944.0, + 639.0, + 1980.0, + 293.0, + 1980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 716.0, + 1944.0, + 1405.0, + 1944.0, + 1405.0, + 1980.0, + 716.0, + 1980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1976.0, + 736.0, + 1976.0, + 736.0, + 2010.0, + 295.0, + 2010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 813.0, + 1976.0, + 1405.0, + 1976.0, + 1405.0, + 2010.0, + 813.0, + 2010.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 6, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1701, + 1406, + 1701, + 1406, + 2007, + 297, + 2007 + ], + "score": 0.984 + }, + { + "category_id": 1, + "poly": [ + 297, + 743, + 1405, + 743, + 1405, + 989, + 297, + 989 + ], + "score": 0.981 + }, + { + "category_id": 5, + "poly": [ + 309, + 1350, + 1388, + 1350, + 1388, + 1515, + 309, + 1515 + ], + "score": 0.978, + "html": "
ModelMixer-S/16 [35]Mixer-B/16 [35]Mixer-L/16 [35]ResNeSt-50 [51]
Token LabelingXX√×X√××√××√
Parameters18M18M18M59M59M 59M207M207M207M27M27M 27M
Top-1 Acc. (%)73.8t75.676.176.478.379.571.6t77.780.181.1t80.981.5
" + }, + { + "category_id": 1, + "poly": [ + 297, + 1002, + 1404, + 1002, + 1404, + 1185, + 297, + 1185 + ], + "score": 0.977 + }, + { + "category_id": 6, + "poly": [ + 297, + 1258, + 1404, + 1258, + 1404, + 1350, + 297, + 1350 + ], + "score": 0.969 + }, + { + "category_id": 3, + "poly": [ + 308, + 317, + 1404, + 317, + 1404, + 574, + 308, + 574 + ], + "score": 0.966 + }, + { + "category_id": 4, + "poly": [ + 298, + 595, + 1407, + 595, + 1407, + 687, + 298, + 687 + ], + "score": 0.961 + }, + { + "category_id": 0, + "poly": [ + 299, + 1636, + 726, + 1636, + 726, + 1669, + 299, + 1669 + ], + "score": 0.917 + }, + { + "category_id": 2, + "poly": [ + 840, + 2062, + 858, + 2062, + 858, + 2085, + 840, + 2085 + ], + "score": 0.807 + }, + { + "category_id": 1, + "poly": [ + 298, + 203, + 1403, + 203, + 1403, + 265, + 298, + 265 + ], + "score": 0.685 + }, + { + "category_id": 0, + "poly": [ + 298, + 203, + 1403, + 203, + 1403, + 265, + 298, + 265 + ], + "score": 0.11 + }, + { + "category_id": 13, + "poly": [ + 799, + 1734, + 911, + 1734, + 911, + 1763, + 799, + 1763 + ], + "score": 0.9, + "latex": "2 2 4 \\times 2 2 4" + }, + { + "category_id": 13, + "poly": [ + 1322, + 1793, + 1397, + 1793, + 1397, + 1824, + 1322, + 1824 + ], + "score": 0.9, + "latex": "8 5 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 836, + 1794, + 959, + 1794, + 959, + 1823, + 836, + 1823 + ], + "score": 0.9, + "latex": "3 8 4 \\times 3 8 4" + }, + { + "category_id": 13, + "poly": [ + 584, + 1946, + 710, + 1946, + 710, + 1975, + 584, + 1975 + ], + "score": 0.89, + "latex": "4 4 8 \\times 4 4 8" + }, + { + "category_id": 13, + "poly": [ + 701, + 1763, + 763, + 1763, + 763, + 1793, + 701, + 1793 + ], + "score": 0.89, + "latex": "3 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 1078, + 1733, + 1154, + 1733, + 1154, + 1763, + 1078, + 1763 + ], + "score": 0.88, + "latex": "8 3 . 3 \\%" + }, + { + "category_id": 13, + "poly": [ + 858, + 1946, + 935, + 1946, + 935, + 1975, + 858, + 1975 + ], + "score": 0.87, + "latex": "8 6 . 2 \\%" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1255.0, + 1405.0, + 1255.0, + 1405.0, + 1294.0, + 291.0, + 1294.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1290.0, + 1404.0, + 1290.0, + 1404.0, + 1323.0, + 295.0, + 1323.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1315.0, + 1064.0, + 1315.0, + 1064.0, + 1357.0, + 294.0, + 1357.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 318.0, + 476.0, + 318.0, + 476.0, + 543.0, + 293.0, + 543.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 562.0, + 354.0, + 628.0, + 354.0, + 628.0, + 384.0, + 562.0, + 384.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 665.0, + 318.0, + 850.0, + 318.0, + 850.0, + 545.0, + 665.0, + 545.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 936.0, + 350.0, + 1003.0, + 350.0, + 1003.0, + 380.0, + 936.0, + 380.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1043.0, + 317.0, + 1224.0, + 317.0, + 1224.0, + 545.0, + 1043.0, + 545.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 401.0, + 366.0, + 462.0, + 366.0, + 462.0, + 396.0, + 401.0, + 396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 775.0, + 366.0, + 836.0, + 366.0, + 836.0, + 396.0, + 775.0, + 396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1149.0, + 366.0, + 1210.0, + 366.0, + 1210.0, + 396.0, + 1149.0, + 396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 405.0, + 354.0, + 405.0, + 354.0, + 432.0, + 327.0, + 432.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 562.0, + 407.0, + 631.0, + 407.0, + 631.0, + 444.0, + 562.0, + 444.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 698.0, + 403.0, + 729.0, + 403.0, + 729.0, + 433.0, + 698.0, + 433.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 789.0, + 422.0, + 858.0, + 422.0, + 858.0, + 455.0, + 789.0, + 455.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 936.0, + 388.0, + 1004.0, + 388.0, + 1004.0, + 422.0, + 936.0, + 422.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1074.0, + 405.0, + 1101.0, + 405.0, + 1101.0, + 432.0, + 1074.0, + 432.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1235.0, + 403.0, + 1300.0, + 403.0, + 1300.0, + 438.0, + 1235.0, + 438.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1333.0, + 383.0, + 1400.0, + 383.0, + 1400.0, + 423.0, + 1333.0, + 423.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 325.0, + 450.0, + 353.0, + 450.0, + 353.0, + 474.0, + 325.0, + 474.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 415.0, + 444.0, + 483.0, + 444.0, + 483.0, + 478.0, + 415.0, + 478.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 791.0, + 450.0, + 856.0, + 450.0, + 856.0, + 480.0, + 791.0, + 480.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1139.0, + 436.0, + 1202.0, + 436.0, + 1202.0, + 469.0, + 1139.0, + 469.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 493.0, + 354.0, + 493.0, + 354.0, + 519.0, + 327.0, + 519.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 414.0, + 483.0, + 482.0, + 483.0, + 482.0, + 516.0, + 414.0, + 516.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1074.0, + 493.0, + 1101.0, + 493.0, + 1101.0, + 520.0, + 1074.0, + 520.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1140.0, + 475.0, + 1202.0, + 475.0, + 1202.0, + 509.0, + 1140.0, + 509.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 536.0, + 354.0, + 536.0, + 354.0, + 562.0, + 327.0, + 562.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 381.0, + 548.0, + 477.0, + 548.0, + 477.0, + 577.0, + 381.0, + 577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 530.0, + 550.0, + 624.0, + 550.0, + 624.0, + 574.0, + 530.0, + 574.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 698.0, + 535.0, + 730.0, + 535.0, + 730.0, + 563.0, + 698.0, + 563.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 753.0, + 546.0, + 854.0, + 546.0, + 854.0, + 577.0, + 753.0, + 577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 902.0, + 550.0, + 999.0, + 550.0, + 999.0, + 574.0, + 902.0, + 574.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1074.0, + 537.0, + 1101.0, + 537.0, + 1101.0, + 562.0, + 1074.0, + 562.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1104.0, + 552.0, + 1405.0, + 552.0, + 1405.0, + 572.0, + 1104.0, + 572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1333.0, + 346.5, + 1397.0, + 346.5, + 1397.0, + 378.5, + 1333.0, + 378.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1236.0, + 451.5, + 1298.0, + 451.5, + 1298.0, + 485.5, + 1236.0, + 485.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 593.0, + 1410.0, + 593.0, + 1410.0, + 630.0, + 293.0, + 630.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 625.0, + 1405.0, + 625.0, + 1405.0, + 659.0, + 295.0, + 659.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 657.0, + 787.0, + 657.0, + 787.0, + 688.0, + 293.0, + 688.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1635.0, + 730.0, + 1635.0, + 730.0, + 1674.0, + 293.0, + 1674.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2058.0, + 862.0, + 2058.0, + 862.0, + 2090.0, + 838.0, + 2090.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 201.0, + 1407.0, + 201.0, + 1407.0, + 240.0, + 293.0, + 240.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 234.0, + 677.0, + 234.0, + 677.0, + 268.0, + 294.0, + 268.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1702.0, + 1407.0, + 1702.0, + 1407.0, + 1733.0, + 295.0, + 1733.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1731.0, + 798.0, + 1731.0, + 798.0, + 1768.0, + 292.0, + 1768.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 912.0, + 1731.0, + 1077.0, + 1731.0, + 1077.0, + 1768.0, + 912.0, + 1768.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1155.0, + 1731.0, + 1406.0, + 1731.0, + 1406.0, + 1768.0, + 1155.0, + 1768.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1761.0, + 700.0, + 1761.0, + 700.0, + 1798.0, + 294.0, + 1798.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 764.0, + 1761.0, + 1408.0, + 1761.0, + 1408.0, + 1798.0, + 764.0, + 1798.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1794.0, + 835.0, + 1794.0, + 835.0, + 1826.0, + 295.0, + 1826.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 960.0, + 1794.0, + 1321.0, + 1794.0, + 1321.0, + 1826.0, + 960.0, + 1826.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1398.0, + 1794.0, + 1406.0, + 1794.0, + 1406.0, + 1826.0, + 1398.0, + 1826.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1825.0, + 1404.0, + 1825.0, + 1404.0, + 1857.0, + 295.0, + 1857.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1854.0, + 1406.0, + 1854.0, + 1406.0, + 1890.0, + 294.0, + 1890.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1887.0, + 1403.0, + 1887.0, + 1403.0, + 1919.0, + 295.0, + 1919.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1915.0, + 1404.0, + 1915.0, + 1404.0, + 1947.0, + 294.0, + 1947.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1945.0, + 583.0, + 1945.0, + 583.0, + 1981.0, + 294.0, + 1981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 711.0, + 1945.0, + 857.0, + 1945.0, + 857.0, + 1981.0, + 711.0, + 1981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 936.0, + 1945.0, + 1406.0, + 1945.0, + 1406.0, + 1981.0, + 936.0, + 1981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1974.0, + 943.0, + 1974.0, + 943.0, + 2009.0, + 295.0, + 2009.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 742.0, + 1405.0, + 742.0, + 1405.0, + 778.0, + 294.0, + 778.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 773.0, + 1406.0, + 773.0, + 1406.0, + 807.0, + 294.0, + 807.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 806.0, + 1405.0, + 806.0, + 1405.0, + 840.0, + 295.0, + 840.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 834.0, + 1405.0, + 834.0, + 1405.0, + 871.0, + 292.0, + 871.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 865.0, + 1408.0, + 865.0, + 1408.0, + 900.0, + 294.0, + 900.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 895.0, + 1405.0, + 895.0, + 1405.0, + 932.0, + 292.0, + 932.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 924.0, + 1406.0, + 924.0, + 1406.0, + 963.0, + 292.0, + 963.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 957.0, + 1271.0, + 957.0, + 1271.0, + 991.0, + 294.0, + 991.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1000.0, + 1406.0, + 1000.0, + 1406.0, + 1037.0, + 292.0, + 1037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1029.0, + 1405.0, + 1029.0, + 1405.0, + 1068.0, + 292.0, + 1068.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1064.0, + 1405.0, + 1064.0, + 1405.0, + 1096.0, + 295.0, + 1096.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1091.0, + 1405.0, + 1091.0, + 1405.0, + 1131.0, + 291.0, + 1131.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1124.0, + 1404.0, + 1124.0, + 1404.0, + 1159.0, + 295.0, + 1159.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1155.0, + 507.0, + 1155.0, + 507.0, + 1188.0, + 295.0, + 1188.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 201.0, + 1407.0, + 201.0, + 1407.0, + 240.0, + 293.0, + 240.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 234.0, + 677.0, + 234.0, + 677.0, + 268.0, + 294.0, + 268.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 7, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1595, + 1403, + 1595, + 1403, + 1901, + 298, + 1901 + ], + "score": 0.984 + }, + { + "category_id": 5, + "poly": [ + 299, + 333, + 1404, + 333, + 1404, + 1331, + 299, + 1331 + ], + "score": 0.983, + "html": "
NetworkParamsFLOPsTrain sizeTest sizeTop-1(%)Real Top-1 (%)
EfficientNet-B5 [34] SNNO30M9.9B45645683.688.3
EfficientNet-B7 [34]66M37.0B60060084.3
Fix-EfficientNet-B8 [34,38]87M89.5B67280085.790.0
NFNet-F3 [3]255M114.8B32041685.789.4
NFNet-F4 [3]316M215.3B38451285.989.4
NFNet-F5[3]377M289.8B41654486.089.2
ViT-B/16 [15]86M55.4B22438477.983.6
ViT-L/16[15]307M190.7B22438476.582.2
T2T-ViT-14 [46]22M5.2B22422481.5
T2T-ViT-14↑384 [46]22M17.1B22438483.31
Cross ViT [6]45M56.6B22448084.1一
Swin-B[25]88M47.0B22438484.2
TNT-B[16]66M14.1B22422482.8
iriirrrrrs DeepViT-S [59]27M6.2B22422482.3
DeepViT-L [59]55M12.5B22422483.1
DeiT-S[36]22M4.6B22422479.985.7
Distilled DeiT-S [36]22M4.6B22422481.286.8
DeiT-B [36]86M17.5B22422481.886.7
DeiT-B↑384 [36]86M55.4B22438483.187.7
Distilled DeiT-B [36]87M17.5B22422483.488.3
BoTNet-S1-128 [31]79.1M19.3B25625684.2
BoTNet-S1-128↑384 [31]79.1M45.8B25638484.7-
CaiT-S36↑384 [37]68M48.0B22438485.4- 89.8
CaiT-M36[37]271M53.7B22422485.189.3
CaiT-M36↑448 [37]271M247.8B22444886.390.2
LV-ViT-S26M6.6B22422483.3
江 LV-ViT-S↑38426M22.2B22438484.488.1 88.9
LV-ViT-M56M16.0B22422484.188.4
W LV-ViT-M↑38456M42.2B22438485.489.5
LV-ViT-L150M59.0B28828885.389.3
0 LV-ViT-L↑448150M157.2B28844885.989.7
LV-ViT-L↑448150M157.2B44844886.289.9
LV-ViT-L↑512151M214.8B44851286.490.1
" + }, + { + "category_id": 1, + "poly": [ + 299, + 1460, + 1403, + 1460, + 1403, + 1582, + 299, + 1582 + ], + "score": 0.979 + }, + { + "category_id": 1, + "poly": [ + 300, + 1914, + 1403, + 1914, + 1403, + 2007, + 300, + 2007 + ], + "score": 0.967 + }, + { + "category_id": 6, + "poly": [ + 295, + 216, + 1406, + 216, + 1406, + 337, + 295, + 337 + ], + "score": 0.959 + }, + { + "category_id": 0, + "poly": [ + 299, + 1398, + 790, + 1398, + 790, + 1431, + 299, + 1431 + ], + "score": 0.915 + }, + { + "category_id": 2, + "poly": [ + 840, + 2061, + 858, + 2061, + 858, + 2084, + 840, + 2084 + ], + "score": 0.797 + }, + { + "category_id": 13, + "poly": [ + 625, + 1809, + 746, + 1809, + 746, + 1838, + 625, + 1838 + ], + "score": 0.9, + "latex": "5 1 2 \\times 5 1 2" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 213.0, + 1409.0, + 213.0, + 1409.0, + 256.0, + 292.0, + 256.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 245.0, + 1407.0, + 245.0, + 1407.0, + 282.0, + 294.0, + 282.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 274.0, + 1406.0, + 274.0, + 1406.0, + 314.0, + 293.0, + 314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 304.0, + 864.0, + 304.0, + 864.0, + 343.0, + 293.0, + 343.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1396.0, + 796.0, + 1396.0, + 796.0, + 1438.0, + 294.0, + 1438.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2060.0, + 860.0, + 2060.0, + 860.0, + 2090.0, + 839.0, + 2090.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1596.0, + 1408.0, + 1596.0, + 1408.0, + 1631.0, + 294.0, + 1631.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1626.0, + 1407.0, + 1626.0, + 1407.0, + 1662.0, + 294.0, + 1662.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1654.0, + 1407.0, + 1654.0, + 1407.0, + 1695.0, + 292.0, + 1695.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1686.0, + 1407.0, + 1686.0, + 1407.0, + 1724.0, + 294.0, + 1724.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1718.0, + 1407.0, + 1718.0, + 1407.0, + 1753.0, + 294.0, + 1753.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1748.0, + 1407.0, + 1748.0, + 1407.0, + 1784.0, + 293.0, + 1784.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1776.0, + 1408.0, + 1776.0, + 1408.0, + 1813.0, + 293.0, + 1813.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1806.0, + 624.0, + 1806.0, + 624.0, + 1843.0, + 293.0, + 1843.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 747.0, + 1806.0, + 1407.0, + 1806.0, + 1407.0, + 1843.0, + 747.0, + 1843.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1839.0, + 1405.0, + 1839.0, + 1405.0, + 1875.0, + 294.0, + 1875.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1870.0, + 1145.0, + 1870.0, + 1145.0, + 1905.0, + 294.0, + 1905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1461.0, + 1405.0, + 1461.0, + 1405.0, + 1493.0, + 293.0, + 1493.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1492.0, + 1405.0, + 1492.0, + 1405.0, + 1524.0, + 295.0, + 1524.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1521.0, + 1407.0, + 1521.0, + 1407.0, + 1557.0, + 294.0, + 1557.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1551.0, + 1116.0, + 1551.0, + 1116.0, + 1586.0, + 293.0, + 1586.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1913.0, + 1405.0, + 1913.0, + 1405.0, + 1951.0, + 294.0, + 1951.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1944.0, + 1403.0, + 1944.0, + 1403.0, + 1981.0, + 294.0, + 1981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1975.0, + 1404.0, + 1975.0, + 1404.0, + 2011.0, + 293.0, + 2011.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 8, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 5, + "poly": [ + 297, + 1001, + 1404, + 1001, + 1404, + 1490, + 297, + 1490 + ], + "score": 0.984, + "html": "
BackboneSegmentation ArchitectureModel SizemIoU (MS)Pixel Acc. (MS)
SNNOResNet-269PSPNet [54]44.981.7
ResNet-101UperNet [44]86M44.9-
ResNet-101Strip Pooling [22]45.682.1
ResNeSt200DeepLabV3+ [8]88M48.41
DeiT-SUperNet52M44.0-
TirriiirrssViT-LargetSETR [56]308M50.383.5
Swin-T[25]UperNet60M46.11
Swin-S [25]UperNet81M49.3=
Swin-B [25]UperNet121M49.7
Swin-B† [25]UperNet121M51.6-
LIA-ATLV-ViT-SFCN30M48.483.0
LV-ViT-SUperNet44M48.683.1
LV-ViT-MUperNet77M50.683.5
LV-ViT-LUperNet209M51.884.1
" + }, + { + "category_id": 1, + "poly": [ + 297, + 640, + 1405, + 640, + 1405, + 854, + 297, + 854 + ], + "score": 0.982 + }, + { + "category_id": 1, + "poly": [ + 298, + 1633, + 1403, + 1633, + 1403, + 1787, + 298, + 1787 + ], + "score": 0.979 + }, + { + "category_id": 1, + "poly": [ + 298, + 1800, + 1404, + 1800, + 1404, + 1984, + 298, + 1984 + ], + "score": 0.979 + }, + { + "category_id": 5, + "poly": [ + 304, + 307, + 1395, + 307, + 1395, + 503, + 304, + 503 + ], + "score": 0.977, + "html": "
MethodToken LabelingModel SizemIoU (SS)P. Acc. (SS)mIoU (MS)P. Acc. (MS)
LV-ViT-S + FCN×30M46.181.947.382.6
LV-ViT-S + FCN30M47.282.448.483.0
LV-ViT-S + UperNet44M46.582.147.682.7
LV-ViT-S + UperNetX44M47.982.648.683.1
" + }, + { + "category_id": 1, + "poly": [ + 298, + 564, + 1399, + 564, + 1399, + 625, + 298, + 625 + ], + "score": 0.955 + }, + { + "category_id": 6, + "poly": [ + 297, + 906, + 1406, + 906, + 1406, + 998, + 297, + 998 + ], + "score": 0.955 + }, + { + "category_id": 6, + "poly": [ + 297, + 216, + 1404, + 216, + 1404, + 306, + 297, + 306 + ], + "score": 0.955 + }, + { + "category_id": 0, + "poly": [ + 299, + 1560, + 742, + 1560, + 742, + 1600, + 299, + 1600 + ], + "score": 0.915 + }, + { + "category_id": 2, + "poly": [ + 836, + 2061, + 865, + 2061, + 865, + 2085, + 836, + 2085 + ], + "score": 0.847 + }, + { + "category_id": 13, + "poly": [ + 736, + 1726, + 811, + 1726, + 811, + 1755, + 736, + 1755 + ], + "score": 0.86, + "latex": "8 4 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 1330, + 1726, + 1403, + 1726, + 1403, + 1755, + 1330, + 1755 + ], + "score": 0.84, + "latex": "8 6 . 4 \\%" + }, + { + "category_id": 13, + "poly": [ + 373, + 941, + 396, + 941, + 396, + 965, + 373, + 965 + ], + "score": 0.62, + "latex": "^ +" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 906.0, + 1408.0, + 906.0, + 1408.0, + 940.0, + 294.0, + 940.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 938.0, + 372.0, + 938.0, + 372.0, + 971.0, + 295.0, + 971.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 397.0, + 938.0, + 1404.0, + 938.0, + 1404.0, + 971.0, + 397.0, + 971.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 968.0, + 781.0, + 968.0, + 781.0, + 1001.0, + 292.0, + 1001.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 216.0, + 1404.0, + 216.0, + 1404.0, + 249.0, + 294.0, + 249.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 247.0, + 1404.0, + 247.0, + 1404.0, + 280.0, + 295.0, + 280.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 276.0, + 809.0, + 276.0, + 809.0, + 310.0, + 295.0, + 310.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1556.0, + 745.0, + 1556.0, + 745.0, + 1604.0, + 291.0, + 1604.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2057.0, + 871.0, + 2057.0, + 871.0, + 2096.0, + 832.0, + 2096.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 640.0, + 1405.0, + 640.0, + 1405.0, + 675.0, + 295.0, + 675.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 669.0, + 1406.0, + 669.0, + 1406.0, + 707.0, + 291.0, + 707.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 698.0, + 1410.0, + 698.0, + 1410.0, + 737.0, + 292.0, + 737.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 729.0, + 1410.0, + 729.0, + 1410.0, + 768.0, + 292.0, + 768.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 760.0, + 1405.0, + 760.0, + 1405.0, + 796.0, + 294.0, + 796.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 792.0, + 1405.0, + 792.0, + 1405.0, + 827.0, + 294.0, + 827.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 824.0, + 524.0, + 824.0, + 524.0, + 856.0, + 296.0, + 856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1632.0, + 1405.0, + 1632.0, + 1405.0, + 1670.0, + 293.0, + 1670.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1662.0, + 1405.0, + 1662.0, + 1405.0, + 1702.0, + 293.0, + 1702.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1694.0, + 1404.0, + 1694.0, + 1404.0, + 1730.0, + 293.0, + 1730.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1725.0, + 735.0, + 1725.0, + 735.0, + 1762.0, + 293.0, + 1762.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 812.0, + 1725.0, + 1329.0, + 1725.0, + 1329.0, + 1762.0, + 812.0, + 1762.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1404.0, + 1725.0, + 1408.0, + 1725.0, + 1408.0, + 1762.0, + 1404.0, + 1762.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1755.0, + 1060.0, + 1755.0, + 1060.0, + 1792.0, + 294.0, + 1792.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1801.0, + 1405.0, + 1801.0, + 1405.0, + 1836.0, + 294.0, + 1836.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1830.0, + 1406.0, + 1830.0, + 1406.0, + 1867.0, + 293.0, + 1867.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1863.0, + 1404.0, + 1863.0, + 1404.0, + 1895.0, + 294.0, + 1895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1891.0, + 1406.0, + 1891.0, + 1406.0, + 1928.0, + 293.0, + 1928.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1922.0, + 1406.0, + 1922.0, + 1406.0, + 1958.0, + 294.0, + 1958.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1953.0, + 525.0, + 1953.0, + 525.0, + 1987.0, + 293.0, + 1987.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 559.0, + 1404.0, + 559.0, + 1404.0, + 598.0, + 294.0, + 598.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 596.0, + 705.0, + 596.0, + 705.0, + 627.0, + 297.0, + 627.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 9, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 0, + "poly": [ + 299, + 200, + 455, + 200, + 455, + 235, + 299, + 235 + ], + "score": 0.899 + }, + { + "category_id": 1, + "poly": [ + 293, + 193, + 1410, + 193, + 1410, + 2024, + 293, + 2024 + ], + "score": 0.861 + }, + { + "category_id": 2, + "poly": [ + 836, + 2061, + 863, + 2061, + 863, + 2086, + 836, + 2086 + ], + "score": 0.78 + }, + { + "category_id": 2, + "poly": [ + 836, + 2061, + 863, + 2061, + 863, + 2086, + 836, + 2086 + ], + "score": 0.172 + }, + { + "category_id": 15, + "poly": [ + 295.0, + 196.0, + 460.0, + 196.0, + 460.0, + 241.0, + 295.0, + 241.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2058.0, + 868.0, + 2058.0, + 868.0, + 2098.0, + 832.0, + 2098.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2058.0, + 868.0, + 2058.0, + 868.0, + 2098.0, + 832.0, + 2098.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 195.0, + 462.0, + 195.0, + 462.0, + 242.0, + 294.0, + 242.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 303.0, + 248.0, + 1406.0, + 248.0, + 1406.0, + 286.0, + 303.0, + 286.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 278.0, + 613.0, + 278.0, + 613.0, + 310.0, + 349.0, + 310.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 303.0, + 324.0, + 1408.0, + 324.0, + 1408.0, + 362.0, + 303.0, + 362.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 354.0, + 976.0, + 354.0, + 976.0, + 386.0, + 351.0, + 386.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 299.0, + 396.0, + 1408.0, + 396.0, + 1408.0, + 443.0, + 299.0, + 443.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 345.0, + 429.0, + 1116.0, + 429.0, + 1116.0, + 465.0, + 345.0, + 465.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 303.0, + 477.0, + 1408.0, + 477.0, + 1408.0, + 515.0, + 303.0, + 515.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 503.0, + 1410.0, + 503.0, + 1410.0, + 543.0, + 347.0, + 543.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 535.0, + 761.0, + 535.0, + 761.0, + 567.0, + 351.0, + 567.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 301.0, + 579.0, + 1406.0, + 579.0, + 1406.0, + 620.0, + 301.0, + 620.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 610.0, + 1358.0, + 610.0, + 1358.0, + 646.0, + 347.0, + 646.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 660.0, + 1404.0, + 660.0, + 1404.0, + 692.0, + 307.0, + 692.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 688.0, + 1006.0, + 688.0, + 1006.0, + 720.0, + 349.0, + 720.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 303.0, + 734.0, + 1406.0, + 734.0, + 1406.0, + 773.0, + 303.0, + 773.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 763.0, + 1408.0, + 763.0, + 1408.0, + 801.0, + 347.0, + 801.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 791.0, + 419.0, + 791.0, + 419.0, + 825.0, + 351.0, + 825.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 841.0, + 1404.0, + 841.0, + 1404.0, + 873.0, + 307.0, + 873.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 865.0, + 1406.0, + 865.0, + 1406.0, + 905.0, + 347.0, + 905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 895.0, + 985.0, + 895.0, + 985.0, + 929.0, + 347.0, + 929.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 305.0, + 942.0, + 1410.0, + 942.0, + 1410.0, + 980.0, + 305.0, + 980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 345.0, + 968.0, + 1412.0, + 968.0, + 1412.0, + 1010.0, + 345.0, + 1010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 998.0, + 498.0, + 998.0, + 498.0, + 1032.0, + 347.0, + 1032.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1048.0, + 1406.0, + 1048.0, + 1406.0, + 1080.0, + 295.0, + 1080.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1074.0, + 1408.0, + 1074.0, + 1408.0, + 1113.0, + 347.0, + 1113.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1104.0, + 1014.0, + 1104.0, + 1014.0, + 1137.0, + 351.0, + 1137.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1151.0, + 1406.0, + 1151.0, + 1406.0, + 1189.0, + 293.0, + 1189.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1179.0, + 1052.0, + 1179.0, + 1052.0, + 1217.0, + 347.0, + 1217.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1229.0, + 1408.0, + 1229.0, + 1408.0, + 1261.0, + 295.0, + 1261.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1257.0, + 1406.0, + 1257.0, + 1406.0, + 1290.0, + 351.0, + 1290.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1283.0, + 423.0, + 1283.0, + 423.0, + 1318.0, + 349.0, + 1318.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1334.0, + 1404.0, + 1334.0, + 1404.0, + 1366.0, + 295.0, + 1366.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1358.0, + 1410.0, + 1358.0, + 1410.0, + 1396.0, + 349.0, + 1396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1388.0, + 471.0, + 1388.0, + 471.0, + 1420.0, + 349.0, + 1420.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1434.0, + 1410.0, + 1434.0, + 1410.0, + 1473.0, + 293.0, + 1473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1463.0, + 1242.0, + 1463.0, + 1242.0, + 1501.0, + 349.0, + 1501.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1513.0, + 1406.0, + 1513.0, + 1406.0, + 1545.0, + 295.0, + 1545.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1539.0, + 1408.0, + 1539.0, + 1408.0, + 1577.0, + 349.0, + 1577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1567.0, + 1352.0, + 1567.0, + 1352.0, + 1605.0, + 349.0, + 1605.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1617.0, + 1408.0, + 1617.0, + 1408.0, + 1650.0, + 295.0, + 1650.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1646.0, + 757.0, + 1646.0, + 757.0, + 1678.0, + 351.0, + 1678.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1694.0, + 1402.0, + 1694.0, + 1402.0, + 1726.0, + 295.0, + 1726.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1720.0, + 1060.0, + 1720.0, + 1060.0, + 1758.0, + 349.0, + 1758.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1768.0, + 1410.0, + 1768.0, + 1410.0, + 1804.0, + 291.0, + 1804.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1796.0, + 1408.0, + 1796.0, + 1408.0, + 1835.0, + 347.0, + 1835.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1843.0, + 1406.0, + 1843.0, + 1406.0, + 1883.0, + 291.0, + 1883.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1873.0, + 1406.0, + 1873.0, + 1406.0, + 1911.0, + 349.0, + 1911.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1901.0, + 1008.0, + 1901.0, + 1008.0, + 1939.0, + 349.0, + 1939.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1947.0, + 1410.0, + 1947.0, + 1410.0, + 1985.0, + 293.0, + 1985.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1975.0, + 1304.0, + 1975.0, + 1304.0, + 2014.0, + 349.0, + 2014.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 10, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 2, + "poly": [ + 836, + 2061, + 865, + 2061, + 865, + 2086, + 836, + 2086 + ], + "score": 0.796 + }, + { + "category_id": 1, + "poly": [ + 293, + 162, + 1410, + 162, + 1410, + 2023, + 293, + 2023 + ], + "score": 0.736 + }, + { + "category_id": 2, + "poly": [ + 836, + 2061, + 865, + 2061, + 865, + 2086, + 836, + 2086 + ], + "score": 0.131 + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2058.0, + 870.0, + 2058.0, + 870.0, + 2097.0, + 832.0, + 2097.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2058.0, + 870.0, + 2058.0, + 870.0, + 2097.0, + 832.0, + 2097.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 206.0, + 1404.0, + 206.0, + 1404.0, + 239.0, + 297.0, + 239.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 345.0, + 231.0, + 699.0, + 231.0, + 699.0, + 267.0, + 345.0, + 267.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 282.0, + 1406.0, + 282.0, + 1406.0, + 314.0, + 295.0, + 314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 310.0, + 1406.0, + 310.0, + 1406.0, + 343.0, + 349.0, + 343.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 345.0, + 335.0, + 603.0, + 335.0, + 603.0, + 371.0, + 345.0, + 371.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 384.0, + 1408.0, + 384.0, + 1408.0, + 422.0, + 293.0, + 422.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 412.0, + 1192.0, + 412.0, + 1192.0, + 445.0, + 351.0, + 445.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 461.0, + 1406.0, + 461.0, + 1406.0, + 494.0, + 295.0, + 494.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 484.0, + 1406.0, + 484.0, + 1406.0, + 525.0, + 347.0, + 525.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 345.0, + 514.0, + 699.0, + 514.0, + 699.0, + 551.0, + 345.0, + 551.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 568.0, + 1404.0, + 568.0, + 1404.0, + 594.0, + 297.0, + 594.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 592.0, + 1408.0, + 592.0, + 1408.0, + 625.0, + 351.0, + 625.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 619.0, + 421.0, + 619.0, + 421.0, + 653.0, + 349.0, + 653.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 668.0, + 1408.0, + 668.0, + 1408.0, + 700.0, + 295.0, + 700.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 696.0, + 1408.0, + 696.0, + 1408.0, + 729.0, + 349.0, + 729.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 719.0, + 423.0, + 719.0, + 423.0, + 760.0, + 347.0, + 760.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 772.0, + 1406.0, + 772.0, + 1406.0, + 804.0, + 295.0, + 804.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 798.0, + 611.0, + 798.0, + 611.0, + 831.0, + 351.0, + 831.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 845.0, + 1406.0, + 845.0, + 1406.0, + 884.0, + 293.0, + 884.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 876.0, + 1018.0, + 876.0, + 1018.0, + 909.0, + 351.0, + 909.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 923.0, + 1406.0, + 923.0, + 1406.0, + 956.0, + 295.0, + 956.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 945.0, + 1406.0, + 945.0, + 1406.0, + 988.0, + 347.0, + 988.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 976.0, + 1322.0, + 976.0, + 1322.0, + 1015.0, + 349.0, + 1015.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1021.0, + 1406.0, + 1021.0, + 1406.0, + 1062.0, + 293.0, + 1062.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1050.0, + 1408.0, + 1050.0, + 1408.0, + 1090.0, + 349.0, + 1090.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1081.0, + 631.0, + 1081.0, + 631.0, + 1113.0, + 347.0, + 1113.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1129.0, + 1406.0, + 1129.0, + 1406.0, + 1162.0, + 295.0, + 1162.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1158.0, + 1234.0, + 1158.0, + 1234.0, + 1191.0, + 351.0, + 1191.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1203.0, + 1410.0, + 1203.0, + 1410.0, + 1242.0, + 293.0, + 1242.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1233.0, + 1406.0, + 1233.0, + 1406.0, + 1266.0, + 349.0, + 1266.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1260.0, + 601.0, + 1260.0, + 601.0, + 1293.0, + 351.0, + 1293.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1309.0, + 1406.0, + 1309.0, + 1406.0, + 1342.0, + 295.0, + 1342.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1336.0, + 968.0, + 1336.0, + 968.0, + 1368.0, + 351.0, + 1368.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1385.0, + 1408.0, + 1385.0, + 1408.0, + 1417.0, + 295.0, + 1417.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 345.0, + 1409.0, + 759.0, + 1409.0, + 759.0, + 1446.0, + 345.0, + 1446.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1458.0, + 1408.0, + 1458.0, + 1408.0, + 1495.0, + 291.0, + 1495.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1489.0, + 1404.0, + 1489.0, + 1404.0, + 1521.0, + 349.0, + 1521.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1513.0, + 835.0, + 1513.0, + 835.0, + 1550.0, + 347.0, + 1550.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1560.0, + 1406.0, + 1560.0, + 1406.0, + 1597.0, + 293.0, + 1597.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1593.0, + 1404.0, + 1593.0, + 1404.0, + 1626.0, + 349.0, + 1626.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1620.0, + 611.0, + 1620.0, + 611.0, + 1650.0, + 349.0, + 1650.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1664.0, + 1408.0, + 1664.0, + 1408.0, + 1703.0, + 293.0, + 1703.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1693.0, + 1018.0, + 1693.0, + 1018.0, + 1730.0, + 347.0, + 1730.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1740.0, + 1406.0, + 1740.0, + 1406.0, + 1779.0, + 293.0, + 1779.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 345.0, + 1769.0, + 891.0, + 1769.0, + 891.0, + 1805.0, + 345.0, + 1805.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1816.0, + 1408.0, + 1816.0, + 1408.0, + 1854.0, + 293.0, + 1854.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1842.0, + 1410.0, + 1842.0, + 1410.0, + 1883.0, + 347.0, + 1883.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1875.0, + 573.0, + 1875.0, + 573.0, + 1908.0, + 351.0, + 1908.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1920.0, + 1408.0, + 1920.0, + 1408.0, + 1959.0, + 293.0, + 1959.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1950.0, + 1408.0, + 1950.0, + 1408.0, + 1983.0, + 349.0, + 1983.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1975.0, + 759.0, + 1975.0, + 759.0, + 2012.0, + 347.0, + 2012.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 11, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 2, + "poly": [ + 835, + 2060, + 865, + 2060, + 865, + 2086, + 835, + 2086 + ], + "score": 0.823 + }, + { + "category_id": 1, + "poly": [ + 293, + 145, + 1409, + 145, + 1409, + 2028, + 293, + 2028 + ], + "score": 0.788 + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2058.0, + 869.0, + 2058.0, + 869.0, + 2097.0, + 832.0, + 2097.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 207.0, + 1408.0, + 207.0, + 1408.0, + 240.0, + 294.0, + 240.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 233.0, + 1358.0, + 233.0, + 1358.0, + 266.0, + 351.0, + 266.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 277.0, + 553.0, + 277.0, + 553.0, + 310.0, + 294.0, + 310.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 608.0, + 273.0, + 904.0, + 273.0, + 904.0, + 314.0, + 608.0, + 314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 961.0, + 279.0, + 1406.0, + 279.0, + 1406.0, + 312.0, + 961.0, + 312.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 344.0, + 304.0, + 735.0, + 304.0, + 735.0, + 341.0, + 344.0, + 341.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 289.0, + 345.0, + 1413.0, + 345.0, + 1413.0, + 386.0, + 289.0, + 386.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 378.0, + 1252.0, + 378.0, + 1252.0, + 411.0, + 351.0, + 411.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 417.0, + 1408.0, + 417.0, + 1408.0, + 459.0, + 292.0, + 459.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 450.0, + 1408.0, + 450.0, + 1408.0, + 483.0, + 351.0, + 483.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 475.0, + 420.0, + 475.0, + 420.0, + 510.0, + 349.0, + 510.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 521.0, + 1406.0, + 521.0, + 1406.0, + 554.0, + 296.0, + 554.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 547.0, + 1408.0, + 547.0, + 1408.0, + 587.0, + 349.0, + 587.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 344.0, + 576.0, + 739.0, + 576.0, + 739.0, + 609.0, + 344.0, + 609.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 615.0, + 1408.0, + 615.0, + 1408.0, + 659.0, + 292.0, + 659.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 651.0, + 1406.0, + 651.0, + 1406.0, + 684.0, + 349.0, + 684.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 346.0, + 676.0, + 610.0, + 676.0, + 610.0, + 706.0, + 346.0, + 706.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 721.0, + 1406.0, + 721.0, + 1406.0, + 754.0, + 296.0, + 754.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 346.0, + 746.0, + 1406.0, + 746.0, + 1406.0, + 785.0, + 346.0, + 785.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 775.0, + 739.0, + 775.0, + 739.0, + 814.0, + 349.0, + 814.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 820.0, + 1406.0, + 820.0, + 1406.0, + 853.0, + 296.0, + 853.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 847.0, + 1402.0, + 847.0, + 1402.0, + 880.0, + 351.0, + 880.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 874.0, + 1187.0, + 874.0, + 1187.0, + 913.0, + 349.0, + 913.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 917.0, + 1410.0, + 917.0, + 1410.0, + 956.0, + 294.0, + 956.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 344.0, + 944.0, + 1408.0, + 944.0, + 1408.0, + 985.0, + 344.0, + 985.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 346.0, + 975.0, + 610.0, + 975.0, + 610.0, + 1006.0, + 346.0, + 1006.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1020.0, + 1406.0, + 1020.0, + 1406.0, + 1053.0, + 296.0, + 1053.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1049.0, + 1313.0, + 1049.0, + 1313.0, + 1082.0, + 351.0, + 1082.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1093.0, + 1404.0, + 1093.0, + 1404.0, + 1126.0, + 294.0, + 1126.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 346.0, + 1115.0, + 1406.0, + 1115.0, + 1406.0, + 1157.0, + 346.0, + 1157.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1146.0, + 612.0, + 1146.0, + 612.0, + 1179.0, + 349.0, + 1179.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1192.0, + 1406.0, + 1192.0, + 1406.0, + 1225.0, + 296.0, + 1225.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1219.0, + 908.0, + 1219.0, + 908.0, + 1252.0, + 349.0, + 1252.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1264.0, + 1406.0, + 1264.0, + 1406.0, + 1297.0, + 296.0, + 1297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1291.0, + 615.0, + 1291.0, + 615.0, + 1324.0, + 349.0, + 1324.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1332.0, + 1408.0, + 1332.0, + 1408.0, + 1374.0, + 292.0, + 1374.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 346.0, + 1361.0, + 1406.0, + 1361.0, + 1406.0, + 1400.0, + 346.0, + 1400.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1390.0, + 545.0, + 1390.0, + 545.0, + 1423.0, + 351.0, + 1423.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1436.0, + 1404.0, + 1436.0, + 1404.0, + 1469.0, + 296.0, + 1469.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 344.0, + 1460.0, + 1136.0, + 1460.0, + 1136.0, + 1498.0, + 344.0, + 1498.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1502.0, + 1406.0, + 1502.0, + 1406.0, + 1541.0, + 294.0, + 1541.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 344.0, + 1531.0, + 1408.0, + 1531.0, + 1408.0, + 1572.0, + 344.0, + 1572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 344.0, + 1562.0, + 1073.0, + 1562.0, + 1073.0, + 1597.0, + 344.0, + 1597.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1607.0, + 1406.0, + 1607.0, + 1406.0, + 1640.0, + 296.0, + 1640.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1636.0, + 1400.0, + 1636.0, + 1400.0, + 1669.0, + 351.0, + 1669.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1679.0, + 1408.0, + 1679.0, + 1408.0, + 1712.0, + 296.0, + 1712.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1708.0, + 1408.0, + 1708.0, + 1408.0, + 1741.0, + 349.0, + 1741.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 348.0, + 1733.0, + 589.0, + 1733.0, + 589.0, + 1764.0, + 348.0, + 1764.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1778.0, + 1408.0, + 1778.0, + 1408.0, + 1812.0, + 296.0, + 1812.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 349.0, + 1803.0, + 1121.0, + 1803.0, + 1121.0, + 1843.0, + 349.0, + 1843.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1851.0, + 1406.0, + 1851.0, + 1406.0, + 1884.0, + 296.0, + 1884.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 351.0, + 1878.0, + 1406.0, + 1878.0, + 1406.0, + 1911.0, + 351.0, + 1911.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 347.0, + 1900.0, + 815.0, + 1900.0, + 815.0, + 1942.0, + 347.0, + 1942.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1948.0, + 1408.0, + 1948.0, + 1408.0, + 1987.0, + 294.0, + 1987.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 344.0, + 1975.0, + 1216.0, + 1975.0, + 1216.0, + 2012.0, + 344.0, + 2012.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 12, + "width": 1700, + "height": 2200 + } + } +] \ No newline at end of file diff --git a/parse/train/H1cKvl-Rb/H1cKvl-Rb_content_list.json b/parse/train/H1cKvl-Rb/H1cKvl-Rb_content_list.json new file mode 100644 index 0000000000000000000000000000000000000000..21bd4231d7ae67a9bcaa6575d71f0bc58d47938e --- /dev/null +++ b/parse/train/H1cKvl-Rb/H1cKvl-Rb_content_list.json @@ -0,0 +1,1744 @@ +[ + { + "type": "text", + "text": "UCB EXPLORATION VIA $Q$ -ENSEMBLES ", + "text_level": 1, + "bbox": [ + 174, + 98, + 658, + 122 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Anonymous authors Paper under double-blind review ", + "bbox": [ + 183, + 150, + 398, + 178 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "ABSTRACT ", + "text_level": 1, + "bbox": [ + 454, + 213, + 544, + 229 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "We show how an ensemble of $Q ^ { * }$ -functions can be leveraged for more effective exploration in deep reinforcement learning. We build on well established algorithms from the bandit setting, and adapt them to the $Q$ -learning setting. We propose an exploration strategy based on upper-confidence bounds (UCB). Our experiments show significant gains on the Atari benchmark. ", + "bbox": [ + 232, + 244, + 764, + 315 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "1 INTRODUCTION ", + "text_level": 1, + "bbox": [ + 176, + 342, + 336, + 357 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Deep reinforcement learning seeks to learn mappings from high-dimensional observations to actions. Deep $Q$ -learning (Mnih et al. (2015)) is a leading technique that has been used successfully, especially for video game benchmarks. However, fundamental challenges remain, for example, improving sample efficiency and ensuring convergence to high quality solutions. Provably optimal solutions exist in the bandit setting and for small MDPs, and at the core of these solutions are exploration schemes. However these provably optimal exploration techniques do not extend to deep RL in a straightforward way. ", + "bbox": [ + 174, + 372, + 825, + 469 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Bootstrapped DQN (Osband et al. (2016)) is a previous attempt at adapting a theoretically verified approach to deep RL. In particular, it draws inspiration from posterior sampling for reinforcement learning (PSRL, Osband et al. (2013); Osband and Van Roy (2016)), which has near-optimal regret bounds. PSRL samples an MDP from its posterior each episode and exactly solves $Q ^ { * }$ , its optimal $Q$ -function. However, in high-dimensional settings, both approximating the posterior over MDPs and solving the sampled MDP are intractable. Bootstrapped DQN avoids having to establish and sample from the posterior over MDPs by instead approximating the posterior over $Q ^ { * }$ . In addition, bootstrapped DQN uses a multi-headed neural network to represent the $Q$ -ensemble. While the authors proposed bootstrapping to estimate the posterior distribution, their empirical findings show best performance is attained by simply relying on different initializations for the different heads, not requiring the sampling-with-replacement process that is prescribed by bootstrapping. ", + "bbox": [ + 174, + 478, + 825, + 631 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "In this paper, we design new algorithms that build on the $Q$ -ensemble approach from Osband et al. (2016). However, instead of using posterior sampling for exploration, we construct uncertainty estimates from the $Q$ -ensemble. Specifically, we first propose the Ensemble Voting algorithm where the agent takes action by a majority vote from the $Q$ -ensemble. Next, we propose the UCB exploration strategy. This strategy is inspired by established UCB algorithms in the bandit setting and constructs uncertainty estimates of the $Q$ -values. In this strategy, agents are optimistic and take actions with the highest UCB. We demonstrate that our algorithms significantly improve performance on the Atari benchmark. ", + "bbox": [ + 173, + 637, + 825, + 748 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "2 BACKGROUND ", + "text_level": 1, + "bbox": [ + 176, + 768, + 326, + 785 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "2.1 NOTATION ", + "text_level": 1, + "bbox": [ + 174, + 799, + 287, + 814 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "We model reinforcement learning as a Markov decision process (MDP). We define an MDP as $( S , A , T , R , p _ { 0 } , \\gamma )$ , in which both the state space $s$ and action space $\\mathcal { A }$ are discrete, $T : S \\times \\mathcal { A } \\times \\mathcal { S } \\mapsto$ $\\mathbb { R } _ { + }$ is the transition distribution, $R : S \\times \\mathcal { A } \\mapsto \\mathbb { R }$ is the reward function, assumed deterministic given the state and action, and $\\gamma \\in ( 0 , 1 ]$ is a discount factor, and $p _ { 0 }$ is the initial state distribution. We denote a transition experience as $\\tau = ( s , a , r , s ^ { \\prime } )$ where $s ^ { \\prime } \\sim T ( s ^ { \\prime } | s , a )$ and $r = R ( s , a )$ . A policy $\\pi : { \\mathcal { S } } \\mapsto A$ specifies the action taken after observing a state. We denote the $Q$ -function for policy $\\pi$ as $\\begin{array} { r } { Q ^ { \\pi } ( s , a ) : = \\mathbb E _ { \\pi } \\big [ \\sum _ { t = 0 } ^ { \\infty } \\gamma ^ { t } r _ { t } | s _ { 0 } = s , a _ { 0 } = a \\big ] } \\end{array}$ where $r _ { t } = R ( s _ { t } , a _ { t } )$ . The optimal $Q ^ { * }$ -function ", + "bbox": [ + 173, + 825, + 825, + 926 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "corresponds to taking the optimal policy ", + "bbox": [ + 174, + 103, + 439, + 118 + ], + "page_idx": 1 + }, + { + "type": "equation", + "img_path": "images/585c41a817ea806365929ea09dd3db34a8a3fee59a73dc7c9db4daae6e8aa0b6.jpg", + "text": "$$\nQ ^ { \\ast } ( s , a ) : = \\operatorname* { s u p } _ { \\pi } Q ^ { \\pi } ( s , a )\n$$", + "text_format": "latex", + "bbox": [ + 413, + 121, + 584, + 145 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "and satisfies the Bellman equation ", + "bbox": [ + 174, + 147, + 400, + 161 + ], + "page_idx": 1 + }, + { + "type": "equation", + "img_path": "images/9029853fe46dcb5f0d5c68c08fc9599dc35a0b40542bbf04df25b8852d736340.jpg", + "text": "$$\nQ ^ { * } ( s , a ) = \\mathbb { E } _ { s ^ { \\prime } \\sim T ( \\cdot \\mid s , a ) } \\big [ r + \\gamma \\cdot \\operatorname* { m a x } _ { a ^ { \\prime } } Q ^ { * } ( s ^ { \\prime } , a ^ { \\prime } ) \\big ] .\n$$", + "text_format": "latex", + "bbox": [ + 336, + 162, + 660, + 186 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "2.2 EXPLORATION IN REINFORCEMENT LEARNING ", + "text_level": 1, + "bbox": [ + 174, + 202, + 537, + 215 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "A notable early optimality result in reinforcement learning was the proof by Watkins and Dayan Watkins (1989); Watkins and Dayan (1992) that an online $Q$ -learning algorithm is guaranteed to converge to the optimal policy, provided that every state is visited an infinite number of times. However, the convergence of Watkins’ Q-learning can be prohibitively slow in MDPs where $\\epsilon$ - greedy action selection explores state space randomly. Later work developed reinforcement learning algorithms with provably fast (polynomial-time) convergence (Kearns and Singh (2002); Brafman and Tennenholtz (2002); Strehl et al. (2006)). At the core of these provably-optimal learning methods is some exploration strategy, which actively encourages the agent to visit novel state-action pairs. For example, R-MAX optimistically assumes that infrequently-visited states provide maximal reward, and delayed $Q$ -learning initializes the $Q$ -function with high values to ensure that each state-action is chosen enough times to drive the value down. ", + "bbox": [ + 174, + 227, + 826, + 380 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "Since the theoretically sound RL algorithms are not computationally practical in the deep RL setting, deep RL implementations often use simple exploration methods such as $\\epsilon$ -greedy and Boltzmann exploration, which are often sample-inefficient and fail to find good policies. One common approach of exploration in deep RL is to construct an exploration bonus, which adds a reward for visiting state-action pairs that are deemed to be novel or informative. In particular, several prior methods define an exploration bonus based on a density model or dynamics model. Examples include VIME by Houthooft et al. (2016), which uses variational inference on the forward-dynamics model, and Tang et al. (2016), Bellemare et al. (2016), Ostrovski et al. (2017), Fu et al. (2017). While these methods yield successful exploration in some problems, a major drawback is that this exploration bonus does not depend on the rewards, so the exploration may focus on irrelevant aspects of the environment, which are unrelated to reward. ", + "bbox": [ + 174, + 387, + 825, + 539 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "2.3 BAYESIAN REINFORCEMENT LEARNING ", + "text_level": 1, + "bbox": [ + 176, + 556, + 488, + 570 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "Earlier works on Bayesian reinforcement learning include Dearden et al. (1998; 1999). Dearden et al. (1998) studied Bayesian $Q$ -learning in the model-free setting and learned the distribution of $Q ^ { * }$ - values through Bayesian updates. The prior and posterior specification relied on several simplifying assumptions, some of which are not compatible with the MDP setting. Dearden et al. (1999) took a model-based approach that updates the posterior distribution of the MDP. The algorithm samples from the MDP posterior multiple times and solving the $Q ^ { * }$ values at every step. This approach is only feasible for RL problems with very small state space and action space. Strens (2000) proposed posterior sampling for reinforcement learning (PSRL). PSRL instead takes a single sample of the MDP from the posterior in each episode and solves the $Q ^ { * }$ values. Recent works including Osband et al. (2013) and Osband and Van Roy (2016) established near-optimal Bayesian regret bounds for episodic RL. Sorg et al. (2012) models the environment and constructs exploration bonus from variance of model parameters. These methods are experimented on low dimensional problems only, because the computational cost of these methods is intractable for high dimensional RL. ", + "bbox": [ + 174, + 582, + 825, + 762 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "2.4 BOOTSTRAPPED DQN ", + "text_level": 1, + "bbox": [ + 176, + 779, + 369, + 794 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "Inspired by PSRL, but wanting to reduce computational cost, prior work developed approximate methods. Osband et al. (2014) proposed randomized least-square value iteration for linearly-parameterized value functions. Bootstrapped DQN Osband et al. (2016) applies to $Q$ -functions parameterized by deep neural networks. Bootstrapped DQN (Osband et al. (2016)) maintains a $Q$ -ensemble, represented by a multi-head neural net structure to parameterize $K \\in \\mathbb { N } _ { + }$ $Q$ -functions. This multi-head structure shares the convolution layers but includes multiple “heads”, each of which defines a $Q$ -function $Q _ { k }$ . ", + "bbox": [ + 174, + 804, + 825, + 888 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "Bootstrapped DQN diversifies the $Q$ -ensemble through two mechanisms. The first mechanism is independent initialization. The second mechanism applies different samples to train each $Q$ -function. ", + "bbox": [ + 174, + 895, + 823, + 924 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "These $Q$ -functions can be trained simultaneously by combining their loss functions with the help of a random mask $m _ { \\tau } \\in \\mathbb { R } _ { + } ^ { K }$ C ", + "bbox": [ + 169, + 103, + 825, + 133 + ], + "page_idx": 2 + }, + { + "type": "equation", + "img_path": "images/554949743a237969dd6d0aa476836d424be4c434fb5a0dcdc1f97be541606197.jpg", + "text": "$$\nL = \\sum _ { \\tau \\in { \\cal B } _ { \\mathrm { m i n i } } } \\sum _ { k = 1 } ^ { K } m _ { \\tau } ^ { k } \\cdot ( Q ^ { k } ( s , a ; \\theta ) - y _ { \\tau } ^ { Q _ { k } } ) ^ { 2 } ,\n$$", + "text_format": "latex", + "bbox": [ + 326, + 150, + 668, + 180 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "where $y _ { \\tau } ^ { Q _ { k } }$ is the target of the $k$ th $Q$ -function. Thus, the transition $\\tau$ updates $Q _ { k }$ only if $m _ { \\tau } ^ { k }$ is nonzero. To avoid the overestimation issue in DQN, bootstrapped DQN calculates the target value $y _ { \\tau } ^ { Q _ { k } }$ using the approach of Double DQN (Van Hasselt et al. (2016)), such that the current $Q _ { k } ( \\cdot ; \\theta _ { t } )$ network determines the optimal action and the target network $Q _ { k } \\big ( \\cdot ; \\theta ^ { - } \\big )$ estimates the value ", + "bbox": [ + 173, + 196, + 826, + 253 + ], + "page_idx": 2 + }, + { + "type": "equation", + "img_path": "images/ccd49d06bffb4624bb90294bf014c922f655ebda5448b2ab077deed9d20903b2.jpg", + "text": "$$\ny _ { \\tau } ^ { Q _ { k } } = r + \\gamma \\operatorname* { m a x } _ { a } Q ^ { k } ( s ^ { \\prime } , \\operatorname * { a r g m a x } _ { a } Q _ { k } ( s ^ { \\prime } , a ; \\theta _ { t } ) ; \\theta ^ { - } ) .\n$$", + "text_format": "latex", + "bbox": [ + 325, + 270, + 671, + 296 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "In their experiments on Atari games, Osband et al. (2016) set the mask $m _ { \\tau } = ( 1 , \\ldots , 1 )$ such that all $\\left\\{ Q _ { k } \\right\\}$ are trained with the same samples and their only difference is initialization. Bootstrapped DQN picks one $Q _ { k }$ uniformly at random at the start of an episode and follows the greedy action $a _ { t } = \\operatorname { a r g m a x } _ { a } Q _ { k } ( s _ { t } , a )$ for the whole episode. ", + "bbox": [ + 173, + 313, + 825, + 369 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "3 ENSEMBLE VOTING ", + "text_level": 1, + "bbox": [ + 176, + 400, + 370, + 416 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "Ignoring computational costs, the ideal Bayesian approach to reinforcement learning is to maintain a posterior over the MDP. However, with limited computation and model capacity, it is more tractable to maintain a posterior of the $Q ^ { * }$ -function. This motivates using a $Q$ -ensemble as a particle filter-based approach to approximate the posterior over $Q ^ { * }$ -function and we display our first proposed method, Ensemble Voting, in Algorithm 1. ", + "bbox": [ + 173, + 439, + 825, + 510 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "Each $Q _ { k }$ in the $Q$ -ensemble $\\{ Q _ { k } \\} _ { k = 1 } ^ { K }$ is parametrized with a deep neural network whose parameters are initialized independently at the start of training. Each $Q _ { k }$ proposes an action that maximizes the $Q$ -value according to $Q _ { k }$ at every time step and the agent chooses the action by a majority vote ", + "bbox": [ + 174, + 515, + 825, + 559 + ], + "page_idx": 2 + }, + { + "type": "equation", + "img_path": "images/be15c40e9293b32696349156add62173ac8516b290f0b01bdbc992e20510a470.jpg", + "text": "$$\na _ { t } = \\mathop { \\mathrm { M a j o r i t y } } \\mathrm { V o t e } ( \\{ \\operatorname { a r g m a x } Q _ { k } ( s _ { t } , a ) \\} _ { k = 1 } ^ { K } ) .\n$$", + "text_format": "latex", + "bbox": [ + 346, + 574, + 650, + 601 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "At each learning interval, a minibatch of transitions is sampled from the replay buffer and each $Q _ { k }$ takes a Bellman update based on this minibatch. For stability, Algorithm 1 also uses a target network for each $Q _ { k }$ as in Double DQN in the batched update. ", + "bbox": [ + 174, + 616, + 825, + 660 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "We point out that the difference among the parameters of the $Q$ -ensemble $\\left\\{ Q _ { k } \\right\\}$ comes only from the independent random initialization. The deep neural network parametrization of the $Q$ -ensemble introduces nonconvexity into the objective function of Bellman update, so the $Q$ -ensemble $\\left\\{ Q _ { k } \\right\\}$ do not converge to the same $Q$ -function during training even though they are trained with the same minibatches at every update. We also experimented with bagging by updating each $Q _ { k }$ using an independently drawn minibatch. However, bagging led to inferior learning performance. This phenomenon that that bagging deteriorates the performance of deep ensembles is also observed in supervised learning settings. Lee et al. (2015) observed that supervised learning trained with deep ensembles with random initializations perform better than bagging for deep ensembles. Lakshminarayanan et al. (2016) used deep ensembles for uncertainty estimates and also observed that bagging deteriorated performance in their experiments. ", + "bbox": [ + 173, + 665, + 825, + 819 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "Lu and Van Roy (2017) develop ensemble sampling for bandit problems with deep neural network parametrized policies and the theoretical justification. We derive a posterior update rule for the $Q ^ { * }$ function and approximations to the posterior update using ensembles in Appendix C. We note that in bootstrapped DQN, ensemble voting is applied for evaluation while Algorithm 1 uses ensemble voting during learning. In the experiments (Sec. 5), we demonstrate that Algorithm 1 is superior to bootstrapped DQN. The action choice of Algorithm 1 is exploitation only. In the next section, we propose our UCB exploration strategy. ", + "bbox": [ + 173, + 825, + 825, + 924 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "Algorithm 1 Ensemble Voting ", + "text_level": 1, + "bbox": [ + 174, + 103, + 377, + 118 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "1: Input: $K \\in \\mathbb { N } _ { + }$ copies of independently initialized $Q ^ { * }$ -functions $\\{ Q _ { k } \\} _ { k = 1 } ^ { K }$ . \n2: Let $B$ be a replay buffer storing transitions for training \n3: for each episode do do \n4: Obtain initial state from environment $s _ { 0 }$ \n5: for step $t = 1 , \\dots$ until end of episode do \n6: Pick an action according to $\\hat { a } _ { t } = \\mathrm { M a j o r i t y V o t e } ( \\{ \\operatorname { a r g m a x } _ { a } Q _ { k } ( s _ { t } , a ) \\} _ { k = 1 } ^ { K } )$ \n7: Execute $a _ { t }$ . Receive state $s _ { t + 1 }$ and reward $r _ { t }$ from the environment \n8: Add $\\left( { { s _ { t } } , { a _ { t } } , { r _ { t } } , { s _ { t + 1 } } } \\right)$ to replay buffer $B$ \n9: At learning interval, sample random minibatch and update $\\left\\{ Q _ { k } \\right\\}$ \n10: end for \n11: end for ", + "bbox": [ + 179, + 121, + 750, + 279 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "4 UCB EXPLORATION STRATEGY USING $Q$ -ENSEMBLES ", + "text_level": 1, + "bbox": [ + 176, + 303, + 661, + 320 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "In this section, we propose optimism-based exploration by adapting the UCB algorithms (Auer et al. (2002); Audibert et al. (2009)) from the bandit setting. The UCB algorithms maintain an upper-confidence bound for each arm, such that the expected reward from pulling each arm is smaller than this bound with high probability. At every time step, the agent optimistically chooses the arm with the highest UCB. Auer et al. (2002) constructed the UCB based on empirical reward and the number of times each arm is chosen. Audibert et al. (2009) incorporated the empirical variance of each arm’s reward into the UCB, such that at time step $t$ , an arm $A _ { t }$ is pulled according to ", + "bbox": [ + 173, + 334, + 826, + 433 + ], + "page_idx": 3 + }, + { + "type": "equation", + "img_path": "images/954f078fad211df1f6f64f010ac189beb0de98bb47a9bc8efbf18c0cbdf9c80b.jpg", + "text": "$$\nA _ { t } = \\underset { i } { \\operatorname { a r g m a x } } \\left\\{ \\hat { r } _ { i , t } + c _ { 1 } \\cdot \\sqrt { \\frac { \\hat { V } _ { i , t } \\log ( t ) } { n _ { i , t } } } + c _ { 2 } \\cdot \\frac { \\log ( t ) } { n _ { i , t } } \\right\\}\n$$", + "text_format": "latex", + "bbox": [ + 316, + 436, + 679, + 479 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "where $\\hat { r } _ { i , t }$ and $\\hat { V } _ { i , t }$ are the empirical reward and variance of arm $i$ at time $t$ , $n _ { i , t }$ is the number of times arm $i$ has been pulled up to time $t$ , and $c _ { 1 } , c _ { 2 }$ are positive constants. ", + "bbox": [ + 173, + 484, + 823, + 515 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "We extend the intuition of UCB algorithms to the RL setting. Using the outputs of the $\\left\\{ Q _ { k } \\right\\}$ functions, we construct a UCB by adding the empirical standard deviation $\\tilde { \\sigma } ( s _ { t } , a )$ of $\\{ Q _ { k } ( s _ { t } , a ) \\} _ { k = 1 } ^ { K }$ to the empirical mean $\\tilde { \\mu } ( s _ { t } , a )$ of $\\{ Q _ { k } ( s _ { t } , a ) \\} _ { k = 1 } ^ { K }$ . The agent chooses the action that maximizes this UCB ", + "bbox": [ + 173, + 520, + 826, + 565 + ], + "page_idx": 3 + }, + { + "type": "equation", + "img_path": "images/b6510d2fb041bf9b71551eb8c5e65fc457aeece694383a9f447bd4f716941e3d.jpg", + "text": "$$\na _ { t } \\in \\mathop { \\operatorname { a r g m a x } } _ { a } \\left\\{ \\tilde { \\mu } ( s _ { t } , a ) + \\lambda \\cdot \\tilde { \\sigma } ( s _ { t } , a ) \\right\\} ,\n$$", + "text_format": "latex", + "bbox": [ + 367, + 569, + 627, + 595 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "where $\\lambda \\in \\mathbb { R } _ { + }$ is a hyperparameter. ", + "bbox": [ + 174, + 599, + 406, + 614 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "We present Algorithm 2, which incorporates the UCB exploration. The hyperparemeter $\\lambda$ controls the degrees of exploration. In Section 5, we compare the performance of our algorithms on Atari games using a consistent set of parameters. ", + "bbox": [ + 174, + 619, + 825, + 662 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "Algorithm 2 UCB Exploration with $Q$ -Ensembles ", + "text_level": 1, + "bbox": [ + 173, + 676, + 506, + 693 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "1: Input: Value function networks $Q$ with $K$ outputs $\\{ Q _ { k } \\} _ { k = 1 } ^ { K }$ . Hyperparameter $\\lambda$ . \n2: Let $B$ be a replay buffer storing experience for training. \n3: for each episode do \n4: Obtain initial state from environment $s _ { 0 }$ \n5: for step $t = 1 , \\dots$ until end of episode do \n6: Pick an action according to $\\begin{array} { r } { \\grave { a _ { t } } \\in \\mathrm { a r g m a x } _ { a } \\left\\{ \\tilde { \\mu } ( s _ { t } , a ) + \\lambda \\cdot \\tilde { \\sigma } ( s _ { t } , a ) \\right\\} } \\end{array}$ \n7: Receive state $s _ { t + 1 }$ and reward $r _ { t }$ from environment, having taken action $a _ { t }$ \n8: Add $\\left( { { s _ { t } } , { a _ { t } } , { r _ { t } } , { s _ { t + 1 } } } \\right)$ to replay buffer $B$ \n9: At learning interval, sample random minibatch and update $\\left\\{ Q _ { k } \\right\\}$ \n10: end for \n11: end for ", + "bbox": [ + 179, + 695, + 741, + 852 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "5 EXPERIMENT ", + "text_level": 1, + "bbox": [ + 174, + 878, + 316, + 895 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "In this section, we conduct experiments to answer the following questions: ", + "bbox": [ + 173, + 909, + 661, + 924 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "1. does Ensemble Voting, Algorithm 1, improve upon existing algorithms including Double DQN and bootstrapped DQN? \n2. is the proposed UCB exploration strategy of Algorithm 2 effective in improving learning compared to Algorithm 1, Double DQN and bootstrapped DQN? \n3. how does UCB exploration compare with prior exploration methods such as the count-based exploration method of Bellemare et al. (2016)? ", + "bbox": [ + 212, + 103, + 825, + 204 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "We evaluate the algorithms on each Atari game of the Arcade Learning Environment (Bellemare et al. (2013)). We use the multi-head neural net architecture of Osband et al. (2016). We fix the common hyperparameters of all algorithms based on a well-tuned double DQN implementation, which uses the Adam optimizer (Kingma and Ba (2014)), different learning rate and exploration schedules compared to Mnih et al. (2015). Appendix A tabulates the hyperparameters. The number of $\\left\\{ Q _ { k } \\right\\}$ functions is $K = 1 0$ . Experiments are conducted on the OpenAI Gym platform (Brockman et al. (2016)) and trained with 40 million frames and 2 trials on each game. ", + "bbox": [ + 174, + 217, + 825, + 315 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "We take the following directions to evaluate the performance of our algorithms: ", + "bbox": [ + 173, + 321, + 692, + 337 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "1. we compare Algorithm 1 against Double DQN and bootstrapped DQN, \n2. we isolate the impact of UCB exploration by comparing Algorithm 2 with $\\lambda = 0 . 1$ , denoted as ucb exploration, against Algorithm 1, Double DQN, and bootstrapped DQN. \n3. we compare Algorithm 1 and Algorithm 2 with the count-based exploration method of Bellemare et al. (2016). \n4. we aggregate the comparison according to different categories of games, to understand when our methods are suprior. ", + "bbox": [ + 210, + 348, + 825, + 472 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "Figure 1 compares the normalized learning curves of all algorithms across Atari games. Overall, Ensemble Voting, Algorithm 1, outperforms both Double DQN and bootstrapped DQN. With exploration, ucb exploration improves further by outperforming Ensemble Voting. ", + "bbox": [ + 174, + 486, + 825, + 527 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "In Appendix B, we tabulate detailed results that compare our algorithms, Ensemble Voting and ucb exploration, against prior methods. In Table 2, we tabulate the maximal mean reward in 100 consecutive episodes for Ensemble Voting, ucb exploration, bootstrapped DQN and Double DQN. Without exploration, Ensemble Voting already achieves higher maximal mean reward than both Double DQN and bootstrapped DQN in a majority of Atari games. Ensemble Voting performs better than Double DQN in 37 games out of the total 49 games evaluated, better than bootstrapped DQN in 41 games. ucb exploration achieves the highest maximal mean reward among these four algorithms in 30 games out of the total 49 games evaluated. Specifically, ucb exploration performs better than Double DQN in 38 out of 49 games evaluated, better than bootstrapped DQN in 45 games, and better than Ensemble Voting in 35 games. Figure 2 displays the learning curves of these five algorithms on a set of six Atari games. Ensemble Voting outperforms Double DQN and bootstrapped DQN. ucb exploration outperforms Ensemble Voting. ", + "bbox": [ + 174, + 535, + 825, + 702 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "In Table 3, we compare our proposed methods with the count-based exploration method ${ \\bf A } 3 { \\bf C } +$ of Bellemare et al. (2016) based on their published results of ${ \\bf A } 3 { \\bf C } +$ trained with 200 million frames. We point out that even though our methods were trained with only 40 million frames, much less than ${ \\bf A } 3 { \\bf C } +$ ’s 200 million frames, UCB exploration achieves the highest average reward in 28 games, Ensemble Voting in 10 games, and ${ \\bf A } 3 { \\bf C } +$ in 10 games. Our approach outperforms ${ \\bf A } 3 { \\bf C } +$ . ", + "bbox": [ + 174, + 708, + 825, + 777 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "Finally to understand why and when the proposed methods are superior, we aggregate the comparison results according to four categories: Human Optimal, Score Explicit, Dense Reward, and Sparse Reward. These categories follow the taxonomy in Table 1 of Ostrovski et al. (2017). Out of all games evaluated, 23 games are Human Optimal, 8 are Score Explicit, 8 are Dense Reward, and 5 are Sparse Reward. The comparison results are tabulated in Table 4, where we see ucb exploration achieves top performance in more games than Ensemble Voting, Double DQN, and Bootstrapped DQN in the categories of Human Optimal, Score Explicit, and Dense Reward. In Sparse Reward, both ucb exploration and Ensemble Voting achieve best performance in 2 games out of total of 5. Thus, we conclude that ucb exploration improves prior methods consistently across different game categories within the Arcade Learning Environment. ", + "bbox": [ + 174, + 785, + 825, + 924 + ], + "page_idx": 4 + }, + { + "type": "image", + "img_path": "images/755944a1fbb499ae90c0dbc15edc278a059fe19214caa71c8855f17a85eb8094.jpg", + "image_caption": [ + "Figure 1: Comparison of algorithms in normalized learning curve. The normalized learning curve is calculated as follows: first, we normalize learning curves for all algorithms in the same game to the interval [0, 1]; next, average the normalized learning curve from all games for each algorithm. " + ], + "image_footnote": [], + "bbox": [ + 187, + 133, + 849, + 426 + ], + "page_idx": 5 + }, + { + "type": "image", + "img_path": "images/187b14c9628fc4f8130883728517c533abd226a73264ac629bbd4c497df9fc9f.jpg", + "image_caption": [ + "Figure 2: Comparison of UCB Exploration and Ensemble Voting against Double DQN and Bootstrapped DQN. " + ], + "image_footnote": [], + "bbox": [ + 173, + 551, + 862, + 847 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "6 CONCLUSION ", + "text_level": 1, + "bbox": [ + 174, + 102, + 318, + 118 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "We proposed a $Q$ -ensemble approach to deep $Q$ -learning, a computationally practical algorithm inspired by Bayesian reinforcement learning that outperforms Double DQN and bootstrapped DQN, as evaluated on Atari. The key ingredient is the UCB exploration strategy, inspired by bandit algorithms. Our experiments show that the exploration strategy achieves improved learning performance on the majority of Atari games. ", + "bbox": [ + 174, + 132, + 825, + 204 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "REFERENCES ", + "text_level": 1, + "bbox": [ + 174, + 224, + 285, + 239 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Jean-Yves Audibert, Rémi Munos, and Csaba Szepesvári. Exploration–exploitation tradeoff using variance estimates in multi-armed bandits. Theor. Comput. Sci., 410(19):1876–1902, 2009. ", + "bbox": [ + 173, + 247, + 825, + 276 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Peter Auer, Nicolo Cesa-Bianchi, and Paul Fischer. Finite-time analysis of the multiarmed bandit problem. Mach. Learn., 47(2-3):235–256, 2002. ", + "bbox": [ + 174, + 284, + 821, + 313 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Marc Bellemare, Sriram Srinivasan, Georg Ostrovski, Tom Schaul, David Saxton, and Remi Munos. Unifying count-based exploration and intrinsic motivation. In NIPS, pages 1471–1479, 2016. ", + "bbox": [ + 173, + 320, + 823, + 351 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Marc G Bellemare, Yavar Naddaf, Joel Veness, and Michael Bowling. The arcade learning environment: An evaluation platform for general agents. J. Artif. Intell. Res., 47:253–279, 2013. ", + "bbox": [ + 173, + 358, + 825, + 388 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Ronen I Brafman and Moshe Tennenholtz. R-max-a general polynomial time algorithm for nearoptimal reinforcement learning. J. Mach. Learn. Res., 3(Oct):213–231, 2002. ", + "bbox": [ + 174, + 396, + 823, + 425 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Greg Brockman, Vicki Cheung, Ludwig Pettersson, Jonas Schneider, John Schulman, Jie Tang, and Wojciech Zaremba. OpenAI Gym. arXiv preprint arXiv:1606.01540, 2016. ", + "bbox": [ + 176, + 433, + 823, + 463 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Richard Dearden, Nir Friedman, and Stuart Russell. Bayesian Q-learning. In AAAI/IAAI, pages 761–768, 1998. ", + "bbox": [ + 176, + 469, + 823, + 500 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Richard Dearden, Nir Friedman, and David Andre. Model based Bayesian exploration. In UAI, pages 150–159, 1999. ", + "bbox": [ + 174, + 507, + 823, + 536 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Justin Fu, John D Co-Reyes, and Sergey Levine. EX2: Exploration with exemplar models for deep reinforcement learning. arXiv preprint arXiv:1703.01260, 2017. ", + "bbox": [ + 173, + 545, + 823, + 575 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Rein Houthooft, Xi Chen, Yan Duan, John Schulman, Filip De Turck, and Pieter Abbeel. VIME: Variational information maximizing exploration. In NIPS, pages 1109–1117, 2016. ", + "bbox": [ + 174, + 582, + 821, + 612 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Michael Kearns and Satinder Singh. Near-optimal reinforcement learning in polynomial time. Mach. Learn., 49(2-3):209–232, 2002. ", + "bbox": [ + 174, + 619, + 823, + 648 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Diederik Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014. ", + "bbox": [ + 174, + 656, + 821, + 685 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blundell. Simple and scalable predictive uncertainty estimation using deep ensembles. arXiv preprint arXiv:1612.01474, 2016. ", + "bbox": [ + 173, + 694, + 821, + 723 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Stefan Lee, Senthil Purushwalkam, Michael Cogswell, David Crandall, and Dhruv Batra. Why M heads are better than one: Training a diverse ensemble of deep networks. arXiv preprint arXiv:1511.06314, 2015. ", + "bbox": [ + 173, + 732, + 825, + 775 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Xiuyuan Lu and Benjamin Van Roy. Ensemble sampling. 2017. ", + "bbox": [ + 173, + 784, + 593, + 799 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A Rusu, Joel Veness, Marc G Bellemare, Alex Graves, Martin Riedmiller, Andreas K Fidjeland, Georg Ostrovski, et al. Human-level control through deep reinforcement learning. Nature, 518(7540):529–533, 2015. ", + "bbox": [ + 176, + 806, + 823, + 849 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Ian Osband and Benjamin Van Roy. Why is posterior sampling better than optimism for reinforcement learning. arXiv preprint arXiv:1607.00215, 2016. ", + "bbox": [ + 173, + 858, + 821, + 887 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Ian Osband, Dan Russo, and Benjamin Van Roy. (More) efficient reinforcement learning via posterior sampling. In NIPS, pages 3003–3011, 2013. ", + "bbox": [ + 174, + 895, + 821, + 924 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Ian Osband, Benjamin Van Roy, and Zheng Wen. Generalization and exploration via randomized value functions. arXiv preprint arXiv:1402.0635, 2014. \nIan Osband, Charles Blundell, Alexander Pritzel, and Benjamin Van Roy. Deep exploration via bootstrapped DQN. In NIPS, pages 4026–4034, 2016. \nGeorg Ostrovski, Marc G Bellemare, Aaron van den Oord, and Remi Munos. Count-based exploration with neural density models. arXiv preprint arXiv:1703.01310, 2017. \nJonathan Sorg, Satinder Singh, and Richard L Lewis. Variance-based rewards for approximate bayesian reinforcement learning. arXiv preprint arXiv:1203.3518, 2012. \nAlexander L Strehl, Lihong Li, Eric Wiewiora, John Langford, and Michael L Littman. Pac model-free reinforcement learning. In ICML, pages 881–888. ACM, 2006. \nMalcolm Strens. A Bayesian framework for reinforcement learning. In ICML, pages 943–950, 2000. \nYi Sun, Faustino Gomez, and Jürgen Schmidhuber. Planning to be surprised: Optimal Bayesian exploration in dynamic environments. In ICAGI, pages 41–51. Springer, 2011. \nHaoran Tang, Rein Houthooft, Davis Foote, Adam Stooke, Xi Chen, Yan Duan, John Schulman, Filip De Turck, and Pieter Abbeel. # Exploration: A study of count-based exploration for deep reinforcement learning. arXiv preprint arXiv:1611.04717, 2016. \nHado Van Hasselt, Arthur Guez, and David Silver. Deep reinforcement learning with double Qlearning. In AAAI, pages 2094–2100, 2016. \nChristopher JCH Watkins and Peter Dayan. Q-learning. Mach. Learn., 8(3-4):279–292, 1992. \nChristopher John Cornish Hellaby Watkins. Learning from delayed rewards. PhD thesis, University of Cambridge England, 1989. ", + "bbox": [ + 169, + 103, + 828, + 497 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "A HYPERPARAMETERS ", + "text_level": 1, + "bbox": [ + 176, + 102, + 382, + 118 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "We tabulate the hyperparameters in our well-tuned implementation of double DQN in Table 1: ", + "bbox": [ + 173, + 143, + 789, + 159 + ], + "page_idx": 8 + }, + { + "type": "table", + "img_path": "images/896390fffca83112777cedcbd8f03dd31ff87650868945a92d74639f978326d5.jpg", + "table_caption": [ + "Table 1: Double DQN hyperparameters. These hyperparameters are selected based on performances of seven Atari games: Beam Rider, Breakout, Pong, Enduro, Qbert, Seaquest, and Space Invaders. $I n t e r p ( \\cdot , \\cdot )$ is linear interpolation between two values. " + ], + "table_footnote": [], + "table_body": "
value
hyperparameter total training frames40 milliondescriptionsLength of training for each game.
minibatch size32
replay buffer size1000000parameter update.The number of most recent frames
agent history length4
length.concatenated as input to the Q net- work. Total number of iterations = total training frames /agent history
target network update10000 frequency
discount factor0.99
action repeat4
update frequency 4
optimizerAdamOptimizer for parameter updates.
β1 0.9Adam optimizer parameter.
β0.99Adam optimizer parameter.
E 10-4Adam optimizer parameter.
learning rate schedule 210-4 Interp(10-4,5 * 10-5) 5*10-5t≤106 otherwise t>5*106Learning rate for Adam optimizer, as a function of iteration t.
exploration scheduleInterp(1,0.1) Interp(0.1,0.01) 0.01t<106 otherwise Probability of random action in e- t>5*106
greedy exploration, as a function of the iteration t . Number of uniform random ac- tions taken before learning starts.
replay start size50000
", + "bbox": [ + 178, + 176, + 859, + 875 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "B RESULTS TABLES ", + "text_level": 1, + "bbox": [ + 176, + 103, + 352, + 117 + ], + "page_idx": 9 + }, + { + "type": "table", + "img_path": "images/4eb43f40bc33336067ff917a144de986a4192279d3974d786f0bb7f1c6ae1bcd.jpg", + "table_caption": [], + "table_footnote": [ + "Table 2: Comparison of maximal mean rewards achieved by agents. Maximal mean reward is calculated in a window of 100 consecutive episodes. Bold denotes the highest value in each row. " + ], + "table_body": "
UCB-Exploration
Bootstrapped DQNDoubleDQNEnsemble Voting 2282.8
Alien1445.12059.72817.6
Amidar430.58667.5683.72663.8
Assault2519.062820.613213.583702.76
Asterix3829.07639.58740.08732.0
Asteroids1009.51002.31149.31007.8
Atlantis1314058.01982677.01786305.02016145.0
Bank Heist795.1789.9869.4906.9
Battle Zone26230.024880.027430.026770.0
Beam Rider8006.587743.747991.99188.26
Bowling28.6230.9232.9238.06
Boxing85.9194.0794.4798.08
Breakout400.22467.45426.78411.31
Centipede5328.775177.516153.286237.18
Chopper Command2153.03260.03544.03677.0
Crazy Climber110926.0124456.0126677.0127754.0
Demon Attack9811.4523562.5530004.459861.9
Double Dunk-10.82-14.58-11.94-4.08
Enduro1314.311439.591999.882752.55
Fishing Derby21.8923.6930.0229.71
Freeway33.5732.9333.9233.96
Frostbite1284.8529.21196.01903.0
Gopher7652.212030.010993.212910.8
Gravitar227.5279.5371.5318.0
Ice Hockey-4.62-4.63-1.73-4.71
Jamesbond594.5594.0602.0710.0
Kangaroo8186.07787.08174.014196.0
Krull8537.528517.918669.179171.61
Kung Fu Master24153.032896.030988.031291.0
Montezuma Revenge2.04.01.04.0
Ms Pacman2508.72498.13039.73425.4
Name This Game8212.49806.99255.19570.5
Pitfall-5.99-7.57-3.37-1.47
Pong21.020.6721.020.95
Private Eye1815.19788.631845.281252.01
Qbert10557.256529.512036.514198.25
Riverraid11528.011834.712785.815622.2
Road Runner52489.049039.054768.053596.0
Robotank21.0329.831.8341.04
Seaquest9320.718056.420458.624001.6
Space Invaders1549.91917.51890.82626.55
Star Gunner20115.052283.041684.047367.0
Tennis-15.11-14.04-11.63-7.8
Time Pilot5088.05548.06153.06490.0
Tutankham167.47223.43208.61200.76
Up N Down9049.111815.319528.319827.3
Venture115.096.078.067.0
Video Pinball364600.85374686.89343380.29372564.11
Wizard Of Wor2860.03877.05451.05873.0
Zaxxon592.08903.03901.03695.0
Times best17930
", + "bbox": [ + 179, + 155, + 854, + 880 + ], + "page_idx": 9 + }, + { + "type": "table", + "img_path": "images/27e3fba1891440807533f9b61b3ee4160ebb19c8a6f9c4e7e5bbc0ef5a241f83.jpg", + "table_caption": [], + "table_footnote": [ + "Table 3: Comparison of Ensemble Voting, UCB Exploration, both trained with 40 million frames and ${ \\bf A } 3 { \\bf C } +$ of Bellemare et al. (2016), trained with 200 million frames " + ], + "table_body": "
Ensemble VotingUCB-ExplorationA3C+
Alien2282.82817.61848.33
Amidar683.72663.8964.77
Assault3213.583702.762607.28
Asterix8740.08732.07262.77
Asteroids1149.31007.82257.92
Atlantis1786305.02016145.01733528.71
Bank Heist869.4906.9991.96
Battle Zone27430.026770.07428.99
Beam Rider7991.99188.265992.08
Bowling32.9238.0668.72
Boxing94.4798.0813.82
Breakout426.78411.31323.21
Centipede6153.286237.185338.24
Chopper Command3544.03677.05388.22
Crazy Climber126677.0127754.0104083.51
Demon Attack30004.459861.919589.95
Double Dunk-11.94-4.08-8.88
Enduro1999.882752.55749.11
Fishing Derby30.0229.7129.46
Freeway33.9233.9627.33
Frostbite1196.01903.0506.61
Gopher10993.212910.85948.40
Gravitar371.5318.0246.02
Ice Hockey-1.73-4.71-7.05
Jamesbond602.0710.01024.16
Kangaroo8174.014196.05475.73
Krull8669.179171.617587.58
Kung Fu Master30988.031291.026593.67
Montezuma Revenge1.04.0142.50
Ms Pacman3039.73425.42380.58
Name This Game9255.19570.56427.51
Pitfall-3.37-1.47-155.97
Pong21.020.9517.33
Private Eye1845.281252.01100.0
Qbert12036.514198.2515804.72
Riverraid12785.815622.210331.56
Road Runner54768.053596.049029.74
Robotank31.8341.046.68
Seaquest20458.624001.62274.06
Space Invaders1890.82626.551466.01
Star Gunner41684.047367.052466.84
Tennis-11.63-7.8-20.49
Time Pilot6153.06490.03816.38
Tutankham208.61200.76132.67
Up N Down19528.319827.38705.64
Venture78.067.00.00
Video Pinball343380.29372564.1135515.92
Wizard Of Wor5451.05873.03657.65
Zaxxon3901.03695.07956.05
Times Best102810
", + "bbox": [ + 243, + 94, + 758, + 818 + ], + "page_idx": 10 + }, + { + "type": "table", + "img_path": "images/005571787932de1656c370acfee80aa9b5104f24d15ee5200d9347b27ed2a0d1.jpg", + "table_caption": [], + "table_footnote": [], + "table_body": "
CategoryTotalBootstrapped DQNDouble DQNEnsemble VotingUCB-Exploration
Human Optimal2303515
Score Explicit80215
Dense Reward80116
Sparse Reward51022
", + "bbox": [ + 173, + 99, + 872, + 174 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Table 4: Comparison of each method across different game categories. The Atari games are separated into four categories: human optimal, score explicit, dense reward, and sparse reward. In each row, we present the number of games in this category, the total number of games where each algorithm achieves the optimal performance according to Table 2. The game categories follow the taxonomy in Table 1 of Ostrovski et al. (2017) ", + "bbox": [ + 173, + 189, + 826, + 260 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "C APPROXIMATING BAYESIAN $Q$ -LEARNING WITH $Q$ -ENSEMBLES ", + "text_level": 1, + "bbox": [ + 173, + 284, + 743, + 301 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "In this section, we first derive a posterior update formula for the $Q ^ { * }$ -function under full exploration assumption and this formula turns out to depend on the transition Markov chain. Next, we approximate the posterior update with $Q$ -ensembles $\\{ { \\bar { Q } } _ { k } \\}$ and demonstrate that the Bellman equation emerges as the approximate update rule for each $Q _ { k }$ . ", + "bbox": [ + 173, + 314, + 825, + 372 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "C.1 POSTERIOR UPDATE FOR THE $Q ^ { * }$ -FUNCTION ", + "text_level": 1, + "bbox": [ + 174, + 387, + 526, + 402 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "An MDP is specified by the transition probability $T$ and the reward function $R$ . Unlike prior works outlined in Section 2.3 which learned the posterior of the MDP, we will consider the joint distribution over $( Q ^ { * } , T )$ . Note that $R$ can be recovered from $Q ^ { * }$ given $T$ . So $( Q ^ { * } , T )$ determines a unique MDP. In this section, we assume that the agent samples $( s , a )$ according to a fixed distribution. The corresponding reward $r$ and next state $s ^ { \\prime }$ given by the MDP append to $( s , a )$ to form a transition ${ \\boldsymbol \\tau } = ( s , a , r , s ^ { \\prime } )$ , for updating the posterior of $( Q ^ { * } , T )$ . Recall that the $Q ^ { * }$ -function satisfies the Bellman equation ", + "bbox": [ + 173, + 412, + 825, + 511 + ], + "page_idx": 11 + }, + { + "type": "equation", + "img_path": "images/a6677133076ccab2959a65d94d7e4fb0872abfdf9ae434ebc50645522db14b11.jpg", + "text": "$$\nQ ( s , a ) = r + \\mathbb { E } _ { s ^ { \\prime } \\sim T ( \\cdot | s , a ) } \\left[ \\gamma \\operatorname* { m a x } _ { a ^ { \\prime } } Q ( s ^ { \\prime } , a ^ { \\prime } ) \\right] .\n$$", + "text_format": "latex", + "bbox": [ + 344, + 510, + 651, + 536 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Denote the joint prior distribution as $p ( Q ^ { * } , T )$ and the posterior as $\\tilde { p }$ . We apply Bayes’ formula to expand the posterior: ", + "bbox": [ + 174, + 539, + 823, + 566 + ], + "page_idx": 11 + }, + { + "type": "equation", + "img_path": "images/49c85f5489553a7e3afb98aca65933136c3196eb0e7aba19aecddb350c6c2167.jpg", + "text": "$$\n\\begin{array} { r l } & { \\tilde { p } ( Q ^ { * } , T | \\tau ) = \\frac { p ( \\tau | Q ^ { * } , T ) \\cdot p ( Q ^ { * } , T ) } { Z ( \\tau ) } } \\\\ & { \\qquad = \\frac { p ( Q ^ { * } , T ) \\cdot p ( s ^ { \\prime } | Q ^ { * } , T , ( s , a ) ) \\cdot p ( r | Q ^ { * } , T , ( s , a , s ^ { \\prime } ) ) \\cdot p ( s , a ) } { Z ( \\tau ) } , } \\end{array}\n$$", + "text_format": "latex", + "bbox": [ + 246, + 569, + 748, + 641 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "where $Z ( \\tau )$ is a normalizing constant and the second equality is because $s$ and $a$ are sampled randomly from $s$ and $\\mathcal { A }$ . Next, we calculate the two conditional probabilities in (1). First, ", + "bbox": [ + 176, + 643, + 823, + 672 + ], + "page_idx": 11 + }, + { + "type": "equation", + "img_path": "images/602cdad18535c501841b89ecb1826559681ee676df0c9a35d39a813d615bf0be.jpg", + "text": "$$\np ( s ^ { \\prime } | Q ^ { * } , T , ( s , a ) ) = p ( s ^ { \\prime } | T , ( s , a ) ) = T ( s ^ { \\prime } | s , a ) ,\n$$", + "text_format": "latex", + "bbox": [ + 334, + 675, + 661, + 694 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "where the first equality is because given $T$ , $Q ^ { * }$ does not influence the transition. Second, ", + "bbox": [ + 173, + 696, + 751, + 712 + ], + "page_idx": 11 + }, + { + "type": "equation", + "img_path": "images/fb0955c02f65e1e496f065faeae25ba9c05cbf13f12d277ebac49da28a4215c6.jpg", + "text": "$$\n\\begin{array} { r l } & { p ( r | Q ^ { * } , T , ( s , a , s ^ { \\prime } ) ) = p ( r | Q ^ { * } , T , ( s , a ) ) } \\\\ & { \\phantom { p s p a c e } = \\mathbb { 1 } _ { \\{ Q ^ { * } ( s , a ) = r + \\gamma \\cdot \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\cdot \\vert s , a ) } \\operatorname* { m a x } _ { a ^ { \\prime } } Q ^ { * } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) \\} } } \\\\ & { \\phantom { p s p a c e } : = \\mathbb { 1 } ( Q ^ { * } , T ) , } \\end{array}\n$$", + "text_format": "latex", + "bbox": [ + 282, + 714, + 715, + 773 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "where $\\mathbb { 1 } _ { \\{ \\cdot \\} }$ is the indicator function and in the last equation we abbreviate it as $\\mathbb { 1 } ( Q ^ { * } , T )$ . Substituting (2) and (3) into (1), we obtain the joint posterior of $Q ^ { * }$ and $T$ after observing an additional randomly sampled transition $\\tau$ ", + "bbox": [ + 176, + 775, + 825, + 818 + ], + "page_idx": 11 + }, + { + "type": "equation", + "img_path": "images/9e1f8436cc01293f7e533e96b6775e8cf9531c631fae4a8f68650d6cb6ddb3a0.jpg", + "text": "$$\n\\tilde { p } ( Q ^ { * } , T | \\tau ) = \\frac { p ( Q ^ { * } , T ) \\cdot T ( s ^ { \\prime } | s , a ) \\cdot p ( s , a ) } { Z ( \\tau ) } \\cdot \\mathbb { 1 } ( Q ^ { * } , T ) .\n$$", + "text_format": "latex", + "bbox": [ + 310, + 820, + 686, + 854 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "C.2 APPROXIMATIONS WITH $Q$ -ENSEMBLES ", + "text_level": 1, + "bbox": [ + 173, + 869, + 493, + 885 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "The exact $Q ^ { * }$ -posterior update (4) is intractable in high-dimensional RL due to the large space of $( Q ^ { * } , T )$ . Thus, we make several approximations to the $Q ^ { * }$ -posterior update. First, we approximate the prior of $Q ^ { * }$ by sampling $K \\in \\mathbb { N } _ { + }$ independently initialized $Q ^ { * }$ -functions $\\{ Q _ { k } \\} _ { k = 1 } ^ { K }$ . Next, we update them as more transitions are sampled. The resulting $\\left\\{ Q _ { k } \\right\\}$ approximate samples drawn from the posterior. The agent chooses the action by taking a majority vote from the actions determined by each $Q _ { k }$ . ", + "bbox": [ + 173, + 895, + 825, + 924 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "", + "bbox": [ + 173, + 102, + 825, + 160 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "We derive the update rule for $\\left\\{ Q _ { k } \\right\\}$ after observing a new transition $\\tau = ( s , a , r , s ^ { \\prime } )$ . At iteration $i$ , given $Q ^ { * } = Q _ { k , i } ( \\cdot ; \\theta _ { k } )$ parametrized by $\\theta _ { k }$ the joint probability of $( Q ^ { * } , T )$ factors into ", + "bbox": [ + 174, + 165, + 825, + 195 + ], + "page_idx": 12 + }, + { + "type": "equation", + "img_path": "images/a9e3490d38f1df33eee5e4e8f26512fd49304477b682900ff7783bd506bbe021.jpg", + "text": "$$\np ( Q _ { k , i } , T ) = p ( Q ^ { * } , T | Q ^ { * } = Q _ { k , i } ) = p ( T | Q _ { k , i } ) .\n$$", + "text_format": "latex", + "bbox": [ + 336, + 200, + 661, + 219 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Substitute (5) into (4) and we obtain the corresponding posterior for each $Q _ { k , i + 1 }$ at iteration $i + 1$ as ", + "bbox": [ + 174, + 223, + 820, + 239 + ], + "page_idx": 12 + }, + { + "type": "equation", + "img_path": "images/2161b1d580d9c2964aed6a05d3774e58926355fa4fa34b423c443d47e0cad512.jpg", + "text": "$$\n\\begin{array} { r l r } { { \\tilde { p } ( Q _ { k , i + 1 } , T | \\tau ) = \\frac { p ( T | Q _ { k , i } ) \\cdot T ( s ^ { \\prime } | s , a ) \\cdot p ( s , a ) } { Z ( \\tau ) } \\cdot \\mathbb { 1 } ( Q _ { k , i + 1 } , T ) . } } \\\\ & { } & { \\tilde { p } ( Q _ { k , i + 1 } | \\tau ) = \\int _ { T } \\tilde { p } ( Q _ { k , i + 1 } , T | \\tau ) \\mathrm { d } T = p ( s , a ) \\cdot \\int _ { T } \\tilde { p } ( T | Q _ { k , i } , \\tau ) \\cdot \\mathbb { 1 } ( Q _ { k , i + 1 } , T ) \\mathrm { d } T . } \\end{array}\n$$", + "text_format": "latex", + "bbox": [ + 210, + 244, + 784, + 314 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "We update $Q _ { k , i }$ to $Q _ { k , i + 1 }$ according to ", + "bbox": [ + 174, + 318, + 431, + 334 + ], + "page_idx": 12 + }, + { + "type": "equation", + "img_path": "images/3572d2c11640739b572990e946c494fde34c0c4e2420f4edaef3189bc4796174.jpg", + "text": "$$\nQ _ { k , i + 1 } \\operatorname * { a r g m a x } _ { Q _ { k , i + 1 } } \\tilde { p } ( Q _ { k , i + 1 } | \\tau ) .\n$$", + "text_format": "latex", + "bbox": [ + 390, + 338, + 607, + 366 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "We first derive a lower bound of the the posterior $\\tilde { p } ( Q _ { k , i + 1 } | \\tau )$ : ", + "bbox": [ + 173, + 372, + 584, + 388 + ], + "page_idx": 12 + }, + { + "type": "equation", + "img_path": "images/aadb27b42c7df13b92fe96ca8b2bd2e3c4f0eb84124981ebf634d4b0c43dae83.jpg", + "text": "$$\n\\begin{array} { r l r } { { \\operatorname { i } ( Q _ { k , i + 1 } \\lvert \\tau \\rangle = p ( s , a ) \\cdot \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } \\mathbb { 1 } ( Q _ { k , i + 1 } , T ) } } \\\\ & { = p ( s , a ) \\cdot \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } \\underset { c \\to + \\infty } { \\operatorname { i m } } \\exp ( - c [ Q _ { k , i + 1 } ( s , a ) - r - \\gamma \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\lvert s , a ) } \\underset { a ^ { \\prime } } { \\operatorname { m a x } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ] ^ { 2 } ) } \\\\ & { = p ( s , a ) \\cdot \\underset { c \\to + \\infty } { \\operatorname* { i m } } \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } \\exp ( - c [ Q _ { k , i + 1 } ( s , a ) - r - \\gamma \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\lvert s , a ) } \\underset { a ^ { \\prime } } { \\operatorname { m a x } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ] ^ { 2 } ) } \\\\ & { \\geq p ( s , a ) \\cdot \\underset { c \\to + \\infty } { \\operatorname* { i m } } \\exp ( - c \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } [ Q _ { k , i + 1 } ( s , a ) - r - \\gamma \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\lvert s , a ) } \\underset { a ^ { \\prime } } { \\operatorname { m a x } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ] ^ { 2 } ) } \\\\ & { = p ( s , a ) \\cdot \\underset { c \\to + \\infty } { \\operatorname* { l i m } } \\underset { c \\to \\tau ( \\tau \\lvert Q _ { k , i } , \\tau ) } { \\operatorname* { i m } } [ Q _ { k , i + 1 } ( s , a ) - r - \\gamma \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\lvert s , a ) } \\underset { a ^ { \\prime } } { \\operatorname* { m a x } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ] ^ { 2 } } \\\\ & = p ( s , a ) \\cdot \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } [ Q _ k , \\end{array}\n$$", + "text_format": "latex", + "bbox": [ + 181, + 392, + 838, + 512 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "where we apply a limit representation of the indicator function in the third equation. The fourth equation is due to the bounded convergence theorem. The inequality is Jensen’s inequality. The last equation (9) replaces the limit with an indicator function. ", + "bbox": [ + 174, + 516, + 826, + 559 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "A sufficient condition for (8) is to maximize the lower-bound of the posterior distribution in (9) by ensuring the indicator function in (9) to hold. We can replace (8) with the following update ", + "bbox": [ + 171, + 565, + 823, + 594 + ], + "page_idx": 12 + }, + { + "type": "equation", + "img_path": "images/1cc85ed7d62f85ce8d16545a0958e89d101ba108cc4e25fe6629d2e2ecef62f7.jpg", + "text": "$$\nQ _ { k , i + 1 } \\underset { Q _ { k , i + 1 } } { \\mathrm { a r g m i n } } \\mathbb { E } _ { T \\sim \\tilde { p } ( T | Q _ { k , i } , \\tau ) } [ Q _ { k , i + 1 } ( s , a ) - ( r + \\gamma \\cdot \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot | s , a ) } \\operatorname* { m a x } _ { a ^ { \\prime } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ) ] ^ { 2 } .\n$$", + "text_format": "latex", + "bbox": [ + 183, + 601, + 813, + 632 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "However, (10) is not tractable because the expectation in (10) is taken with respect to the posterior $\\tilde { p } ( T | Q _ { k , i } , \\tau )$ of the transition $T$ . To overcome this challenge, we approximate the posterior update by reusing the one-sample next state $s ^ { \\prime }$ from $\\tau$ . Solving the exact minimal for each $Q _ { k , i + 1 }$ is impractical, thus we take a gradient step on $Q _ { k , i + 1 }$ according to the following gradient ", + "bbox": [ + 173, + 642, + 825, + 700 + ], + "page_idx": 12 + }, + { + "type": "equation", + "img_path": "images/4bea0dec3faa79e0caee8d86ff93eb1847812a86435a690e5db1076c1a7d9095.jpg", + "text": "$$\n\\theta _ { k } \\theta _ { k } + \\eta \\cdot ( Q _ { k } ( s , a ; \\theta _ { k } ) - ( r + \\gamma \\cdot \\operatorname* { m a x } _ { a ^ { \\prime } } Q _ { k } ( s ^ { \\prime } , a ^ { \\prime } ; \\theta _ { k } ) ) ) \\nabla _ { \\theta _ { k } } Q _ { k } ( s , a ; \\theta _ { k } ) ,\n$$", + "text_format": "latex", + "bbox": [ + 243, + 705, + 753, + 729 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "where $\\eta$ is the step size. Instead of updating $Q _ { k }$ after each transition, we use an experience replay buffer $B$ to store observed transitions and sample a minibatch $B _ { \\mathrm { m i n i } }$ of transitions $( s , a , r , s ^ { \\prime } )$ for each update. In this case, the batched update of each $Q _ { k , i }$ to $Q _ { k , i + 1 }$ becomes a standard Bellman update ", + "bbox": [ + 173, + 734, + 825, + 791 + ], + "page_idx": 12 + }, + { + "type": "equation", + "img_path": "images/d70b3f36248f6004e8bc44efa652a042d00c1404f0bb9b03800759a5a683cd41.jpg", + "text": "$$\n\\theta _ { k } \\gets \\theta _ { k } + \\eta \\cdot \\mathbb { E } _ { ( s , a , r , s ^ { \\prime } ) \\in B _ { \\operatorname* { m i n } } } \\big [ \\big ( Q _ { k } \\big ( s , a ; \\theta _ { k } \\big ) - \\big ( r + \\gamma \\cdot \\operatorname* { m a x } _ { a ^ { \\prime } } Q _ { k } \\big ( s ^ { \\prime } , a ^ { \\prime } ; \\theta _ { k } \\big ) \\big ) \\big ) \\nabla _ { \\theta _ { k } } Q _ { k } \\big ( s , a ; \\theta _ { k } \\big ) \\big ] .\n$$", + "text_format": "latex", + "bbox": [ + 183, + 795, + 812, + 819 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "D INFOGAIN EXPLORATION ", + "text_level": 1, + "bbox": [ + 176, + 837, + 421, + 853 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "In this section, we also studied an “InfoGain” exploration bonus, which encourages agents to gain information about the $Q ^ { * }$ -function and examine its effectiveness. We found it had some benefits on top of Ensemble Voting, but no uniform additional benefits once already using Q-ensembles on top of Double DQN. We describe the approach and our experimental findings. ", + "bbox": [ + 174, + 867, + 825, + 924 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Similar to Sun et al. (2011), we define the information gain from observing an additional transition $\\tau _ { n }$ as ", + "bbox": [ + 171, + 103, + 823, + 132 + ], + "page_idx": 13 + }, + { + "type": "equation", + "img_path": "images/92d0a80953a40c9a25364420585521da683a2bbc55554267f030e06126cf824a.jpg", + "text": "$$\nH _ { \\tau _ { t } | \\tau _ { 1 } , \\dots , \\tau _ { n - 1 } } = D _ { K L } ( \\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { n } ) | | \\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { n - 1 } ) )\n$$", + "text_format": "latex", + "bbox": [ + 290, + 132, + 704, + 151 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "where $\\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { n } )$ is the posterior distribution of $Q ^ { * }$ after observing a sequence of transitions $\\left( \\tau _ { 1 } , \\dots , \\tau _ { n } \\right)$ . The total information gain is ", + "bbox": [ + 173, + 156, + 820, + 185 + ], + "page_idx": 13 + }, + { + "type": "equation", + "img_path": "images/f989e42d3a3df5cd3eacb6baeb2a0fdf4ed66ccb011a106b222986f026240b2d.jpg", + "text": "$$\nH _ { \\tau _ { 1 } , \\dots , \\tau _ { N } } = \\sum _ { n = 1 } ^ { N } H _ { \\tau _ { n } | \\tau _ { 1 } , \\dots , \\tau _ { n - 1 } } .\n$$", + "text_format": "latex", + "bbox": [ + 379, + 194, + 617, + 223 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "Our Ensemble Voting, Algorithm 1, does not maintain the posterior $\\tilde { p }$ , thus we cannot calculate (11) explicitly. Instead, inspired by Lakshminarayanan et al. (2016), we define an InfoGain exploration bonus that measures the disagreement among $\\left\\{ Q _ { k } \\right\\}$ . Note that ", + "bbox": [ + 173, + 229, + 825, + 273 + ], + "page_idx": 13 + }, + { + "type": "equation", + "img_path": "images/da85240ec8e2f545f5e402235f477ab4c68b55ff03768b899da2490a4bc78ae3.jpg", + "text": "$$\nH _ { \\tau _ { 1 } , \\dots , \\tau _ { N } } + \\mathsf { H } ( \\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { N } ) ) = \\mathsf { H } ( p ( Q ^ { * } ) ) ,\n$$", + "text_format": "latex", + "bbox": [ + 339, + 281, + 656, + 299 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "where $\\mathsf { H } ( \\cdot )$ is the entropy. If $H _ { \\tau _ { 1 } , \\dots , \\tau _ { N } }$ is small, then the posterior distribution has high entropy and high residual information. Since $\\left\\{ Q _ { k } \\right\\}$ are approximate samples from the posterior, high entropy of the posterior leads to large discrepancy among $\\left\\{ Q _ { k } \\right\\}$ . Thus, the exploration bonus is monotonous with respect to the residual information in the posterior $\\mathsf { H } ( \\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { N } ) )$ . We first compute the Boltzmann distribution for each $Q _ { k }$ ", + "bbox": [ + 173, + 306, + 825, + 377 + ], + "page_idx": 13 + }, + { + "type": "equation", + "img_path": "images/24be8a27e56f39b989b40f301208b39a1e67cb305d0df0b0a77cb478192364bf.jpg", + "text": "$$\nP _ { \\mathsf { T } , k } ( a | s ) = \\frac { \\exp \\left( Q _ { k } ( s , a ) / \\mathsf { T } \\right) } { \\sum _ { a ^ { \\prime } } \\exp \\left( Q _ { k } ( s , a ^ { \\prime } ) / \\mathsf { T } \\right) } ,\n$$", + "text_format": "latex", + "bbox": [ + 370, + 385, + 625, + 424 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "where $\\mathsf T > 0$ is a temperature parameter. Next, calculate the average Boltzmann distribution ", + "bbox": [ + 169, + 431, + 776, + 445 + ], + "page_idx": 13 + }, + { + "type": "equation", + "img_path": "images/fd7f46f57b39c6d76b6ea0551a04fbcf7f0fff4f0f12971f16ea7d9371731d99.jpg", + "text": "$$\nP _ { \\mathsf { T } , \\mathrm { a v g } } = \\frac { 1 } { K } \\cdot \\sum _ { k = 1 } ^ { K } P _ { \\mathsf { T } , k } ( a | s ) .\n$$", + "text_format": "latex", + "bbox": [ + 388, + 454, + 607, + 484 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "The InfoGain exploration bonus is the average KL-divergence from $\\{ P _ { \\mathsf { T } , k } \\} _ { k = 1 } ^ { K }$ to $P _ { \\mathrm { { T , a v g } } }$ ", + "bbox": [ + 171, + 492, + 754, + 510 + ], + "page_idx": 13 + }, + { + "type": "equation", + "img_path": "images/0b2a7fedeb71301cadf9a63b4396515b553a6a5e5eaf3244c7354a20783153fb.jpg", + "text": "$$\nb _ { \\mathsf { T } } ( s ) = \\frac { 1 } { K } \\cdot \\sum _ { k = 1 } ^ { K } \\mathrm { D } _ { K L } [ P _ { \\mathsf { T } , k } | | P _ { \\mathsf { T } , \\mathrm { a v g } } ] .\n$$", + "text_format": "latex", + "bbox": [ + 361, + 517, + 637, + 547 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "The modified reward is ", + "bbox": [ + 174, + 555, + 326, + 569 + ], + "page_idx": 13 + }, + { + "type": "equation", + "img_path": "images/2271a374b1bd2f8522eea625ab54732f3037b9931d39f79a368d1c20741583c3.jpg", + "text": "$$\n\\hat { r } ( s , a , s ^ { \\prime } ) = r ( s , a ) + \\rho \\cdot b \\tau ( s ) ,\n$$", + "text_format": "latex", + "bbox": [ + 392, + 570, + 606, + 588 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "where $\\rho \\in \\mathbb { R } _ { + }$ is a hyperparameter that controls the degree of exploration. ", + "bbox": [ + 174, + 593, + 658, + 608 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "The exploration bonus $b _ { \\mathsf { T } } ( s _ { t } )$ encourages the agent to explore where $\\left\\{ Q _ { k } \\right\\}$ disagree. The temperature parameter $\\top$ controls the sensitivity to discrepancies among $\\{ Q _ { k } \\}$ . When $\\mathsf { T } \\to + \\infty$ , $\\{ P _ { \\top , k } \\}$ converge to the uniform distribution on the action space and $b _ { \\mathsf { T } } ( s ) \\to 0$ . When $\\top$ is small, the differences among $\\left\\{ Q _ { k } \\right\\}$ are magnified and $b _ { \\mathsf { T } } ( s )$ is large. ", + "bbox": [ + 173, + 614, + 825, + 671 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "We display Algorithrim 3, which incorporates our InfoGain exploration bonus into Algorithm 2. The hyperparameters $\\lambda$ , $\\top$ and $\\rho$ vary for each game. ", + "bbox": [ + 171, + 676, + 823, + 707 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "Algorithm 3 UCB + InfoGain Exploration with $Q$ -Ensembles ", + "text_level": 1, + "bbox": [ + 173, + 723, + 581, + 738 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "1: Input: Value function networks $Q$ with $K$ outputs $\\{ Q _ { k } \\} _ { k = 1 } ^ { K }$ . Hyperparameters $\\tau , \\lambda$ , and $\\rho$ . \n2: Let $B$ be a replay buffer storing experience for training. \n3: for each episode do \n4: Obtain initial state from environment $s _ { 0 }$ \n5: for step $t = 1 , \\dots$ until end of episode do \n6: Pick an action according to $\\begin{array} { r } { \\grave { a _ { t } } \\in \\mathrm { a r g m a x } _ { a } \\left\\{ \\tilde { \\mu } ( s _ { t } , a ) + \\lambda \\cdot \\tilde { \\sigma } ( s _ { t } , a ) \\right\\} } \\end{array}$ \n7: Receive state $s _ { t + 1 }$ and reward $r _ { t }$ from environment, having taken action $a _ { t }$ \n8: Calculate exploration bonus $b _ { \\mathsf { T } } ( s _ { t } )$ according to (12) \n9: Add $( s _ { t } , a _ { t } , r _ { t } + \\rho \\cdot b _ { \\mathsf { T } } ( s _ { t } ) , s _ { t + 1 } )$ to replay buffer $B$ \n10: At learning interval, sample random minibatch and update $\\left\\{ Q _ { k } \\right\\}$ \n11: end for \n12: end for ", + "bbox": [ + 178, + 741, + 808, + 911 + ], + "page_idx": 13 + }, + { + "type": "image", + "img_path": "images/939ce04cb216d1efd18ae513f715268a3148f6d9a0bb94bd5582028e5311ce05.jpg", + "image_caption": [ + "Figure 3: Comparison of all algorithms in normalized curve. The normalized learning curve is calculated as follows: first, we normalize learning curves for all algorithms in the same game to the interval [0, 1]; next, average the normalized learning curve from all games for each algorithm. " + ], + "image_footnote": [], + "bbox": [ + 186, + 102, + 849, + 414 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "D.1 PERFORMANCE OF UCB $^ +$ INFOGAIN EXPLORATION ", + "text_level": 1, + "bbox": [ + 174, + 502, + 573, + 517 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "We demonstrate the performance of the combined UCB+InfoGain exploration in Figure 3 and Figure 3. We augment the previous figures in Section 5 with the performance of ucb+infogain exploration, where we set $\\lambda = 0 . 1 , \\rho = 1$ , and ${ \\mathsf T } = 1$ in Algorithm 3. ", + "bbox": [ + 174, + 529, + 823, + 571 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "Figure 3 shows that combining UCB and InfoGain exploration does not lead to uniform improvement in the normalized learning curve. ", + "bbox": [ + 171, + 577, + 823, + 606 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "At the individual game level, Figure 3 shows that the impact of InfoGain exploration varies. UCB exploration achieves sufficient exploration in games including Demon Attack and Kangaroo and Riverraid, while InfoGain exploration further improves learning on Enduro, Seaquest, and Up N Down. The effect of InfoGain exploration depends on the choice of the temperature $\\top$ . The optimal temperature parameter varies across games. In Figure 5, we display the behavior of ucb+infogain exploration with different temperature values. Thus, we see the InfoGain exploration bonus, tuned with the appropriate temperature parameter, can lead to improved learning for games that require extra exploration, such as ChopperCommand, KungFuMaster, Seaquest, UpNDown. ", + "bbox": [ + 173, + 613, + 825, + 738 + ], + "page_idx": 14 + }, + { + "type": "image", + "img_path": "images/298f295b77aa13e16d25b06d19768431fd0cd7bebf49664568a091f6ffef1383.jpg", + "image_caption": [ + "Figure 4: Comparison of algorithms against Double DQN and bootstrapped DQN. " + ], + "image_footnote": [], + "bbox": [ + 174, + 349, + 864, + 641 + ], + "page_idx": 15 + }, + { + "type": "image", + "img_path": "images/bb37003bba1e0413ac880ed76c095033248f764b2a5e78eee9d9c0b3878a47c9.jpg", + "image_caption": [ + "D.2 UCB $^ +$ INFOGAIN EXPLORATION WITH DIFFERENT TEMPERATURES", + "Figure 5: Comparison of UCB+InfoGain exploration with different temperatures versus UCB exploration. " + ], + "image_footnote": [], + "bbox": [ + 202, + 126, + 782, + 792 + ], + "page_idx": 16 + } +] \ No newline at end of file diff --git a/parse/train/H1cKvl-Rb/H1cKvl-Rb_model.json b/parse/train/H1cKvl-Rb/H1cKvl-Rb_model.json new file mode 100644 index 0000000000000000000000000000000000000000..d4484773bd551b406a708112652305da338930c9 --- /dev/null +++ b/parse/train/H1cKvl-Rb/H1cKvl-Rb_model.json @@ -0,0 +1,25815 @@ +[ + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1053, + 1404, + 1053, + 1404, + 1389, + 298, + 1389 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 298, + 822, + 1404, + 822, + 1404, + 1036, + 298, + 1036 + ], + "score": 0.98 + }, + { + "category_id": 1, + "poly": [ + 297, + 1403, + 1404, + 1403, + 1404, + 1649, + 297, + 1649 + ], + "score": 0.979 + }, + { + "category_id": 1, + "poly": [ + 297, + 1819, + 1403, + 1819, + 1403, + 2039, + 297, + 2039 + ], + "score": 0.978 + }, + { + "category_id": 1, + "poly": [ + 397, + 541, + 1302, + 541, + 1302, + 695, + 397, + 695 + ], + "score": 0.977 + }, + { + "category_id": 0, + "poly": [ + 299, + 219, + 1122, + 219, + 1122, + 270, + 299, + 270 + ], + "score": 0.943 + }, + { + "category_id": 0, + "poly": [ + 300, + 1692, + 557, + 1692, + 557, + 1729, + 300, + 1729 + ], + "score": 0.908 + }, + { + "category_id": 0, + "poly": [ + 299, + 1761, + 491, + 1761, + 491, + 1794, + 299, + 1794 + ], + "score": 0.9 + }, + { + "category_id": 0, + "poly": [ + 302, + 753, + 573, + 753, + 573, + 788, + 302, + 788 + ], + "score": 0.897 + }, + { + "category_id": 2, + "poly": [ + 298, + 75, + 857, + 75, + 857, + 105, + 298, + 105 + ], + "score": 0.888 + }, + { + "category_id": 0, + "poly": [ + 773, + 472, + 927, + 472, + 927, + 506, + 773, + 506 + ], + "score": 0.879 + }, + { + "category_id": 1, + "poly": [ + 313, + 331, + 680, + 331, + 680, + 392, + 313, + 392 + ], + "score": 0.83 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 856, + 2088, + 856, + 2112, + 841, + 2112 + ], + "score": 0.726 + }, + { + "category_id": 0, + "poly": [ + 315, + 332, + 556, + 332, + 556, + 360, + 315, + 360 + ], + "score": 0.097 + }, + { + "category_id": 13, + "poly": [ + 938, + 2003, + 1102, + 2003, + 1102, + 2036, + 938, + 2036 + ], + "score": 0.93, + "latex": "r _ { t } = R ( s _ { t } , a _ { t } )" + }, + { + "category_id": 13, + "poly": [ + 1151, + 1942, + 1287, + 1942, + 1287, + 1976, + 1151, + 1976 + ], + "score": 0.92, + "latex": "r = R ( s , a )" + }, + { + "category_id": 13, + "poly": [ + 929, + 1944, + 1098, + 1944, + 1098, + 1975, + 929, + 1975 + ], + "score": 0.92, + "latex": "s ^ { \\prime } \\sim T ( s ^ { \\prime } | s , a )" + }, + { + "category_id": 13, + "poly": [ + 580, + 1912, + 693, + 1912, + 693, + 1945, + 580, + 1945 + ], + "score": 0.92, + "latex": "\\gamma \\in ( 0 , 1 ]" + }, + { + "category_id": 13, + "poly": [ + 673, + 1943, + 847, + 1943, + 847, + 1976, + 673, + 1976 + ], + "score": 0.92, + "latex": "\\tau = ( s , a , r , s ^ { \\prime } )" + }, + { + "category_id": 13, + "poly": [ + 655, + 1882, + 837, + 1882, + 837, + 1911, + 655, + 1911 + ], + "score": 0.9, + "latex": "R : S \\times \\mathcal { A } \\mapsto \\mathbb { R }" + }, + { + "category_id": 13, + "poly": [ + 296, + 1851, + 502, + 1851, + 502, + 1883, + 296, + 1883 + ], + "score": 0.9, + "latex": "( S , A , T , R , p _ { 0 } , \\gamma )" + }, + { + "category_id": 13, + "poly": [ + 742, + 543, + 779, + 543, + 779, + 575, + 742, + 575 + ], + "score": 0.9, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1260, + 2003, + 1296, + 2003, + 1296, + 2035, + 1260, + 2035 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1222, + 1237, + 1258, + 1237, + 1258, + 1267, + 1222, + 1267 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1204, + 1851, + 1404, + 1851, + 1404, + 1880, + 1204, + 1880 + ], + "score": 0.89, + "latex": "T : S \\times \\mathcal { A } \\times \\mathcal { S } \\mapsto" + }, + { + "category_id": 13, + "poly": [ + 298, + 1884, + 339, + 1884, + 339, + 1914, + 298, + 1914 + ], + "score": 0.89, + "latex": "\\mathbb { R } _ { + }" + }, + { + "category_id": 13, + "poly": [ + 1230, + 1146, + 1267, + 1146, + 1267, + 1176, + 1230, + 1176 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 297, + 1973, + 427, + 1973, + 427, + 2001, + 297, + 2001 + ], + "score": 0.89, + "latex": "\\pi : { \\mathcal { S } } \\mapsto A" + }, + { + "category_id": 13, + "poly": [ + 352, + 2001, + 857, + 2001, + 857, + 2039, + 352, + 2039 + ], + "score": 0.87, + "latex": "\\begin{array} { r } { Q ^ { \\pi } ( s , a ) : = \\mathbb E _ { \\pi } \\big [ \\sum _ { t = 0 } ^ { \\infty } \\gamma ^ { t } r _ { t } | s _ { 0 } = s , a _ { 0 } = a \\big ] } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 832, + 1497, + 857, + 1497, + 857, + 1527, + 832, + 1527 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 507, + 1466, + 532, + 1466, + 532, + 1497, + 507, + 1497 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 912, + 604, + 937, + 604, + 937, + 635, + 912, + 635 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 940, + 1406, + 965, + 1406, + 965, + 1436, + 940, + 1436 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 605, + 1558, + 630, + 1558, + 630, + 1588, + 605, + 1588 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1125, + 1268, + 1151, + 1268, + 1151, + 1298, + 1125, + 1298 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1156, + 1975, + 1180, + 1975, + 1180, + 2004, + 1156, + 2004 + ], + "score": 0.84, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 829, + 1852, + 851, + 1852, + 851, + 1878, + 829, + 1878 + ], + "score": 0.83, + "latex": "s" + }, + { + "category_id": 13, + "poly": [ + 1039, + 1853, + 1063, + 1853, + 1063, + 1878, + 1039, + 1878 + ], + "score": 0.83, + "latex": "\\mathcal { A }" + }, + { + "category_id": 13, + "poly": [ + 361, + 855, + 386, + 855, + 386, + 886, + 361, + 886 + ], + "score": 0.82, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 298, + 1175, + 322, + 1175, + 322, + 1206, + 298, + 1206 + ], + "score": 0.82, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 977, + 1917, + 1006, + 1917, + 1006, + 1943, + 977, + 1943 + ], + "score": 0.8, + "latex": "p _ { 0 }" + }, + { + "category_id": 13, + "poly": [ + 808, + 220, + 854, + 220, + 854, + 273, + 808, + 273 + ], + "score": 0.73, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 297, + 2008, + 318, + 2008, + 318, + 2031, + 297, + 2031 + ], + "score": 0.71, + "latex": "\\pi" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 217.0, + 807.0, + 217.0, + 807.0, + 277.0, + 293.0, + 277.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 855.0, + 217.0, + 1127.0, + 217.0, + 1127.0, + 277.0, + 855.0, + 277.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1689.0, + 561.0, + 1689.0, + 561.0, + 1736.0, + 291.0, + 1736.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1760.0, + 495.0, + 1760.0, + 495.0, + 1799.0, + 294.0, + 1799.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 750.0, + 579.0, + 750.0, + 579.0, + 797.0, + 294.0, + 797.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 298.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 770.0, + 471.0, + 932.0, + 471.0, + 932.0, + 509.0, + 770.0, + 509.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 840.0, + 2088.0, + 857.0, + 2088.0, + 857.0, + 2116.0, + 840.0, + 2116.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 311.0, + 325.0, + 562.0, + 325.0, + 562.0, + 369.0, + 311.0, + 369.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1054.0, + 1405.0, + 1054.0, + 1405.0, + 1088.0, + 294.0, + 1088.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1081.0, + 1407.0, + 1081.0, + 1407.0, + 1122.0, + 293.0, + 1122.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1111.0, + 1406.0, + 1111.0, + 1406.0, + 1150.0, + 293.0, + 1150.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1143.0, + 1229.0, + 1143.0, + 1229.0, + 1181.0, + 293.0, + 1181.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1268.0, + 1143.0, + 1406.0, + 1143.0, + 1406.0, + 1181.0, + 1268.0, + 1181.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1175.0, + 1405.0, + 1175.0, + 1405.0, + 1210.0, + 323.0, + 1210.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1205.0, + 1406.0, + 1205.0, + 1406.0, + 1243.0, + 293.0, + 1243.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1233.0, + 1221.0, + 1233.0, + 1221.0, + 1273.0, + 292.0, + 1273.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1259.0, + 1233.0, + 1408.0, + 1233.0, + 1408.0, + 1273.0, + 1259.0, + 1273.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1266.0, + 1124.0, + 1266.0, + 1124.0, + 1304.0, + 293.0, + 1304.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1152.0, + 1266.0, + 1405.0, + 1266.0, + 1405.0, + 1304.0, + 1152.0, + 1304.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1295.0, + 1405.0, + 1295.0, + 1405.0, + 1334.0, + 293.0, + 1334.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1325.0, + 1405.0, + 1325.0, + 1405.0, + 1360.0, + 296.0, + 1360.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1355.0, + 1244.0, + 1355.0, + 1244.0, + 1398.0, + 292.0, + 1398.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 822.0, + 1410.0, + 822.0, + 1410.0, + 860.0, + 294.0, + 860.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 853.0, + 360.0, + 853.0, + 360.0, + 891.0, + 293.0, + 891.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 387.0, + 853.0, + 1405.0, + 853.0, + 1405.0, + 891.0, + 387.0, + 891.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 881.0, + 1405.0, + 881.0, + 1405.0, + 923.0, + 292.0, + 923.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 916.0, + 1405.0, + 916.0, + 1405.0, + 950.0, + 294.0, + 950.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 946.0, + 1404.0, + 946.0, + 1404.0, + 979.0, + 293.0, + 979.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 977.0, + 1406.0, + 977.0, + 1406.0, + 1011.0, + 294.0, + 1011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1005.0, + 532.0, + 1005.0, + 532.0, + 1043.0, + 294.0, + 1043.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1405.0, + 939.0, + 1405.0, + 939.0, + 1438.0, + 295.0, + 1438.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 966.0, + 1405.0, + 1407.0, + 1405.0, + 1407.0, + 1438.0, + 966.0, + 1438.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1433.0, + 1404.0, + 1433.0, + 1404.0, + 1471.0, + 294.0, + 1471.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1466.0, + 506.0, + 1466.0, + 506.0, + 1500.0, + 295.0, + 1500.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 533.0, + 1466.0, + 1404.0, + 1466.0, + 1404.0, + 1500.0, + 533.0, + 1500.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1496.0, + 831.0, + 1496.0, + 831.0, + 1530.0, + 295.0, + 1530.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 858.0, + 1496.0, + 1402.0, + 1496.0, + 1402.0, + 1530.0, + 858.0, + 1530.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1526.0, + 1405.0, + 1526.0, + 1405.0, + 1562.0, + 292.0, + 1562.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1556.0, + 604.0, + 1556.0, + 604.0, + 1593.0, + 294.0, + 1593.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 631.0, + 1556.0, + 1405.0, + 1556.0, + 1405.0, + 1593.0, + 631.0, + 1593.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1585.0, + 1406.0, + 1585.0, + 1406.0, + 1623.0, + 291.0, + 1623.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1617.0, + 434.0, + 1617.0, + 434.0, + 1649.0, + 295.0, + 1649.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1818.0, + 1407.0, + 1818.0, + 1407.0, + 1854.0, + 294.0, + 1854.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 503.0, + 1849.0, + 828.0, + 1849.0, + 828.0, + 1884.0, + 503.0, + 1884.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 852.0, + 1849.0, + 1038.0, + 1849.0, + 1038.0, + 1884.0, + 852.0, + 1884.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1064.0, + 1849.0, + 1203.0, + 1849.0, + 1203.0, + 1884.0, + 1064.0, + 1884.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1879.0, + 297.0, + 1879.0, + 297.0, + 1915.0, + 294.0, + 1915.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 340.0, + 1879.0, + 654.0, + 1879.0, + 654.0, + 1915.0, + 340.0, + 1915.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 1879.0, + 1405.0, + 1879.0, + 1405.0, + 1915.0, + 838.0, + 1915.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1913.0, + 579.0, + 1913.0, + 579.0, + 1944.0, + 296.0, + 1944.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 694.0, + 1913.0, + 976.0, + 1913.0, + 976.0, + 1944.0, + 694.0, + 1944.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1007.0, + 1913.0, + 1404.0, + 1913.0, + 1404.0, + 1944.0, + 1007.0, + 1944.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1938.0, + 672.0, + 1938.0, + 672.0, + 1979.0, + 294.0, + 1979.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 848.0, + 1938.0, + 928.0, + 1938.0, + 928.0, + 1979.0, + 848.0, + 1979.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1099.0, + 1938.0, + 1150.0, + 1938.0, + 1150.0, + 1979.0, + 1099.0, + 1979.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1288.0, + 1938.0, + 1405.0, + 1938.0, + 1405.0, + 1979.0, + 1288.0, + 1979.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1969.0, + 296.0, + 1969.0, + 296.0, + 2010.0, + 291.0, + 2010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 428.0, + 1969.0, + 1155.0, + 1969.0, + 1155.0, + 2010.0, + 428.0, + 2010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1181.0, + 1969.0, + 1405.0, + 1969.0, + 1405.0, + 2010.0, + 1181.0, + 2010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 289.0, + 1996.0, + 296.0, + 1996.0, + 296.0, + 2047.0, + 289.0, + 2047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 319.0, + 1996.0, + 351.0, + 1996.0, + 351.0, + 2047.0, + 319.0, + 2047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 858.0, + 1996.0, + 937.0, + 1996.0, + 937.0, + 2047.0, + 858.0, + 2047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1103.0, + 1996.0, + 1259.0, + 1996.0, + 1259.0, + 2047.0, + 1103.0, + 2047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1297.0, + 1996.0, + 1409.0, + 1996.0, + 1409.0, + 2047.0, + 1297.0, + 2047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 395.0, + 541.0, + 741.0, + 541.0, + 741.0, + 576.0, + 395.0, + 576.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 780.0, + 541.0, + 1305.0, + 541.0, + 1305.0, + 576.0, + 780.0, + 576.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 571.0, + 1304.0, + 571.0, + 1304.0, + 607.0, + 394.0, + 607.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 601.0, + 911.0, + 601.0, + 911.0, + 639.0, + 393.0, + 639.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 938.0, + 601.0, + 1305.0, + 601.0, + 1305.0, + 639.0, + 938.0, + 639.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 633.0, + 1306.0, + 633.0, + 1306.0, + 670.0, + 394.0, + 670.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 665.0, + 925.0, + 665.0, + 925.0, + 697.0, + 394.0, + 697.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 315.0, + 331.0, + 560.0, + 331.0, + 560.0, + 364.0, + 315.0, + 364.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 311.0, + 361.0, + 682.0, + 361.0, + 682.0, + 394.0, + 311.0, + 394.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 0, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1283, + 1405, + 1283, + 1405, + 1680, + 298, + 1680 + ], + "score": 0.985 + }, + { + "category_id": 1, + "poly": [ + 298, + 854, + 1404, + 854, + 1404, + 1188, + 298, + 1188 + ], + "score": 0.984 + }, + { + "category_id": 1, + "poly": [ + 298, + 501, + 1406, + 501, + 1406, + 838, + 298, + 838 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 299, + 1772, + 1404, + 1772, + 1404, + 1957, + 299, + 1957 + ], + "score": 0.978 + }, + { + "category_id": 8, + "poly": [ + 573, + 362, + 1124, + 362, + 1124, + 410, + 573, + 410 + ], + "score": 0.946 + }, + { + "category_id": 1, + "poly": [ + 299, + 1972, + 1401, + 1972, + 1401, + 2034, + 299, + 2034 + ], + "score": 0.94 + }, + { + "category_id": 1, + "poly": [ + 299, + 326, + 682, + 326, + 682, + 356, + 299, + 356 + ], + "score": 0.925 + }, + { + "category_id": 2, + "poly": [ + 297, + 74, + 857, + 74, + 857, + 106, + 297, + 106 + ], + "score": 0.918 + }, + { + "category_id": 1, + "poly": [ + 298, + 230, + 748, + 230, + 748, + 262, + 298, + 262 + ], + "score": 0.917 + }, + { + "category_id": 8, + "poly": [ + 703, + 268, + 995, + 268, + 995, + 319, + 703, + 319 + ], + "score": 0.914 + }, + { + "category_id": 0, + "poly": [ + 299, + 446, + 914, + 446, + 914, + 476, + 299, + 476 + ], + "score": 0.906 + }, + { + "category_id": 0, + "poly": [ + 300, + 1715, + 628, + 1715, + 628, + 1748, + 300, + 1748 + ], + "score": 0.906 + }, + { + "category_id": 0, + "poly": [ + 301, + 1226, + 831, + 1226, + 831, + 1256, + 301, + 1256 + ], + "score": 0.895 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 858, + 2088, + 858, + 2112, + 841, + 2112 + ], + "score": 0.701 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 859, + 2088, + 859, + 2112, + 841, + 2112 + ], + "score": 0.172 + }, + { + "category_id": 14, + "poly": [ + 703, + 267, + 995, + 267, + 995, + 321, + 703, + 321 + ], + "score": 0.92, + "latex": "Q ^ { \\ast } ( s , a ) : = \\operatorname* { s u p } _ { \\pi } Q ^ { \\pi } ( s , a )" + }, + { + "category_id": 13, + "poly": [ + 914, + 1528, + 952, + 1528, + 952, + 1558, + 914, + 1558 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 14, + "poly": [ + 573, + 360, + 1124, + 360, + 1124, + 413, + 573, + 413 + ], + "score": 0.89, + "latex": "Q ^ { * } ( s , a ) = \\mathbb { E } _ { s ^ { \\prime } \\sim T ( \\cdot \\mid s , a ) } \\big [ r + \\gamma \\cdot \\operatorname* { m a x } _ { a ^ { \\prime } } Q ^ { * } ( s ^ { \\prime } , a ^ { \\prime } ) \\big ] ." + }, + { + "category_id": 13, + "poly": [ + 1360, + 1926, + 1398, + 1926, + 1398, + 1958, + 1360, + 1958 + ], + "score": 0.89, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 929, + 1436, + 967, + 1436, + 967, + 1467, + 929, + 1467 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1359, + 1314, + 1395, + 1314, + 1395, + 1345, + 1359, + 1345 + ], + "score": 0.88, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1061, + 1835, + 1086, + 1835, + 1086, + 1866, + 1061, + 1866 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 690, + 1974, + 714, + 1974, + 714, + 2004, + 690, + 2004 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1230, + 1926, + 1255, + 1926, + 1255, + 1958, + 1230, + 1958 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1127, + 1866, + 1152, + 1866, + 1152, + 1896, + 1127, + 1896 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 627, + 1314, + 652, + 1314, + 652, + 1346, + 627, + 1346 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 966, + 535, + 991, + 535, + 991, + 565, + 966, + 565 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 718, + 778, + 742, + 778, + 742, + 809, + 718, + 809 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1274, + 2003, + 1298, + 2003, + 1298, + 2035, + 1274, + 2035 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 436, + 778, + 461, + 778, + 461, + 809, + 436, + 809 + ], + "score": 0.84, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 870, + 1895, + 970, + 1895, + 970, + 1928, + 870, + 1928 + ], + "score": 0.73, + "latex": "K \\in \\mathbb { N } _ { + }" + }, + { + "category_id": 13, + "poly": [ + 1120, + 891, + 1135, + 891, + 1135, + 912, + 1120, + 912 + ], + "score": 0.54, + "latex": "\\epsilon" + }, + { + "category_id": 13, + "poly": [ + 975, + 1896, + 999, + 1896, + 999, + 1927, + 975, + 1927 + ], + "score": 0.44, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1381, + 600, + 1396, + 600, + 1396, + 621, + 1381, + 621 + ], + "score": 0.38, + "latex": "\\epsilon" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 72.0, + 859.0, + 72.0, + 859.0, + 109.0, + 297.0, + 109.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 445.0, + 916.0, + 445.0, + 916.0, + 480.0, + 295.0, + 480.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1715.0, + 631.0, + 1715.0, + 631.0, + 1751.0, + 295.0, + 1751.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1226.0, + 836.0, + 1226.0, + 836.0, + 1261.0, + 294.0, + 1261.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2085.0, + 862.0, + 2085.0, + 862.0, + 2120.0, + 839.0, + 2120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2085.0, + 862.0, + 2085.0, + 862.0, + 2120.0, + 838.0, + 2120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1285.0, + 1405.0, + 1285.0, + 1405.0, + 1317.0, + 296.0, + 1317.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1313.0, + 626.0, + 1313.0, + 626.0, + 1349.0, + 295.0, + 1349.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 653.0, + 1313.0, + 1358.0, + 1313.0, + 1358.0, + 1349.0, + 653.0, + 1349.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1396.0, + 1313.0, + 1407.0, + 1313.0, + 1407.0, + 1349.0, + 1396.0, + 1349.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1340.0, + 1406.0, + 1340.0, + 1406.0, + 1384.0, + 292.0, + 1384.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1374.0, + 1407.0, + 1374.0, + 1407.0, + 1410.0, + 295.0, + 1410.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1405.0, + 1406.0, + 1405.0, + 1406.0, + 1441.0, + 293.0, + 1441.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1435.0, + 928.0, + 1435.0, + 928.0, + 1471.0, + 295.0, + 1471.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 968.0, + 1435.0, + 1405.0, + 1435.0, + 1405.0, + 1471.0, + 968.0, + 1471.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1466.0, + 1406.0, + 1466.0, + 1406.0, + 1502.0, + 295.0, + 1502.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1496.0, + 1406.0, + 1496.0, + 1406.0, + 1532.0, + 292.0, + 1532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1527.0, + 913.0, + 1527.0, + 913.0, + 1563.0, + 295.0, + 1563.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 953.0, + 1527.0, + 1406.0, + 1527.0, + 1406.0, + 1563.0, + 953.0, + 1563.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1557.0, + 1405.0, + 1557.0, + 1405.0, + 1593.0, + 295.0, + 1593.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1588.0, + 1405.0, + 1588.0, + 1405.0, + 1624.0, + 295.0, + 1624.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1617.0, + 1408.0, + 1617.0, + 1408.0, + 1655.0, + 293.0, + 1655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1650.0, + 1276.0, + 1650.0, + 1276.0, + 1682.0, + 296.0, + 1682.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 850.0, + 1409.0, + 850.0, + 1409.0, + 893.0, + 292.0, + 893.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 885.0, + 1119.0, + 885.0, + 1119.0, + 920.0, + 296.0, + 920.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1136.0, + 885.0, + 1405.0, + 885.0, + 1405.0, + 920.0, + 1136.0, + 920.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 914.0, + 1405.0, + 914.0, + 1405.0, + 951.0, + 292.0, + 951.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 942.0, + 1406.0, + 942.0, + 1406.0, + 984.0, + 292.0, + 984.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 975.0, + 1406.0, + 975.0, + 1406.0, + 1010.0, + 293.0, + 1010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1006.0, + 1406.0, + 1006.0, + 1406.0, + 1041.0, + 294.0, + 1041.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1037.0, + 1405.0, + 1037.0, + 1405.0, + 1072.0, + 294.0, + 1072.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1067.0, + 1406.0, + 1067.0, + 1406.0, + 1104.0, + 292.0, + 1104.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1099.0, + 1404.0, + 1099.0, + 1404.0, + 1133.0, + 294.0, + 1133.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1130.0, + 1405.0, + 1130.0, + 1405.0, + 1164.0, + 294.0, + 1164.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1160.0, + 792.0, + 1160.0, + 792.0, + 1190.0, + 294.0, + 1190.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 501.0, + 1404.0, + 501.0, + 1404.0, + 538.0, + 293.0, + 538.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 532.0, + 965.0, + 532.0, + 965.0, + 569.0, + 293.0, + 569.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 992.0, + 532.0, + 1406.0, + 532.0, + 1406.0, + 569.0, + 992.0, + 569.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 564.0, + 1409.0, + 564.0, + 1409.0, + 599.0, + 295.0, + 599.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 594.0, + 1380.0, + 594.0, + 1380.0, + 629.0, + 295.0, + 629.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1397.0, + 594.0, + 1409.0, + 594.0, + 1409.0, + 629.0, + 1397.0, + 629.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 622.0, + 1407.0, + 622.0, + 1407.0, + 663.0, + 293.0, + 663.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 656.0, + 1406.0, + 656.0, + 1406.0, + 691.0, + 296.0, + 691.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 686.0, + 1404.0, + 686.0, + 1404.0, + 720.0, + 295.0, + 720.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 717.0, + 1406.0, + 717.0, + 1406.0, + 751.0, + 295.0, + 751.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 745.0, + 1408.0, + 745.0, + 1408.0, + 781.0, + 292.0, + 781.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 778.0, + 435.0, + 778.0, + 435.0, + 812.0, + 295.0, + 812.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 462.0, + 778.0, + 717.0, + 778.0, + 717.0, + 812.0, + 462.0, + 812.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 743.0, + 778.0, + 1406.0, + 778.0, + 1406.0, + 812.0, + 743.0, + 812.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 810.0, + 808.0, + 810.0, + 808.0, + 841.0, + 296.0, + 841.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1773.0, + 1408.0, + 1773.0, + 1408.0, + 1809.0, + 295.0, + 1809.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1803.0, + 1406.0, + 1803.0, + 1406.0, + 1839.0, + 294.0, + 1839.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1833.0, + 1060.0, + 1833.0, + 1060.0, + 1872.0, + 293.0, + 1872.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1087.0, + 1833.0, + 1406.0, + 1833.0, + 1406.0, + 1872.0, + 1087.0, + 1872.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1864.0, + 1126.0, + 1864.0, + 1126.0, + 1900.0, + 294.0, + 1900.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1153.0, + 1864.0, + 1406.0, + 1864.0, + 1406.0, + 1900.0, + 1153.0, + 1900.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1892.0, + 869.0, + 1892.0, + 869.0, + 1928.0, + 294.0, + 1928.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 971.0, + 1892.0, + 974.0, + 1892.0, + 974.0, + 1928.0, + 971.0, + 1928.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1000.0, + 1892.0, + 1405.0, + 1892.0, + 1405.0, + 1928.0, + 1000.0, + 1928.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1923.0, + 1229.0, + 1923.0, + 1229.0, + 1961.0, + 293.0, + 1961.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1256.0, + 1923.0, + 1359.0, + 1923.0, + 1359.0, + 1961.0, + 1256.0, + 1961.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1399.0, + 1923.0, + 1408.0, + 1923.0, + 1408.0, + 1961.0, + 1399.0, + 1961.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1971.0, + 689.0, + 1971.0, + 689.0, + 2007.0, + 295.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 715.0, + 1971.0, + 1405.0, + 1971.0, + 1405.0, + 2007.0, + 715.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 2002.0, + 1273.0, + 2002.0, + 1273.0, + 2037.0, + 295.0, + 2037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1299.0, + 2002.0, + 1406.0, + 2002.0, + 1406.0, + 2037.0, + 1299.0, + 2037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 324.0, + 682.0, + 324.0, + 682.0, + 361.0, + 295.0, + 361.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 223.0, + 750.0, + 223.0, + 750.0, + 272.0, + 294.0, + 272.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 1, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1466, + 1404, + 1466, + 1404, + 1805, + 297, + 1805 + ], + "score": 0.986 + }, + { + "category_id": 1, + "poly": [ + 297, + 1819, + 1404, + 1819, + 1404, + 2036, + 297, + 2036 + ], + "score": 0.98 + }, + { + "category_id": 1, + "poly": [ + 297, + 967, + 1405, + 967, + 1405, + 1123, + 297, + 1123 + ], + "score": 0.978 + }, + { + "category_id": 1, + "poly": [ + 297, + 690, + 1405, + 690, + 1405, + 816, + 297, + 816 + ], + "score": 0.978 + }, + { + "category_id": 1, + "poly": [ + 297, + 434, + 1407, + 434, + 1407, + 561, + 297, + 561 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 298, + 1134, + 1404, + 1134, + 1404, + 1232, + 298, + 1232 + ], + "score": 0.972 + }, + { + "category_id": 1, + "poly": [ + 299, + 1358, + 1404, + 1358, + 1404, + 1454, + 299, + 1454 + ], + "score": 0.97 + }, + { + "category_id": 1, + "poly": [ + 291, + 228, + 1403, + 228, + 1403, + 297, + 291, + 297 + ], + "score": 0.949 + }, + { + "category_id": 8, + "poly": [ + 560, + 332, + 1139, + 332, + 1139, + 400, + 560, + 400 + ], + "score": 0.949 + }, + { + "category_id": 8, + "poly": [ + 552, + 596, + 1143, + 596, + 1143, + 653, + 552, + 653 + ], + "score": 0.947 + }, + { + "category_id": 8, + "poly": [ + 590, + 1266, + 1105, + 1266, + 1105, + 1324, + 590, + 1324 + ], + "score": 0.932 + }, + { + "category_id": 2, + "poly": [ + 297, + 74, + 857, + 74, + 857, + 106, + 297, + 106 + ], + "score": 0.925 + }, + { + "category_id": 0, + "poly": [ + 300, + 883, + 632, + 883, + 632, + 919, + 300, + 919 + ], + "score": 0.917 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 859, + 2088, + 859, + 2112, + 841, + 2112 + ], + "score": 0.651 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 859, + 2088, + 859, + 2112, + 841, + 2112 + ], + "score": 0.432 + }, + { + "category_id": 13, + "poly": [ + 887, + 527, + 999, + 527, + 999, + 561, + 887, + 561 + ], + "score": 0.93, + "latex": "Q _ { k } \\big ( \\cdot ; \\theta ^ { - } \\big )" + }, + { + "category_id": 13, + "poly": [ + 611, + 1135, + 718, + 1135, + 718, + 1172, + 611, + 1172 + ], + "score": 0.93, + "latex": "\\{ Q _ { k } \\} _ { k = 1 } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 1136, + 1468, + 1201, + 1468, + 1201, + 1502, + 1136, + 1502 + ], + "score": 0.93, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 1199, + 498, + 1300, + 498, + 1300, + 531, + 1199, + 531 + ], + "score": 0.93, + "latex": "Q _ { k } ( \\cdot ; \\theta _ { t } )" + }, + { + "category_id": 13, + "poly": [ + 1103, + 693, + 1291, + 693, + 1291, + 726, + 1103, + 726 + ], + "score": 0.92, + "latex": "m _ { \\tau } = ( 1 , \\ldots , 1 )" + }, + { + "category_id": 13, + "poly": [ + 454, + 259, + 569, + 259, + 569, + 296, + 454, + 296 + ], + "score": 0.92, + "latex": "m _ { \\tau } \\in \\mathbb { R } _ { + } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 369, + 434, + 415, + 434, + 415, + 470, + 369, + 470 + ], + "score": 0.92, + "latex": "y _ { \\tau } ^ { Q _ { k } }" + }, + { + "category_id": 13, + "poly": [ + 334, + 723, + 398, + 723, + 398, + 757, + 334, + 757 + ], + "score": 0.92, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 1335, + 1529, + 1400, + 1529, + 1400, + 1564, + 1335, + 1564 + ], + "score": 0.92, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 14, + "poly": [ + 553, + 595, + 1144, + 595, + 1144, + 654, + 553, + 654 + ], + "score": 0.91, + "latex": "y _ { \\tau } ^ { Q _ { k } } = r + \\gamma \\operatorname* { m a x } _ { a } Q ^ { k } ( s ^ { \\prime } , \\operatorname * { a r g m a x } _ { a } Q _ { k } ( s ^ { \\prime } , a ; \\theta _ { t } ) ; \\theta ^ { - } ) ." + }, + { + "category_id": 14, + "poly": [ + 558, + 332, + 1137, + 332, + 1137, + 398, + 558, + 398 + ], + "score": 0.91, + "latex": "L = \\sum _ { \\tau \\in { \\cal B } _ { \\mathrm { m i n i } } } \\sum _ { k = 1 } ^ { K } m _ { \\tau } ^ { k } \\cdot ( Q ^ { k } ( s , a ; \\theta ) - y _ { \\tau } ^ { Q _ { k } } ) ^ { 2 } ," + }, + { + "category_id": 13, + "poly": [ + 1241, + 435, + 1281, + 435, + 1281, + 469, + 1241, + 469 + ], + "score": 0.91, + "latex": "m _ { \\tau } ^ { k }" + }, + { + "category_id": 13, + "poly": [ + 1285, + 465, + 1331, + 465, + 1331, + 499, + 1285, + 499 + ], + "score": 0.91, + "latex": "y _ { \\tau } ^ { Q _ { k } }" + }, + { + "category_id": 14, + "poly": [ + 591, + 1266, + 1107, + 1266, + 1107, + 1324, + 591, + 1324 + ], + "score": 0.9, + "latex": "a _ { t } = \\mathop { \\mathrm { M a j o r i t y } } \\mathrm { V o t e } ( \\{ \\operatorname { a r g m a x } Q _ { k } ( s _ { t } , a ) \\} _ { k = 1 } ^ { K } ) ." + }, + { + "category_id": 13, + "poly": [ + 539, + 1199, + 577, + 1199, + 577, + 1231, + 539, + 1231 + ], + "score": 0.9, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 1122, + 438, + 1160, + 438, + 1160, + 468, + 1122, + 468 + ], + "score": 0.9, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 298, + 785, + 576, + 785, + 576, + 817, + 298, + 817 + ], + "score": 0.9, + "latex": "a _ { t } = \\operatorname { a r g m a x } _ { a } Q _ { k } ( s _ { t } , a )" + }, + { + "category_id": 13, + "poly": [ + 786, + 1061, + 823, + 1061, + 823, + 1092, + 786, + 1092 + ], + "score": 0.9, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 359, + 1138, + 397, + 1138, + 397, + 1170, + 359, + 1170 + ], + "score": 0.9, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 934, + 1169, + 973, + 1169, + 973, + 1200, + 934, + 1200 + ], + "score": 0.89, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 593, + 1030, + 630, + 1030, + 630, + 1061, + 593, + 1061 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 394, + 1422, + 433, + 1422, + 433, + 1453, + 394, + 1453 + ], + "score": 0.89, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 1257, + 1592, + 1296, + 1592, + 1296, + 1623, + 1257, + 1623 + ], + "score": 0.89, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 1362, + 1361, + 1400, + 1361, + 1400, + 1393, + 1362, + 1393 + ], + "score": 0.88, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 483, + 754, + 522, + 754, + 522, + 785, + 483, + 785 + ], + "score": 0.88, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 1363, + 1851, + 1400, + 1851, + 1400, + 1882, + 1363, + 1882 + ], + "score": 0.88, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 469, + 1139, + 495, + 1139, + 495, + 1169, + 469, + 1169 + ], + "score": 0.87, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1262, + 1501, + 1287, + 1501, + 1287, + 1531, + 1262, + 1531 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 990, + 1470, + 1016, + 1470, + 1016, + 1500, + 990, + 1500 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 996, + 1031, + 1021, + 1031, + 1021, + 1061, + 996, + 1061 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1188, + 1531, + 1214, + 1531, + 1214, + 1562, + 1188, + 1562 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 661, + 438, + 685, + 438, + 685, + 468, + 661, + 468 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 298, + 1199, + 322, + 1199, + 322, + 1231, + 298, + 1231 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 616, + 1562, + 641, + 1562, + 641, + 1592, + 616, + 1592 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 368, + 232, + 393, + 232, + 393, + 263, + 368, + 263 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1011, + 443, + 1030, + 443, + 1030, + 464, + 1011, + 464 + ], + "score": 0.78, + "latex": "\\tau" + }, + { + "category_id": 13, + "poly": [ + 618, + 438, + 635, + 438, + 635, + 465, + 618, + 465 + ], + "score": 0.49, + "latex": "k" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 72.0, + 859.0, + 72.0, + 859.0, + 109.0, + 297.0, + 109.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 879.0, + 637.0, + 879.0, + 637.0, + 925.0, + 291.0, + 925.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2085.0, + 862.0, + 2085.0, + 862.0, + 2118.0, + 838.0, + 2118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2085.0, + 862.0, + 2085.0, + 862.0, + 2118.0, + 838.0, + 2118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1467.0, + 989.0, + 1467.0, + 989.0, + 1506.0, + 294.0, + 1506.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1017.0, + 1467.0, + 1135.0, + 1467.0, + 1135.0, + 1506.0, + 1017.0, + 1506.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1202.0, + 1467.0, + 1405.0, + 1467.0, + 1405.0, + 1506.0, + 1202.0, + 1506.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1498.0, + 1261.0, + 1498.0, + 1261.0, + 1533.0, + 295.0, + 1533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1288.0, + 1498.0, + 1404.0, + 1498.0, + 1404.0, + 1533.0, + 1288.0, + 1533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1526.0, + 1187.0, + 1526.0, + 1187.0, + 1567.0, + 292.0, + 1567.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1215.0, + 1526.0, + 1334.0, + 1526.0, + 1334.0, + 1567.0, + 1215.0, + 1567.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1401.0, + 1526.0, + 1404.0, + 1526.0, + 1404.0, + 1567.0, + 1401.0, + 1567.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1561.0, + 615.0, + 1561.0, + 615.0, + 1596.0, + 295.0, + 1596.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 642.0, + 1561.0, + 1405.0, + 1561.0, + 1405.0, + 1596.0, + 642.0, + 1596.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1589.0, + 1256.0, + 1589.0, + 1256.0, + 1628.0, + 294.0, + 1628.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1297.0, + 1589.0, + 1406.0, + 1589.0, + 1406.0, + 1628.0, + 1297.0, + 1628.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1619.0, + 1407.0, + 1619.0, + 1407.0, + 1658.0, + 294.0, + 1658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1652.0, + 1404.0, + 1652.0, + 1404.0, + 1687.0, + 295.0, + 1687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1679.0, + 1405.0, + 1679.0, + 1405.0, + 1720.0, + 292.0, + 1720.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1713.0, + 1406.0, + 1713.0, + 1406.0, + 1748.0, + 295.0, + 1748.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1740.0, + 1406.0, + 1740.0, + 1406.0, + 1781.0, + 291.0, + 1781.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1773.0, + 818.0, + 1773.0, + 818.0, + 1809.0, + 294.0, + 1809.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1821.0, + 1405.0, + 1821.0, + 1405.0, + 1856.0, + 295.0, + 1856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1852.0, + 1362.0, + 1852.0, + 1362.0, + 1885.0, + 293.0, + 1885.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1880.0, + 1407.0, + 1880.0, + 1407.0, + 1919.0, + 294.0, + 1919.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1909.0, + 1405.0, + 1909.0, + 1405.0, + 1949.0, + 292.0, + 1949.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1941.0, + 1405.0, + 1941.0, + 1405.0, + 1980.0, + 294.0, + 1980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1973.0, + 1404.0, + 1973.0, + 1404.0, + 2008.0, + 295.0, + 2008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 2003.0, + 732.0, + 2003.0, + 732.0, + 2041.0, + 294.0, + 2041.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 970.0, + 1406.0, + 970.0, + 1406.0, + 1004.0, + 296.0, + 1004.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 998.0, + 1406.0, + 998.0, + 1406.0, + 1034.0, + 292.0, + 1034.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1030.0, + 592.0, + 1030.0, + 592.0, + 1064.0, + 295.0, + 1064.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 631.0, + 1030.0, + 995.0, + 1030.0, + 995.0, + 1064.0, + 631.0, + 1064.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1022.0, + 1030.0, + 1405.0, + 1030.0, + 1405.0, + 1064.0, + 1022.0, + 1064.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1058.0, + 785.0, + 1058.0, + 785.0, + 1097.0, + 292.0, + 1097.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 824.0, + 1058.0, + 1410.0, + 1058.0, + 1410.0, + 1097.0, + 824.0, + 1097.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1089.0, + 680.0, + 1089.0, + 680.0, + 1125.0, + 294.0, + 1125.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 693.0, + 1102.0, + 693.0, + 1102.0, + 726.0, + 295.0, + 726.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1292.0, + 693.0, + 1405.0, + 693.0, + 1405.0, + 726.0, + 1292.0, + 726.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 722.0, + 333.0, + 722.0, + 333.0, + 759.0, + 295.0, + 759.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 399.0, + 722.0, + 1406.0, + 722.0, + 1406.0, + 759.0, + 399.0, + 759.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 753.0, + 482.0, + 753.0, + 482.0, + 789.0, + 294.0, + 789.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 523.0, + 753.0, + 1405.0, + 753.0, + 1405.0, + 789.0, + 523.0, + 789.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 785.0, + 297.0, + 785.0, + 297.0, + 818.0, + 292.0, + 818.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 577.0, + 785.0, + 833.0, + 785.0, + 833.0, + 818.0, + 577.0, + 818.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 432.0, + 368.0, + 432.0, + 368.0, + 473.0, + 292.0, + 473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 416.0, + 432.0, + 617.0, + 432.0, + 617.0, + 473.0, + 416.0, + 473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 636.0, + 432.0, + 660.0, + 432.0, + 660.0, + 473.0, + 636.0, + 473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 686.0, + 432.0, + 1010.0, + 432.0, + 1010.0, + 473.0, + 686.0, + 473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1031.0, + 432.0, + 1121.0, + 432.0, + 1121.0, + 473.0, + 1031.0, + 473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1161.0, + 432.0, + 1240.0, + 432.0, + 1240.0, + 473.0, + 1161.0, + 473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1282.0, + 432.0, + 1409.0, + 432.0, + 1409.0, + 473.0, + 1282.0, + 473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 461.0, + 1284.0, + 461.0, + 1284.0, + 505.0, + 291.0, + 505.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1332.0, + 461.0, + 1408.0, + 461.0, + 1408.0, + 505.0, + 1332.0, + 505.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 496.0, + 1198.0, + 496.0, + 1198.0, + 533.0, + 294.0, + 533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1301.0, + 496.0, + 1407.0, + 496.0, + 1407.0, + 533.0, + 1301.0, + 533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 528.0, + 886.0, + 528.0, + 886.0, + 561.0, + 295.0, + 561.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1000.0, + 528.0, + 1220.0, + 528.0, + 1220.0, + 561.0, + 1000.0, + 561.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 287.0, + 1127.0, + 358.0, + 1127.0, + 358.0, + 1185.0, + 287.0, + 1185.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 398.0, + 1127.0, + 468.0, + 1127.0, + 468.0, + 1185.0, + 398.0, + 1185.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 496.0, + 1127.0, + 610.0, + 1127.0, + 610.0, + 1185.0, + 496.0, + 1185.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 719.0, + 1127.0, + 1413.0, + 1127.0, + 1413.0, + 1185.0, + 719.0, + 1185.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1167.0, + 933.0, + 1167.0, + 933.0, + 1204.0, + 293.0, + 1204.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 974.0, + 1167.0, + 1405.0, + 1167.0, + 1405.0, + 1204.0, + 974.0, + 1204.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1199.0, + 538.0, + 1199.0, + 538.0, + 1234.0, + 323.0, + 1234.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 578.0, + 1199.0, + 1356.0, + 1199.0, + 1356.0, + 1234.0, + 578.0, + 1234.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1355.0, + 1361.0, + 1355.0, + 1361.0, + 1399.0, + 293.0, + 1399.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1391.0, + 1405.0, + 1391.0, + 1405.0, + 1426.0, + 294.0, + 1426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1419.0, + 393.0, + 1419.0, + 393.0, + 1456.0, + 294.0, + 1456.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 434.0, + 1419.0, + 900.0, + 1419.0, + 900.0, + 1456.0, + 434.0, + 1456.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 228.0, + 367.0, + 228.0, + 367.0, + 265.0, + 295.0, + 265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 228.0, + 1406.0, + 228.0, + 1406.0, + 265.0, + 394.0, + 265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 288.0, + 247.0, + 453.0, + 247.0, + 453.0, + 306.0, + 288.0, + 306.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 570.0, + 247.0, + 582.0, + 247.0, + 582.0, + 306.0, + 570.0, + 306.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 2, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 296, + 738, + 1406, + 738, + 1406, + 955, + 296, + 955 + ], + "score": 0.982 + }, + { + "category_id": 1, + "poly": [ + 297, + 1147, + 1406, + 1147, + 1406, + 1246, + 297, + 1246 + ], + "score": 0.975 + }, + { + "category_id": 1, + "poly": [ + 298, + 1366, + 1405, + 1366, + 1405, + 1460, + 298, + 1460 + ], + "score": 0.973 + }, + { + "category_id": 1, + "poly": [ + 297, + 1069, + 1400, + 1069, + 1400, + 1134, + 297, + 1134 + ], + "score": 0.954 + }, + { + "category_id": 0, + "poly": [ + 300, + 669, + 1127, + 669, + 1127, + 707, + 300, + 707 + ], + "score": 0.932 + }, + { + "category_id": 1, + "poly": [ + 298, + 1320, + 693, + 1320, + 693, + 1353, + 298, + 1353 + ], + "score": 0.926 + }, + { + "category_id": 2, + "poly": [ + 298, + 75, + 857, + 75, + 857, + 105, + 298, + 105 + ], + "score": 0.91 + }, + { + "category_id": 8, + "poly": [ + 627, + 1256, + 1071, + 1256, + 1071, + 1306, + 627, + 1306 + ], + "score": 0.909 + }, + { + "category_id": 0, + "poly": [ + 299, + 1934, + 541, + 1934, + 541, + 1970, + 299, + 1970 + ], + "score": 0.909 + }, + { + "category_id": 1, + "poly": [ + 296, + 2002, + 1127, + 2002, + 1127, + 2035, + 296, + 2035 + ], + "score": 0.906 + }, + { + "category_id": 8, + "poly": [ + 540, + 966, + 1158, + 966, + 1158, + 1054, + 540, + 1054 + ], + "score": 0.906 + }, + { + "category_id": 2, + "poly": [ + 840, + 2089, + 857, + 2089, + 857, + 2111, + 840, + 2111 + ], + "score": 0.79 + }, + { + "category_id": 1, + "poly": [ + 306, + 268, + 1275, + 268, + 1275, + 614, + 306, + 614 + ], + "score": 0.775 + }, + { + "category_id": 1, + "poly": [ + 307, + 1531, + 1263, + 1531, + 1263, + 1876, + 307, + 1876 + ], + "score": 0.752 + }, + { + "category_id": 0, + "poly": [ + 298, + 229, + 643, + 229, + 643, + 262, + 298, + 262 + ], + "score": 0.689 + }, + { + "category_id": 5, + "poly": [ + 306, + 268, + 1275, + 268, + 1275, + 614, + 306, + 614 + ], + "score": 0.319, + "html": "
2:1: Input: K ∈ N+ copies of independently initialized Q*-functions {Qk}k=1· Let B be a replay buffer storing transitions for training
for each episode do do
3: 4:Obtain initial state from environment so
5:for step t = 1,... until end of episode do
6: Pick an action aording toat = MajorityVote({argmaxq Qk(st,α)}-1)
7:Execute at. Receive state St+1 and reward rt from the environment
8:Add (st, at, rt, St+1) to replay buffer B
9:At learning interval, sample random minibatch and update {Qk }
10: 11: end forend for
" + }, + { + "category_id": 5, + "poly": [ + 307, + 1531, + 1263, + 1531, + 1263, + 1876, + 307, + 1876 + ], + "score": 0.303, + "html": "
2:
Let B be a replay buffer storing experience for training.
3: for each episode do
Obtain initial state from environment so
for step t = 1,... until end of episode do
Pick an action according to at ∈ argmaxa {μ(St,a) + λ · δ(st, a)}
6: 7: Receive state St+1 and reward rt from environment, having taken action ät
8: Add (St,at, rt, St+1) to replay buffer B
9: At learning interval, sample random minibatch and update {Qk}
10: end for
11: end for
" + }, + { + "category_id": 8, + "poly": [ + 541, + 966, + 1158, + 966, + 1158, + 1054, + 541, + 1054 + ], + "score": 0.137 + }, + { + "category_id": 0, + "poly": [ + 296, + 1491, + 862, + 1491, + 862, + 1525, + 296, + 1525 + ], + "score": 0.12 + }, + { + "category_id": 13, + "poly": [ + 1024, + 1179, + 1114, + 1179, + 1114, + 1212, + 1024, + 1212 + ], + "score": 0.93, + "latex": "\\tilde { \\sigma } ( s _ { t } , a )" + }, + { + "category_id": 14, + "poly": [ + 539, + 963, + 1158, + 963, + 1158, + 1056, + 539, + 1056 + ], + "score": 0.93, + "latex": "A _ { t } = \\underset { i } { \\operatorname { a r g m a x } } \\left\\{ \\hat { r } _ { i , t } + c _ { 1 } \\cdot \\sqrt { \\frac { \\hat { V } _ { i , t } \\log ( t ) } { n _ { i , t } } } + c _ { 2 } \\cdot \\frac { \\log ( t ) } { n _ { i , t } } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 602, + 1210, + 779, + 1210, + 779, + 1248, + 602, + 1248 + ], + "score": 0.93, + "latex": "\\{ Q _ { k } ( s _ { t } , a ) \\} _ { k = 1 } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 476, + 1212, + 566, + 1212, + 566, + 1246, + 476, + 1246 + ], + "score": 0.93, + "latex": "\\tilde { \\mu } ( s _ { t } , a )" + }, + { + "category_id": 14, + "poly": [ + 626, + 1254, + 1069, + 1254, + 1069, + 1312, + 626, + 1312 + ], + "score": 0.93, + "latex": "a _ { t } \\in \\mathop { \\operatorname { a r g m a x } } _ { a } \\left\\{ \\tilde { \\mu } ( s _ { t } , a ) + \\lambda \\cdot \\tilde { \\sigma } ( s _ { t } , a ) \\right\\} ," + }, + { + "category_id": 13, + "poly": [ + 466, + 1065, + 509, + 1065, + 509, + 1104, + 466, + 1104 + ], + "score": 0.92, + "latex": "\\hat { V } _ { i , t }" + }, + { + "category_id": 13, + "poly": [ + 371, + 1321, + 462, + 1321, + 462, + 1353, + 371, + 1353 + ], + "score": 0.92, + "latex": "\\lambda \\in \\mathbb { R } _ { + }" + }, + { + "category_id": 13, + "poly": [ + 912, + 1531, + 1018, + 1531, + 1018, + 1568, + 912, + 1568 + ], + "score": 0.91, + "latex": "\\{ Q _ { k } \\} _ { k = 1 } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 1152, + 1179, + 1328, + 1179, + 1328, + 1213, + 1152, + 1213 + ], + "score": 0.91, + "latex": "\\{ Q _ { k } ( s _ { t } , a ) \\} _ { k = 1 } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 483, + 485, + 663, + 485, + 663, + 518, + 483, + 518 + ], + "score": 0.91, + "latex": "\\left( { { s _ { t } } , { a _ { t } } , { r _ { t } } , { s _ { t + 1 } } } \\right)" + }, + { + "category_id": 13, + "poly": [ + 427, + 272, + 528, + 272, + 528, + 304, + 427, + 304 + ], + "score": 0.9, + "latex": "K \\in \\mathbb { N } _ { + }" + }, + { + "category_id": 13, + "poly": [ + 1075, + 269, + 1182, + 269, + 1182, + 306, + 1075, + 306 + ], + "score": 0.9, + "latex": "\\{ Q _ { k } \\} _ { k = 1 } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 800, + 1106, + 863, + 1106, + 863, + 1134, + 800, + 1134 + ], + "score": 0.9, + "latex": "c _ { 1 } , c _ { 2 }" + }, + { + "category_id": 13, + "poly": [ + 478, + 394, + 588, + 394, + 588, + 424, + 478, + 424 + ], + "score": 0.9, + "latex": "t = 1 , \\dots" + }, + { + "category_id": 13, + "poly": [ + 716, + 459, + 769, + 459, + 769, + 486, + 716, + 486 + ], + "score": 0.9, + "latex": "s _ { t + 1 }" + }, + { + "category_id": 13, + "poly": [ + 736, + 1684, + 1185, + 1684, + 1185, + 1722, + 736, + 1722 + ], + "score": 0.89, + "latex": "\\begin{array} { r } { \\grave { a _ { t } } \\in \\mathrm { a r g m a x } _ { a } \\left\\{ \\tilde { \\mu } ( s _ { t } , a ) + \\lambda \\cdot \\tilde { \\sigma } ( s _ { t } , a ) \\right\\} } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 484, + 1748, + 662, + 1748, + 662, + 1780, + 484, + 1780 + ], + "score": 0.89, + "latex": "\\left( { { s _ { t } } , { a _ { t } } , { r _ { t } } , { s _ { t + 1 } } } \\right)" + }, + { + "category_id": 13, + "poly": [ + 373, + 1071, + 413, + 1071, + 413, + 1104, + 373, + 1104 + ], + "score": 0.89, + "latex": "\\hat { r } _ { i , t }" + }, + { + "category_id": 13, + "poly": [ + 920, + 273, + 957, + 273, + 957, + 303, + 920, + 303 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1015, + 923, + 1049, + 923, + 1049, + 952, + 1015, + 952 + ], + "score": 0.88, + "latex": "A _ { t }" + }, + { + "category_id": 13, + "poly": [ + 486, + 1657, + 596, + 1657, + 596, + 1687, + 486, + 1687 + ], + "score": 0.88, + "latex": "t = 1 , \\dots" + }, + { + "category_id": 13, + "poly": [ + 1226, + 1148, + 1289, + 1148, + 1289, + 1179, + 1226, + 1179 + ], + "score": 0.88, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 1080, + 1777, + 1146, + 1777, + 1146, + 1811, + 1080, + 1811 + ], + "score": 0.87, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 1080, + 514, + 1146, + 514, + 1146, + 548, + 1080, + 548 + ], + "score": 0.86, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 580, + 1721, + 634, + 1721, + 634, + 1748, + 580, + 1748 + ], + "score": 0.84, + "latex": "s _ { t + 1 }" + }, + { + "category_id": 13, + "poly": [ + 705, + 1535, + 731, + 1535, + 731, + 1565, + 705, + 1565 + ], + "score": 0.84, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 903, + 459, + 928, + 459, + 928, + 484, + 903, + 484 + ], + "score": 0.84, + "latex": "r _ { t }" + }, + { + "category_id": 13, + "poly": [ + 804, + 369, + 832, + 369, + 832, + 393, + 804, + 393 + ], + "score": 0.83, + "latex": "s _ { 0 }" + }, + { + "category_id": 13, + "poly": [ + 767, + 1722, + 793, + 1722, + 793, + 1747, + 767, + 1747 + ], + "score": 0.82, + "latex": "r _ { t }" + }, + { + "category_id": 13, + "poly": [ + 523, + 458, + 550, + 458, + 550, + 484, + 523, + 484 + ], + "score": 0.82, + "latex": "a _ { t }" + }, + { + "category_id": 13, + "poly": [ + 791, + 1535, + 820, + 1535, + 820, + 1561, + 791, + 1561 + ], + "score": 0.81, + "latex": "K" + }, + { + "category_id": 13, + "poly": [ + 804, + 1633, + 832, + 1633, + 832, + 1656, + 804, + 1656 + ], + "score": 0.81, + "latex": "s _ { 0 }" + }, + { + "category_id": 13, + "poly": [ + 907, + 671, + 939, + 671, + 939, + 708, + 907, + 708 + ], + "score": 0.8, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 845, + 486, + 872, + 486, + 872, + 512, + 845, + 512 + ], + "score": 0.78, + "latex": "B" + }, + { + "category_id": 13, + "poly": [ + 1284, + 1369, + 1304, + 1369, + 1304, + 1395, + 1284, + 1395 + ], + "score": 0.78, + "latex": "\\lambda" + }, + { + "category_id": 13, + "poly": [ + 388, + 303, + 414, + 303, + 414, + 329, + 388, + 329 + ], + "score": 0.78, + "latex": "B" + }, + { + "category_id": 13, + "poly": [ + 1233, + 1724, + 1260, + 1724, + 1260, + 1747, + 1233, + 1747 + ], + "score": 0.77, + "latex": "a _ { t }" + }, + { + "category_id": 13, + "poly": [ + 845, + 1748, + 871, + 1748, + 871, + 1774, + 845, + 1774 + ], + "score": 0.77, + "latex": "B" + }, + { + "category_id": 13, + "poly": [ + 388, + 1566, + 414, + 1566, + 414, + 1593, + 388, + 1593 + ], + "score": 0.75, + "latex": "B" + }, + { + "category_id": 13, + "poly": [ + 729, + 1105, + 742, + 1105, + 742, + 1129, + 729, + 1129 + ], + "score": 0.75, + "latex": "t" + }, + { + "category_id": 13, + "poly": [ + 734, + 421, + 1266, + 421, + 1266, + 458, + 734, + 458 + ], + "score": 0.74, + "latex": "\\hat { a } _ { t } = \\mathrm { M a j o r i t y V o t e } ( \\{ \\operatorname { a r g m a x } _ { a } Q _ { k } ( s _ { t } , a ) \\} _ { k = 1 } ^ { K } )" + }, + { + "category_id": 13, + "poly": [ + 908, + 925, + 921, + 925, + 921, + 950, + 908, + 950 + ], + "score": 0.73, + "latex": "t" + }, + { + "category_id": 13, + "poly": [ + 415, + 1105, + 427, + 1105, + 427, + 1128, + 415, + 1128 + ], + "score": 0.73, + "latex": "i" + }, + { + "category_id": 13, + "poly": [ + 1033, + 1073, + 1046, + 1073, + 1046, + 1098, + 1033, + 1098 + ], + "score": 0.71, + "latex": "i" + }, + { + "category_id": 13, + "poly": [ + 705, + 1495, + 729, + 1495, + 729, + 1524, + 705, + 1524 + ], + "score": 0.7, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1162, + 1075, + 1206, + 1075, + 1206, + 1105, + 1162, + 1105 + ], + "score": 0.66, + "latex": "n _ { i , t }" + }, + { + "category_id": 13, + "poly": [ + 1218, + 1535, + 1238, + 1535, + 1238, + 1561, + 1218, + 1561 + ], + "score": 0.6, + "latex": "\\lambda" + }, + { + "category_id": 13, + "poly": [ + 1137, + 1075, + 1151, + 1075, + 1151, + 1099, + 1137, + 1099 + ], + "score": 0.45, + "latex": "t" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 663.0, + 906.0, + 663.0, + 906.0, + 716.0, + 291.0, + 716.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 940.0, + 663.0, + 1129.0, + 663.0, + 1129.0, + 716.0, + 940.0, + 716.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 298.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1930.0, + 545.0, + 1930.0, + 545.0, + 1975.0, + 291.0, + 1975.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2086.0, + 862.0, + 2086.0, + 862.0, + 2118.0, + 838.0, + 2118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 221.0, + 646.0, + 221.0, + 646.0, + 272.0, + 293.0, + 272.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1484.0, + 704.0, + 1484.0, + 704.0, + 1534.0, + 294.0, + 1534.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 730.0, + 1484.0, + 865.0, + 1484.0, + 865.0, + 1534.0, + 730.0, + 1534.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 740.0, + 1406.0, + 740.0, + 1406.0, + 775.0, + 294.0, + 775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 767.0, + 1407.0, + 767.0, + 1407.0, + 807.0, + 291.0, + 807.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 798.0, + 1407.0, + 798.0, + 1407.0, + 837.0, + 291.0, + 837.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 829.0, + 1406.0, + 829.0, + 1406.0, + 866.0, + 294.0, + 866.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 860.0, + 1406.0, + 860.0, + 1406.0, + 895.0, + 294.0, + 895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 890.0, + 1406.0, + 890.0, + 1406.0, + 926.0, + 293.0, + 926.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 920.0, + 907.0, + 920.0, + 907.0, + 957.0, + 294.0, + 957.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 922.0, + 920.0, + 1014.0, + 920.0, + 1014.0, + 957.0, + 922.0, + 957.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1050.0, + 920.0, + 1298.0, + 920.0, + 1298.0, + 957.0, + 1050.0, + 957.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1143.0, + 1225.0, + 1143.0, + 1225.0, + 1185.0, + 292.0, + 1185.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1290.0, + 1143.0, + 1409.0, + 1143.0, + 1409.0, + 1185.0, + 1290.0, + 1185.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 287.0, + 1167.0, + 1023.0, + 1167.0, + 1023.0, + 1225.0, + 287.0, + 1225.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1115.0, + 1167.0, + 1151.0, + 1167.0, + 1151.0, + 1225.0, + 1115.0, + 1225.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1329.0, + 1167.0, + 1412.0, + 1167.0, + 1412.0, + 1225.0, + 1329.0, + 1225.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1204.0, + 475.0, + 1204.0, + 475.0, + 1255.0, + 292.0, + 1255.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 567.0, + 1204.0, + 601.0, + 1204.0, + 601.0, + 1255.0, + 567.0, + 1255.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 780.0, + 1204.0, + 1409.0, + 1204.0, + 1409.0, + 1255.0, + 780.0, + 1255.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1366.0, + 1283.0, + 1366.0, + 1283.0, + 1403.0, + 295.0, + 1403.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1305.0, + 1366.0, + 1405.0, + 1366.0, + 1405.0, + 1403.0, + 1305.0, + 1403.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1398.0, + 1403.0, + 1398.0, + 1403.0, + 1432.0, + 296.0, + 1432.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1430.0, + 780.0, + 1430.0, + 780.0, + 1463.0, + 294.0, + 1463.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1067.0, + 372.0, + 1067.0, + 372.0, + 1105.0, + 293.0, + 1105.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 414.0, + 1067.0, + 465.0, + 1067.0, + 465.0, + 1105.0, + 414.0, + 1105.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 510.0, + 1067.0, + 1032.0, + 1067.0, + 1032.0, + 1105.0, + 510.0, + 1105.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1047.0, + 1067.0, + 1136.0, + 1067.0, + 1136.0, + 1105.0, + 1047.0, + 1105.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1152.0, + 1067.0, + 1161.0, + 1067.0, + 1161.0, + 1105.0, + 1152.0, + 1105.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1207.0, + 1067.0, + 1406.0, + 1067.0, + 1406.0, + 1105.0, + 1207.0, + 1105.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1103.0, + 414.0, + 1103.0, + 414.0, + 1135.0, + 297.0, + 1135.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 428.0, + 1103.0, + 728.0, + 1103.0, + 728.0, + 1135.0, + 428.0, + 1135.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 743.0, + 1103.0, + 799.0, + 1103.0, + 799.0, + 1135.0, + 743.0, + 1135.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 864.0, + 1103.0, + 1117.0, + 1103.0, + 1117.0, + 1135.0, + 864.0, + 1135.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1315.0, + 370.0, + 1315.0, + 370.0, + 1360.0, + 295.0, + 1360.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 463.0, + 1315.0, + 697.0, + 1315.0, + 697.0, + 1360.0, + 463.0, + 1360.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1998.0, + 1130.0, + 1998.0, + 1130.0, + 2040.0, + 293.0, + 2040.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 304.0, + 263.0, + 426.0, + 263.0, + 426.0, + 314.0, + 304.0, + 314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 529.0, + 263.0, + 919.0, + 263.0, + 919.0, + 314.0, + 529.0, + 314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 958.0, + 263.0, + 1074.0, + 263.0, + 1074.0, + 314.0, + 958.0, + 314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1183.0, + 263.0, + 1194.0, + 263.0, + 1194.0, + 314.0, + 1183.0, + 314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 298.0, + 387.0, + 298.0, + 387.0, + 338.0, + 306.0, + 338.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 415.0, + 298.0, + 959.0, + 298.0, + 959.0, + 338.0, + 415.0, + 338.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 309.0, + 332.0, + 607.0, + 332.0, + 607.0, + 364.0, + 309.0, + 364.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 362.0, + 340.0, + 362.0, + 340.0, + 395.0, + 307.0, + 395.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 356.0, + 803.0, + 356.0, + 803.0, + 399.0, + 382.0, + 399.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 833.0, + 356.0, + 837.0, + 356.0, + 837.0, + 399.0, + 833.0, + 399.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 392.0, + 340.0, + 392.0, + 340.0, + 426.0, + 307.0, + 426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 390.0, + 477.0, + 390.0, + 477.0, + 426.0, + 382.0, + 426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 589.0, + 390.0, + 857.0, + 390.0, + 857.0, + 426.0, + 589.0, + 426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 424.0, + 340.0, + 424.0, + 340.0, + 456.0, + 307.0, + 456.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 421.0, + 417.0, + 733.0, + 417.0, + 733.0, + 464.0, + 421.0, + 464.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1267.0, + 417.0, + 1273.0, + 417.0, + 1273.0, + 464.0, + 1267.0, + 464.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 453.0, + 340.0, + 453.0, + 340.0, + 486.0, + 307.0, + 486.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 424.0, + 450.0, + 522.0, + 450.0, + 522.0, + 489.0, + 424.0, + 489.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 551.0, + 450.0, + 715.0, + 450.0, + 715.0, + 489.0, + 551.0, + 489.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 770.0, + 450.0, + 902.0, + 450.0, + 902.0, + 489.0, + 770.0, + 489.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 929.0, + 450.0, + 1180.0, + 450.0, + 1180.0, + 489.0, + 929.0, + 489.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 310.0, + 481.0, + 341.0, + 481.0, + 341.0, + 519.0, + 310.0, + 519.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 425.0, + 483.0, + 482.0, + 483.0, + 482.0, + 520.0, + 425.0, + 520.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 664.0, + 483.0, + 844.0, + 483.0, + 844.0, + 520.0, + 664.0, + 520.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 514.0, + 340.0, + 514.0, + 340.0, + 547.0, + 306.0, + 547.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 424.0, + 512.0, + 1079.0, + 512.0, + 1079.0, + 551.0, + 424.0, + 551.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 542.0, + 341.0, + 542.0, + 341.0, + 580.0, + 297.0, + 580.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 380.0, + 543.0, + 478.0, + 543.0, + 478.0, + 575.0, + 380.0, + 575.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 572.0, + 439.0, + 572.0, + 439.0, + 609.0, + 296.0, + 609.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 303.0, + 1524.0, + 704.0, + 1524.0, + 704.0, + 1577.0, + 303.0, + 1577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 732.0, + 1524.0, + 790.0, + 1524.0, + 790.0, + 1577.0, + 732.0, + 1577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 821.0, + 1524.0, + 911.0, + 1524.0, + 911.0, + 1577.0, + 821.0, + 1577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1019.0, + 1524.0, + 1217.0, + 1524.0, + 1217.0, + 1577.0, + 1019.0, + 1577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1239.0, + 1524.0, + 1252.0, + 1524.0, + 1252.0, + 1577.0, + 1239.0, + 1577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1561.0, + 387.0, + 1561.0, + 387.0, + 1602.0, + 306.0, + 1602.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 415.0, + 1561.0, + 969.0, + 1561.0, + 969.0, + 1602.0, + 415.0, + 1602.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 308.0, + 1595.0, + 572.0, + 1595.0, + 572.0, + 1626.0, + 308.0, + 1626.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 308.0, + 1625.0, + 339.0, + 1625.0, + 339.0, + 1658.0, + 308.0, + 1658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 383.0, + 1622.0, + 803.0, + 1622.0, + 803.0, + 1661.0, + 383.0, + 1661.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 308.0, + 1656.0, + 339.0, + 1656.0, + 339.0, + 1687.0, + 308.0, + 1687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 1654.0, + 485.0, + 1654.0, + 485.0, + 1687.0, + 382.0, + 1687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 597.0, + 1654.0, + 864.0, + 1654.0, + 864.0, + 1687.0, + 597.0, + 1687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 308.0, + 1687.0, + 339.0, + 1687.0, + 339.0, + 1717.0, + 308.0, + 1717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 421.0, + 1681.0, + 735.0, + 1681.0, + 735.0, + 1724.0, + 421.0, + 1724.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 309.0, + 1714.0, + 341.0, + 1714.0, + 341.0, + 1749.0, + 309.0, + 1749.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 423.0, + 1712.0, + 579.0, + 1712.0, + 579.0, + 1754.0, + 423.0, + 1754.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 635.0, + 1712.0, + 766.0, + 1712.0, + 766.0, + 1754.0, + 635.0, + 1754.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 794.0, + 1712.0, + 1232.0, + 1712.0, + 1232.0, + 1754.0, + 794.0, + 1754.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1261.0, + 1712.0, + 1265.0, + 1712.0, + 1265.0, + 1754.0, + 1261.0, + 1754.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 308.0, + 1747.0, + 339.0, + 1747.0, + 339.0, + 1778.0, + 308.0, + 1778.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 425.0, + 1746.0, + 483.0, + 1746.0, + 483.0, + 1781.0, + 425.0, + 1781.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 663.0, + 1746.0, + 844.0, + 1746.0, + 844.0, + 1781.0, + 663.0, + 1781.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 1778.0, + 339.0, + 1778.0, + 339.0, + 1809.0, + 307.0, + 1809.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 423.0, + 1773.0, + 1079.0, + 1773.0, + 1079.0, + 1814.0, + 423.0, + 1814.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 1805.0, + 340.0, + 1805.0, + 340.0, + 1841.0, + 298.0, + 1841.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 1807.0, + 479.0, + 1807.0, + 479.0, + 1839.0, + 382.0, + 1839.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1836.0, + 439.0, + 1836.0, + 439.0, + 1870.0, + 294.0, + 1870.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 3, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1178, + 1404, + 1178, + 1404, + 1545, + 298, + 1545 + ], + "score": 0.98 + }, + { + "category_id": 1, + "poly": [ + 298, + 1728, + 1404, + 1728, + 1404, + 2034, + 298, + 2034 + ], + "score": 0.979 + }, + { + "category_id": 1, + "poly": [ + 298, + 480, + 1405, + 480, + 1405, + 696, + 298, + 696 + ], + "score": 0.976 + }, + { + "category_id": 1, + "poly": [ + 298, + 1560, + 1405, + 1560, + 1405, + 1713, + 298, + 1713 + ], + "score": 0.968 + }, + { + "category_id": 1, + "poly": [ + 362, + 228, + 1404, + 228, + 1404, + 450, + 362, + 450 + ], + "score": 0.936 + }, + { + "category_id": 1, + "poly": [ + 361, + 769, + 1405, + 769, + 1405, + 1041, + 361, + 1041 + ], + "score": 0.934 + }, + { + "category_id": 1, + "poly": [ + 298, + 1070, + 1405, + 1070, + 1405, + 1163, + 298, + 1163 + ], + "score": 0.932 + }, + { + "category_id": 2, + "poly": [ + 298, + 75, + 857, + 75, + 857, + 105, + 298, + 105 + ], + "score": 0.903 + }, + { + "category_id": 1, + "poly": [ + 297, + 709, + 1180, + 709, + 1180, + 743, + 297, + 743 + ], + "score": 0.85 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 858, + 2088, + 858, + 2112, + 841, + 2112 + ], + "score": 0.73 + }, + { + "category_id": 13, + "poly": [ + 1200, + 602, + 1265, + 602, + 1265, + 636, + 1200, + 636 + ], + "score": 0.93, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 298, + 633, + 393, + 633, + 393, + 661, + 298, + 661 + ], + "score": 0.89, + "latex": "K = 1 0" + }, + { + "category_id": 13, + "poly": [ + 1209, + 821, + 1300, + 821, + 1300, + 850, + 1209, + 850 + ], + "score": 0.89, + "latex": "\\lambda = 0 . 1" + }, + { + "category_id": 13, + "poly": [ + 960, + 1591, + 1034, + 1591, + 1034, + 1620, + 960, + 1620 + ], + "score": 0.87, + "latex": "{ \\bf A } 3 { \\bf C } +" + }, + { + "category_id": 13, + "poly": [ + 1297, + 1561, + 1372, + 1561, + 1372, + 1590, + 1297, + 1590 + ], + "score": 0.85, + "latex": "{ \\bf A } 3 { \\bf C } +" + }, + { + "category_id": 13, + "poly": [ + 688, + 1683, + 761, + 1683, + 761, + 1711, + 688, + 1711 + ], + "score": 0.84, + "latex": "{ \\bf A } 3 { \\bf C } +" + }, + { + "category_id": 13, + "poly": [ + 351, + 1652, + 423, + 1652, + 423, + 1681, + 351, + 1681 + ], + "score": 0.81, + "latex": "{ \\bf A } 3 { \\bf C } +" + }, + { + "category_id": 13, + "poly": [ + 1213, + 1683, + 1283, + 1683, + 1283, + 1711, + 1213, + 1711 + ], + "score": 0.78, + "latex": "{ \\bf A } 3 { \\bf C } +" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 298.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2085.0, + 861.0, + 2085.0, + 861.0, + 2120.0, + 839.0, + 2120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1179.0, + 1404.0, + 1179.0, + 1404.0, + 1213.0, + 294.0, + 1213.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1209.0, + 1405.0, + 1209.0, + 1405.0, + 1242.0, + 293.0, + 1242.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1236.0, + 1409.0, + 1236.0, + 1409.0, + 1276.0, + 291.0, + 1276.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1270.0, + 1405.0, + 1270.0, + 1405.0, + 1304.0, + 294.0, + 1304.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1298.0, + 1406.0, + 1298.0, + 1406.0, + 1336.0, + 293.0, + 1336.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1330.0, + 1406.0, + 1330.0, + 1406.0, + 1368.0, + 293.0, + 1368.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1358.0, + 1405.0, + 1358.0, + 1405.0, + 1398.0, + 292.0, + 1398.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1392.0, + 1404.0, + 1392.0, + 1404.0, + 1426.0, + 294.0, + 1426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1419.0, + 1406.0, + 1419.0, + 1406.0, + 1460.0, + 292.0, + 1460.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1451.0, + 1408.0, + 1451.0, + 1408.0, + 1489.0, + 293.0, + 1489.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1482.0, + 1406.0, + 1482.0, + 1406.0, + 1520.0, + 293.0, + 1520.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1512.0, + 1120.0, + 1512.0, + 1120.0, + 1550.0, + 294.0, + 1550.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1726.0, + 1405.0, + 1726.0, + 1405.0, + 1764.0, + 293.0, + 1764.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1759.0, + 1405.0, + 1759.0, + 1405.0, + 1795.0, + 293.0, + 1795.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1788.0, + 1405.0, + 1788.0, + 1405.0, + 1823.0, + 292.0, + 1823.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1819.0, + 1406.0, + 1819.0, + 1406.0, + 1856.0, + 292.0, + 1856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1851.0, + 1406.0, + 1851.0, + 1406.0, + 1886.0, + 294.0, + 1886.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1881.0, + 1406.0, + 1881.0, + 1406.0, + 1917.0, + 294.0, + 1917.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1913.0, + 1405.0, + 1913.0, + 1405.0, + 1944.0, + 296.0, + 1944.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1942.0, + 1409.0, + 1942.0, + 1409.0, + 1976.0, + 292.0, + 1976.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1972.0, + 1406.0, + 1972.0, + 1406.0, + 2008.0, + 294.0, + 2008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 2003.0, + 951.0, + 2003.0, + 951.0, + 2038.0, + 292.0, + 2038.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 482.0, + 1407.0, + 482.0, + 1407.0, + 513.0, + 296.0, + 513.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 507.0, + 1407.0, + 507.0, + 1407.0, + 547.0, + 293.0, + 547.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 541.0, + 1405.0, + 541.0, + 1405.0, + 574.0, + 293.0, + 574.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 572.0, + 1405.0, + 572.0, + 1405.0, + 607.0, + 295.0, + 607.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 603.0, + 1199.0, + 603.0, + 1199.0, + 637.0, + 295.0, + 637.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1266.0, + 603.0, + 1405.0, + 603.0, + 1405.0, + 637.0, + 1266.0, + 637.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 630.0, + 297.0, + 630.0, + 297.0, + 669.0, + 293.0, + 669.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 630.0, + 1407.0, + 630.0, + 1407.0, + 669.0, + 394.0, + 669.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 660.0, + 936.0, + 660.0, + 936.0, + 699.0, + 293.0, + 699.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1559.0, + 1296.0, + 1559.0, + 1296.0, + 1596.0, + 293.0, + 1596.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1373.0, + 1559.0, + 1406.0, + 1559.0, + 1406.0, + 1596.0, + 1373.0, + 1596.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1588.0, + 959.0, + 1588.0, + 959.0, + 1626.0, + 293.0, + 1626.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1035.0, + 1588.0, + 1408.0, + 1588.0, + 1408.0, + 1626.0, + 1035.0, + 1626.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1619.0, + 1405.0, + 1619.0, + 1405.0, + 1656.0, + 295.0, + 1656.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1648.0, + 350.0, + 1648.0, + 350.0, + 1688.0, + 293.0, + 1688.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 424.0, + 1648.0, + 1410.0, + 1648.0, + 1410.0, + 1688.0, + 424.0, + 1688.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1680.0, + 687.0, + 1680.0, + 687.0, + 1718.0, + 293.0, + 1718.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 762.0, + 1680.0, + 1212.0, + 1680.0, + 1212.0, + 1718.0, + 762.0, + 1718.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1284.0, + 1680.0, + 1294.0, + 1680.0, + 1294.0, + 1718.0, + 1284.0, + 1718.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 360.0, + 227.0, + 1405.0, + 227.0, + 1405.0, + 269.0, + 360.0, + 269.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 260.0, + 741.0, + 260.0, + 741.0, + 299.0, + 392.0, + 299.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 357.0, + 305.0, + 1404.0, + 305.0, + 1404.0, + 349.0, + 357.0, + 349.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 340.0, + 1122.0, + 340.0, + 1122.0, + 375.0, + 394.0, + 375.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 357.0, + 386.0, + 1405.0, + 386.0, + 1405.0, + 426.0, + 357.0, + 426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 420.0, + 923.0, + 420.0, + 923.0, + 453.0, + 394.0, + 453.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 361.0, + 771.0, + 1192.0, + 771.0, + 1192.0, + 807.0, + 361.0, + 807.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 357.0, + 819.0, + 1208.0, + 819.0, + 1208.0, + 859.0, + 357.0, + 859.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1301.0, + 819.0, + 1406.0, + 819.0, + 1406.0, + 859.0, + 1301.0, + 859.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 853.0, + 1355.0, + 853.0, + 1355.0, + 885.0, + 393.0, + 885.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 357.0, + 898.0, + 1407.0, + 898.0, + 1407.0, + 938.0, + 357.0, + 938.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 931.0, + 666.0, + 931.0, + 666.0, + 964.0, + 394.0, + 964.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 357.0, + 978.0, + 1404.0, + 978.0, + 1404.0, + 1014.0, + 357.0, + 1014.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 1011.0, + 672.0, + 1011.0, + 672.0, + 1047.0, + 392.0, + 1047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1069.0, + 1408.0, + 1069.0, + 1408.0, + 1107.0, + 295.0, + 1107.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1101.0, + 1405.0, + 1101.0, + 1405.0, + 1135.0, + 296.0, + 1135.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1129.0, + 1288.0, + 1129.0, + 1288.0, + 1171.0, + 293.0, + 1171.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 707.0, + 1182.0, + 707.0, + 1182.0, + 747.0, + 295.0, + 747.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 4, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 3, + "poly": [ + 321, + 296, + 1446, + 296, + 1446, + 940, + 321, + 940 + ], + "score": 0.972 + }, + { + "category_id": 3, + "poly": [ + 296, + 1216, + 1469, + 1216, + 1469, + 1866, + 296, + 1866 + ], + "score": 0.971 + }, + { + "category_id": 4, + "poly": [ + 296, + 984, + 1408, + 984, + 1408, + 1080, + 296, + 1080 + ], + "score": 0.947 + }, + { + "category_id": 4, + "poly": [ + 298, + 1901, + 1405, + 1901, + 1405, + 1965, + 298, + 1965 + ], + "score": 0.946 + }, + { + "category_id": 2, + "poly": [ + 297, + 75, + 857, + 75, + 857, + 105, + 297, + 105 + ], + "score": 0.905 + }, + { + "category_id": 2, + "poly": [ + 840, + 2088, + 859, + 2088, + 859, + 2113, + 840, + 2113 + ], + "score": 0.79 + }, + { + "category_id": 15, + "poly": [ + 547.0, + 296.0, + 1011.0, + 296.0, + 1011.0, + 339.0, + 547.0, + 339.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 330.0, + 317.0, + 382.0, + 317.0, + 382.0, + 357.0, + 330.0, + 357.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 328.0, + 380.0, + 383.0, + 380.0, + 383.0, + 424.0, + 328.0, + 424.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 445.0, + 383.0, + 445.0, + 383.0, + 489.0, + 327.0, + 489.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 511.0, + 383.0, + 511.0, + 383.0, + 555.0, + 327.0, + 555.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1245.0, + 524.0, + 1446.0, + 524.0, + 1446.0, + 572.0, + 1245.0, + 572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 575.0, + 383.0, + 575.0, + 383.0, + 619.0, + 327.0, + 619.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1244.0, + 560.0, + 1436.0, + 560.0, + 1436.0, + 636.0, + 1244.0, + 636.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 640.0, + 385.0, + 640.0, + 385.0, + 684.0, + 327.0, + 684.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1208.0, + 637.0, + 1241.0, + 637.0, + 1241.0, + 655.0, + 1208.0, + 655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1245.0, + 627.0, + 1382.0, + 627.0, + 1382.0, + 670.0, + 1245.0, + 670.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 706.0, + 385.0, + 706.0, + 385.0, + 750.0, + 327.0, + 750.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 456.0, + 729.0, + 479.0, + 729.0, + 479.0, + 750.0, + 456.0, + 750.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 452.0, + 745.0, + 462.0, + 745.0, + 462.0, + 760.0, + 452.0, + 760.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 432.0, + 763.0, + 452.0, + 763.0, + 452.0, + 786.0, + 432.0, + 786.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 770.0, + 382.0, + 770.0, + 382.0, + 814.0, + 327.0, + 814.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 396.0, + 787.0, + 408.0, + 787.0, + 408.0, + 799.0, + 396.0, + 799.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 835.0, + 385.0, + 835.0, + 385.0, + 878.0, + 327.0, + 878.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 359.0, + 863.0, + 416.0, + 863.0, + 416.0, + 906.0, + 359.0, + 906.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 459.0, + 863.0, + 515.0, + 863.0, + 515.0, + 906.0, + 459.0, + 906.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 557.0, + 865.0, + 610.0, + 865.0, + 610.0, + 903.0, + 557.0, + 903.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 654.0, + 865.0, + 709.0, + 865.0, + 709.0, + 903.0, + 654.0, + 903.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 750.0, + 862.0, + 807.0, + 862.0, + 807.0, + 905.0, + 750.0, + 905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 846.0, + 862.0, + 905.0, + 862.0, + 905.0, + 905.0, + 846.0, + 905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 947.0, + 865.0, + 1001.0, + 865.0, + 1001.0, + 903.0, + 947.0, + 903.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1043.0, + 862.0, + 1099.0, + 862.0, + 1099.0, + 905.0, + 1043.0, + 905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1140.0, + 862.0, + 1197.0, + 862.0, + 1197.0, + 905.0, + 1140.0, + 905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 725.0, + 895.0, + 832.0, + 895.0, + 832.0, + 940.0, + 725.0, + 940.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1121.0, + 892.0, + 1181.0, + 892.0, + 1181.0, + 934.0, + 1121.0, + 934.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 395.0, + 790.0, + 439.0, + 790.0, + 439.0, + 807.5, + 395.0, + 807.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 444.0, + 1230.0, + 552.0, + 1230.0, + 552.0, + 1255.0, + 444.0, + 1255.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 800.0, + 1228.0, + 867.0, + 1228.0, + 867.0, + 1256.0, + 800.0, + 1256.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1128.0, + 1228.0, + 1212.0, + 1228.0, + 1212.0, + 1259.0, + 1128.0, + 1259.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1243.0, + 356.0, + 1243.0, + 356.0, + 1265.0, + 306.0, + 1265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 648.0, + 1240.0, + 693.0, + 1240.0, + 693.0, + 1268.0, + 648.0, + 1268.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 1243.0, + 1028.0, + 1243.0, + 1028.0, + 1265.0, + 978.0, + 1265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 1273.0, + 1028.0, + 1273.0, + 1028.0, + 1295.0, + 978.0, + 1295.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1283.0, + 356.0, + 1283.0, + 356.0, + 1305.0, + 306.0, + 1305.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 648.0, + 1283.0, + 692.0, + 1283.0, + 692.0, + 1307.0, + 648.0, + 1307.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 304.0, + 1322.0, + 357.0, + 1322.0, + 357.0, + 1348.0, + 304.0, + 1348.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 596.0, + 1317.0, + 619.0, + 1317.0, + 619.0, + 1331.0, + 596.0, + 1331.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 648.0, + 1322.0, + 693.0, + 1322.0, + 693.0, + 1350.0, + 648.0, + 1350.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 977.0, + 1301.0, + 1030.0, + 1301.0, + 1030.0, + 1359.0, + 977.0, + 1359.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1365.0, + 356.0, + 1365.0, + 356.0, + 1387.0, + 306.0, + 1387.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 648.0, + 1363.0, + 693.0, + 1363.0, + 693.0, + 1390.0, + 648.0, + 1390.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 983.0, + 1363.0, + 1030.0, + 1363.0, + 1030.0, + 1390.0, + 983.0, + 1390.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 304.0, + 1404.0, + 357.0, + 1404.0, + 357.0, + 1430.0, + 304.0, + 1430.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 650.0, + 1406.0, + 692.0, + 1406.0, + 692.0, + 1429.0, + 650.0, + 1429.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 743.0, + 1406.0, + 750.0, + 1406.0, + 750.0, + 1412.0, + 743.0, + 1412.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1395.0, + 1028.0, + 1395.0, + 1028.0, + 1419.0, + 985.0, + 1419.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 583.0, + 1421.0, + 597.0, + 1421.0, + 597.0, + 1430.0, + 583.0, + 1430.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 771.0, + 1439.0, + 778.0, + 1439.0, + 778.0, + 1446.0, + 771.0, + 1446.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1426.0, + 1028.0, + 1426.0, + 1028.0, + 1449.0, + 985.0, + 1449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 304.0, + 1445.0, + 357.0, + 1445.0, + 357.0, + 1471.0, + 304.0, + 1471.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 656.0, + 1446.0, + 692.0, + 1446.0, + 692.0, + 1469.0, + 656.0, + 1469.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 762.0, + 1451.0, + 768.0, + 1451.0, + 768.0, + 1458.0, + 762.0, + 1458.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 491.0, + 1463.0, + 522.0, + 1463.0, + 522.0, + 1477.0, + 491.0, + 1477.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 597.0, + 1459.0, + 631.0, + 1459.0, + 631.0, + 1472.0, + 597.0, + 1472.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 742.0, + 1458.0, + 764.0, + 1458.0, + 764.0, + 1477.0, + 742.0, + 1477.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1456.0, + 1030.0, + 1456.0, + 1030.0, + 1480.0, + 985.0, + 1480.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 737.0, + 1473.0, + 753.0, + 1473.0, + 753.0, + 1489.0, + 737.0, + 1489.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 339.0, + 1490.0, + 356.0, + 1490.0, + 356.0, + 1508.0, + 339.0, + 1508.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 674.0, + 1490.0, + 690.0, + 1490.0, + 690.0, + 1508.0, + 674.0, + 1508.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1010.0, + 1490.0, + 1028.0, + 1490.0, + 1028.0, + 1507.0, + 1010.0, + 1507.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 340.0, + 1501.0, + 652.0, + 1501.0, + 652.0, + 1527.0, + 340.0, + 1527.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 1502.0, + 989.0, + 1502.0, + 989.0, + 1527.0, + 678.0, + 1527.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1014.0, + 1502.0, + 1254.0, + 1502.0, + 1254.0, + 1527.0, + 1014.0, + 1527.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1255.0, + 1503.0, + 1321.0, + 1503.0, + 1321.0, + 1525.0, + 1255.0, + 1525.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 466.0, + 1524.0, + 530.0, + 1524.0, + 530.0, + 1546.0, + 466.0, + 1546.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 608.0, + 1520.0, + 642.0, + 1520.0, + 642.0, + 1544.0, + 608.0, + 1544.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 802.0, + 1524.0, + 865.0, + 1524.0, + 865.0, + 1546.0, + 802.0, + 1546.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 943.0, + 1520.0, + 978.0, + 1520.0, + 978.0, + 1544.0, + 943.0, + 1544.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1137.0, + 1521.0, + 1202.0, + 1521.0, + 1202.0, + 1548.0, + 1137.0, + 1548.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1279.0, + 1520.0, + 1313.0, + 1520.0, + 1313.0, + 1544.0, + 1279.0, + 1544.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 458.0, + 1544.0, + 536.0, + 1544.0, + 536.0, + 1570.0, + 458.0, + 1570.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 790.0, + 1541.0, + 876.0, + 1541.0, + 876.0, + 1574.0, + 790.0, + 1574.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1126.0, + 1545.0, + 1214.0, + 1545.0, + 1214.0, + 1570.0, + 1126.0, + 1570.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1555.0, + 357.0, + 1555.0, + 357.0, + 1581.0, + 306.0, + 1581.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 640.0, + 1555.0, + 693.0, + 1555.0, + 693.0, + 1581.0, + 640.0, + 1581.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 975.0, + 1555.0, + 1030.0, + 1555.0, + 1030.0, + 1581.0, + 975.0, + 1581.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1588.0, + 356.0, + 1588.0, + 356.0, + 1610.0, + 306.0, + 1610.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 641.0, + 1606.0, + 692.0, + 1606.0, + 692.0, + 1628.0, + 641.0, + 1628.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 304.0, + 1617.0, + 357.0, + 1617.0, + 357.0, + 1643.0, + 304.0, + 1643.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 1618.0, + 1027.0, + 1618.0, + 1027.0, + 1640.0, + 978.0, + 1640.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 919.0, + 1634.0, + 960.0, + 1634.0, + 960.0, + 1655.0, + 919.0, + 1655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 1649.0, + 356.0, + 1649.0, + 356.0, + 1702.0, + 307.0, + 1702.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 532.0, + 1663.0, + 563.0, + 1663.0, + 563.0, + 1682.0, + 532.0, + 1682.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 585.0, + 1648.0, + 612.0, + 1648.0, + 612.0, + 1667.0, + 585.0, + 1667.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 642.0, + 1656.0, + 692.0, + 1656.0, + 692.0, + 1678.0, + 642.0, + 1678.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 977.0, + 1678.0, + 1030.0, + 1678.0, + 1030.0, + 1704.0, + 977.0, + 1704.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1193.0, + 1697.0, + 1203.0, + 1697.0, + 1203.0, + 1706.0, + 1193.0, + 1706.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1351.0, + 1650.0, + 1473.0, + 1650.0, + 1473.0, + 1714.0, + 1351.0, + 1714.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 314.0, + 1710.0, + 356.0, + 1710.0, + 356.0, + 1734.0, + 314.0, + 1734.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 641.0, + 1702.0, + 693.0, + 1702.0, + 693.0, + 1728.0, + 641.0, + 1728.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 875.0, + 1714.0, + 976.0, + 1714.0, + 976.0, + 1756.0, + 875.0, + 1756.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1117.0, + 1722.0, + 1136.0, + 1722.0, + 1136.0, + 1736.0, + 1117.0, + 1736.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1280.0, + 1704.0, + 1306.0, + 1704.0, + 1306.0, + 1723.0, + 1280.0, + 1723.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1352.0, + 1710.0, + 1459.0, + 1710.0, + 1459.0, + 1731.0, + 1352.0, + 1731.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 312.0, + 1740.0, + 356.0, + 1740.0, + 356.0, + 1764.0, + 312.0, + 1764.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 855.0, + 1742.0, + 872.0, + 1742.0, + 872.0, + 1758.0, + 855.0, + 1758.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1740.0, + 1028.0, + 1740.0, + 1028.0, + 1764.0, + 985.0, + 1764.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 312.0, + 1771.0, + 355.0, + 1771.0, + 355.0, + 1794.0, + 312.0, + 1794.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 649.0, + 1752.0, + 692.0, + 1752.0, + 692.0, + 1775.0, + 649.0, + 1775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 811.0, + 1770.0, + 823.0, + 1770.0, + 823.0, + 1781.0, + 811.0, + 1781.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 845.0, + 1757.0, + 856.0, + 1757.0, + 856.0, + 1768.0, + 845.0, + 1768.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1077.0, + 1761.0, + 1096.0, + 1761.0, + 1096.0, + 1777.0, + 1077.0, + 1777.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 357.0, + 1790.0, + 371.0, + 1790.0, + 371.0, + 1811.0, + 357.0, + 1811.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 337.0, + 1804.0, + 353.0, + 1804.0, + 353.0, + 1822.0, + 337.0, + 1822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 674.0, + 1804.0, + 690.0, + 1804.0, + 690.0, + 1822.0, + 674.0, + 1822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1010.0, + 1804.0, + 1026.0, + 1804.0, + 1026.0, + 1822.0, + 1010.0, + 1822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 344.0, + 1817.0, + 652.0, + 1817.0, + 652.0, + 1842.0, + 344.0, + 1842.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 681.0, + 1818.0, + 782.0, + 1818.0, + 782.0, + 1840.0, + 681.0, + 1840.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 783.0, + 1818.0, + 987.0, + 1818.0, + 987.0, + 1840.0, + 783.0, + 1840.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1016.0, + 1820.0, + 1048.0, + 1820.0, + 1048.0, + 1839.0, + 1016.0, + 1839.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1050.0, + 1820.0, + 1081.0, + 1820.0, + 1081.0, + 1839.0, + 1050.0, + 1839.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1084.0, + 1817.0, + 1324.0, + 1817.0, + 1324.0, + 1842.0, + 1084.0, + 1842.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 465.0, + 1835.0, + 532.0, + 1835.0, + 532.0, + 1866.0, + 465.0, + 1866.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 608.0, + 1834.0, + 642.0, + 1834.0, + 642.0, + 1857.0, + 608.0, + 1857.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 800.0, + 1835.0, + 867.0, + 1835.0, + 867.0, + 1866.0, + 800.0, + 1866.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 945.0, + 1837.0, + 978.0, + 1837.0, + 978.0, + 1856.0, + 945.0, + 1856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1136.0, + 1835.0, + 1203.0, + 1835.0, + 1203.0, + 1866.0, + 1136.0, + 1866.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1280.0, + 1837.0, + 1312.0, + 1837.0, + 1312.0, + 1856.0, + 1280.0, + 1856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1279.0, + 1284.5, + 1299.0, + 1284.5, + 1299.0, + 1299.5, + 1279.0, + 1299.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 934.0, + 1435.5, + 958.0, + 1435.5, + 958.0, + 1455.0, + 934.0, + 1455.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 426.0, + 1455.0, + 499.0, + 1455.0, + 499.0, + 1480.0, + 426.0, + 1480.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1101.0, + 1477.5, + 1155.0, + 1477.5, + 1155.0, + 1501.5, + 1101.0, + 1501.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 913.0, + 1654.5, + 945.0, + 1654.5, + 945.0, + 1668.5, + 913.0, + 1668.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 892.0, + 1676.0, + 922.0, + 1676.0, + 922.0, + 1688.0, + 892.0, + 1688.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 370.0, + 1676.0, + 567.0, + 1676.0, + 567.0, + 1736.5, + 370.0, + 1736.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1119.25, + 1742.0, + 1135.25, + 1742.0, + 1135.25, + 1752.5, + 1119.25, + 1752.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1086.0, + 1749.0, + 1131.0, + 1749.0, + 1131.0, + 1767.0, + 1086.0, + 1767.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 356.0, + 1768.0, + 400.0, + 1768.0, + 400.0, + 1782.0, + 356.0, + 1782.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1016.0, + 1782.5, + 1091.0, + 1782.5, + 1091.0, + 1809.5, + 1016.0, + 1809.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 985.0, + 1405.0, + 985.0, + 1405.0, + 1020.0, + 295.0, + 1020.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1016.0, + 1405.0, + 1016.0, + 1405.0, + 1050.0, + 294.0, + 1050.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1048.0, + 1338.0, + 1048.0, + 1338.0, + 1082.0, + 295.0, + 1082.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1898.0, + 1408.0, + 1898.0, + 1408.0, + 1938.0, + 293.0, + 1938.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1932.0, + 475.0, + 1932.0, + 475.0, + 1969.0, + 292.0, + 1969.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 297.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 840.0, + 2087.0, + 862.0, + 2087.0, + 862.0, + 2117.0, + 840.0, + 2117.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 5, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 294, + 1405, + 294, + 1405, + 450, + 298, + 450 + ], + "score": 0.965 + }, + { + "category_id": 2, + "poly": [ + 298, + 75, + 857, + 75, + 857, + 105, + 298, + 105 + ], + "score": 0.904 + }, + { + "category_id": 0, + "poly": [ + 299, + 226, + 544, + 226, + 544, + 262, + 299, + 262 + ], + "score": 0.881 + }, + { + "category_id": 0, + "poly": [ + 299, + 496, + 488, + 496, + 488, + 528, + 299, + 528 + ], + "score": 0.863 + }, + { + "category_id": 2, + "poly": [ + 841, + 2087, + 858, + 2087, + 858, + 2111, + 841, + 2111 + ], + "score": 0.734 + }, + { + "category_id": 1, + "poly": [ + 295, + 545, + 1403, + 545, + 1403, + 609, + 295, + 609 + ], + "score": 0.722 + }, + { + "category_id": 1, + "poly": [ + 297, + 1529, + 1399, + 1529, + 1399, + 1594, + 297, + 1594 + ], + "score": 0.647 + }, + { + "category_id": 1, + "poly": [ + 298, + 1447, + 1399, + 1447, + 1399, + 1511, + 298, + 1511 + ], + "score": 0.644 + }, + { + "category_id": 1, + "poly": [ + 298, + 1366, + 1401, + 1366, + 1401, + 1429, + 298, + 1429 + ], + "score": 0.642 + }, + { + "category_id": 1, + "poly": [ + 296, + 1725, + 1010, + 1725, + 1010, + 1759, + 296, + 1759 + ], + "score": 0.64 + }, + { + "category_id": 1, + "poly": [ + 299, + 1118, + 1401, + 1118, + 1401, + 1183, + 299, + 1183 + ], + "score": 0.617 + }, + { + "category_id": 1, + "poly": [ + 300, + 1036, + 1401, + 1036, + 1401, + 1101, + 300, + 1101 + ], + "score": 0.617 + }, + { + "category_id": 1, + "poly": [ + 299, + 1283, + 1399, + 1283, + 1399, + 1349, + 299, + 1349 + ], + "score": 0.613 + }, + { + "category_id": 1, + "poly": [ + 300, + 955, + 1400, + 955, + 1400, + 1020, + 300, + 1020 + ], + "score": 0.611 + }, + { + "category_id": 1, + "poly": [ + 297, + 1201, + 1401, + 1201, + 1401, + 1267, + 297, + 1267 + ], + "score": 0.605 + }, + { + "category_id": 1, + "poly": [ + 299, + 626, + 1398, + 626, + 1398, + 691, + 299, + 691 + ], + "score": 0.6 + }, + { + "category_id": 1, + "poly": [ + 298, + 873, + 1402, + 873, + 1402, + 938, + 298, + 938 + ], + "score": 0.59 + }, + { + "category_id": 1, + "poly": [ + 297, + 790, + 1404, + 790, + 1404, + 856, + 297, + 856 + ], + "score": 0.56 + }, + { + "category_id": 1, + "poly": [ + 299, + 1971, + 1398, + 1971, + 1398, + 2035, + 299, + 2035 + ], + "score": 0.536 + }, + { + "category_id": 1, + "poly": [ + 297, + 708, + 1400, + 708, + 1400, + 773, + 297, + 773 + ], + "score": 0.536 + }, + { + "category_id": 1, + "poly": [ + 301, + 1776, + 1400, + 1776, + 1400, + 1871, + 301, + 1871 + ], + "score": 0.525 + }, + { + "category_id": 1, + "poly": [ + 297, + 1889, + 1398, + 1889, + 1398, + 1953, + 297, + 1953 + ], + "score": 0.493 + }, + { + "category_id": 1, + "poly": [ + 296, + 1612, + 1405, + 1612, + 1405, + 1706, + 296, + 1706 + ], + "score": 0.283 + }, + { + "category_id": 13, + "poly": [ + 475, + 296, + 501, + 296, + 501, + 328, + 475, + 328 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 827, + 296, + 852, + 296, + 852, + 328, + 827, + 328 + ], + "score": 0.83, + "latex": "Q" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 298.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 219.0, + 550.0, + 219.0, + 550.0, + 271.0, + 292.0, + 271.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 493.0, + 491.0, + 493.0, + 491.0, + 533.0, + 296.0, + 533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2086.0, + 860.0, + 2086.0, + 860.0, + 2118.0, + 839.0, + 2118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 295.0, + 474.0, + 295.0, + 474.0, + 332.0, + 296.0, + 332.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 502.0, + 295.0, + 826.0, + 295.0, + 826.0, + 332.0, + 502.0, + 332.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 853.0, + 295.0, + 1405.0, + 295.0, + 1405.0, + 332.0, + 853.0, + 332.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 325.0, + 1405.0, + 325.0, + 1405.0, + 363.0, + 295.0, + 363.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 353.0, + 1410.0, + 353.0, + 1410.0, + 395.0, + 292.0, + 395.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 387.0, + 1405.0, + 387.0, + 1405.0, + 424.0, + 296.0, + 424.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 417.0, + 576.0, + 417.0, + 576.0, + 455.0, + 295.0, + 455.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 540.0, + 1407.0, + 540.0, + 1407.0, + 586.0, + 293.0, + 586.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 576.0, + 1338.0, + 576.0, + 1338.0, + 612.0, + 323.0, + 612.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1531.0, + 1404.0, + 1531.0, + 1404.0, + 1567.0, + 296.0, + 1567.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1562.0, + 1287.0, + 1562.0, + 1287.0, + 1595.0, + 322.0, + 1595.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1443.0, + 1404.0, + 1443.0, + 1404.0, + 1487.0, + 293.0, + 1487.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1479.0, + 597.0, + 1479.0, + 597.0, + 1510.0, + 323.0, + 1510.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1366.0, + 1406.0, + 1366.0, + 1406.0, + 1401.0, + 294.0, + 1401.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 1396.0, + 683.0, + 1396.0, + 683.0, + 1431.0, + 321.0, + 1431.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1723.0, + 1013.0, + 1723.0, + 1013.0, + 1763.0, + 295.0, + 1763.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1113.0, + 1405.0, + 1113.0, + 1405.0, + 1160.0, + 293.0, + 1160.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1150.0, + 502.0, + 1150.0, + 502.0, + 1184.0, + 323.0, + 1184.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1031.0, + 1405.0, + 1031.0, + 1405.0, + 1078.0, + 293.0, + 1078.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1068.0, + 504.0, + 1068.0, + 504.0, + 1101.0, + 324.0, + 1101.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1284.0, + 1404.0, + 1284.0, + 1404.0, + 1320.0, + 294.0, + 1320.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 1316.0, + 1250.0, + 1316.0, + 1250.0, + 1348.0, + 326.0, + 1348.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 955.0, + 1404.0, + 955.0, + 1404.0, + 991.0, + 296.0, + 991.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 986.0, + 1167.0, + 986.0, + 1167.0, + 1022.0, + 321.0, + 1022.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1198.0, + 1406.0, + 1198.0, + 1406.0, + 1243.0, + 292.0, + 1243.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1234.0, + 1046.0, + 1234.0, + 1046.0, + 1265.0, + 323.0, + 1265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 625.0, + 1404.0, + 625.0, + 1404.0, + 665.0, + 294.0, + 665.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 657.0, + 868.0, + 657.0, + 868.0, + 692.0, + 321.0, + 692.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 872.0, + 1406.0, + 872.0, + 1406.0, + 911.0, + 293.0, + 911.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 906.0, + 1187.0, + 906.0, + 1187.0, + 939.0, + 323.0, + 939.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 789.0, + 1408.0, + 789.0, + 1408.0, + 828.0, + 294.0, + 828.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 823.0, + 1314.0, + 823.0, + 1314.0, + 855.0, + 322.0, + 855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1968.0, + 1403.0, + 1968.0, + 1403.0, + 2008.0, + 294.0, + 2008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 2004.0, + 823.0, + 2004.0, + 823.0, + 2032.0, + 324.0, + 2032.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 708.0, + 1404.0, + 708.0, + 1404.0, + 746.0, + 293.0, + 746.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 325.0, + 739.0, + 1362.0, + 739.0, + 1362.0, + 775.0, + 325.0, + 775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1775.0, + 1406.0, + 1775.0, + 1406.0, + 1814.0, + 295.0, + 1814.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1809.0, + 1404.0, + 1809.0, + 1404.0, + 1843.0, + 322.0, + 1843.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1838.0, + 1140.0, + 1838.0, + 1140.0, + 1876.0, + 322.0, + 1876.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1888.0, + 1404.0, + 1888.0, + 1404.0, + 1927.0, + 293.0, + 1927.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1920.0, + 882.0, + 1920.0, + 882.0, + 1954.0, + 323.0, + 1954.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1609.0, + 1405.0, + 1609.0, + 1405.0, + 1649.0, + 293.0, + 1649.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1644.0, + 1405.0, + 1644.0, + 1405.0, + 1678.0, + 323.0, + 1678.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1675.0, + 611.0, + 1675.0, + 611.0, + 1704.0, + 323.0, + 1704.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 6, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 2, + "poly": [ + 298, + 75, + 857, + 75, + 857, + 106, + 298, + 106 + ], + "score": 0.896 + }, + { + "category_id": 2, + "poly": [ + 840, + 2088, + 859, + 2088, + 859, + 2112, + 840, + 2112 + ], + "score": 0.772 + }, + { + "category_id": 1, + "poly": [ + 291, + 228, + 1409, + 228, + 1409, + 1096, + 291, + 1096 + ], + "score": 0.672 + }, + { + "category_id": 15, + "poly": [ + 298.0, + 72.0, + 858.0, + 72.0, + 858.0, + 108.0, + 298.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2085.0, + 861.0, + 2085.0, + 861.0, + 2116.0, + 839.0, + 2116.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 230.0, + 1406.0, + 230.0, + 1406.0, + 269.0, + 293.0, + 269.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 262.0, + 946.0, + 262.0, + 946.0, + 296.0, + 322.0, + 296.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 310.0, + 1404.0, + 310.0, + 1404.0, + 352.0, + 292.0, + 352.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 344.0, + 930.0, + 344.0, + 930.0, + 381.0, + 322.0, + 381.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 396.0, + 1403.0, + 396.0, + 1403.0, + 435.0, + 294.0, + 435.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 428.0, + 1090.0, + 428.0, + 1090.0, + 463.0, + 322.0, + 463.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 477.0, + 1403.0, + 477.0, + 1403.0, + 518.0, + 292.0, + 518.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 512.0, + 1134.0, + 512.0, + 1134.0, + 547.0, + 322.0, + 547.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 561.0, + 1404.0, + 561.0, + 1404.0, + 600.0, + 294.0, + 600.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 595.0, + 1029.0, + 595.0, + 1029.0, + 629.0, + 321.0, + 629.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 646.0, + 1407.0, + 646.0, + 1407.0, + 681.0, + 296.0, + 681.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 693.0, + 1404.0, + 693.0, + 1404.0, + 738.0, + 294.0, + 738.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 730.0, + 1200.0, + 730.0, + 1200.0, + 765.0, + 323.0, + 765.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 778.0, + 1408.0, + 778.0, + 1408.0, + 821.0, + 292.0, + 821.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 808.0, + 1406.0, + 808.0, + 1406.0, + 852.0, + 320.0, + 852.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 844.0, + 1045.0, + 844.0, + 1045.0, + 879.0, + 322.0, + 879.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 890.0, + 1408.0, + 890.0, + 1408.0, + 934.0, + 292.0, + 934.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 927.0, + 813.0, + 927.0, + 813.0, + 962.0, + 323.0, + 962.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 976.0, + 1345.0, + 976.0, + 1345.0, + 1014.0, + 294.0, + 1014.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1029.0, + 1404.0, + 1029.0, + 1404.0, + 1069.0, + 293.0, + 1069.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1062.0, + 661.0, + 1062.0, + 661.0, + 1097.0, + 322.0, + 1097.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 7, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 5, + "poly": [ + 305, + 390, + 1462, + 390, + 1462, + 1926, + 305, + 1926 + ], + "score": 0.969, + "html": "
value
hyperparameter total training frames40 milliondescriptionsLength of training for each game.
minibatch size32
replay buffer size1000000parameter update.The number of most recent frames
agent history length4
length.concatenated as input to the Q net- work. Total number of iterations = total training frames /agent history
target network update10000 frequency
discount factor0.99
action repeat4
update frequency 4
optimizerAdamOptimizer for parameter updates.
β1 0.9Adam optimizer parameter.
β0.99Adam optimizer parameter.
E 10-4Adam optimizer parameter.
learning rate schedule 210-4 Interp(10-4,5 * 10-5) 5*10-5t≤106 otherwise t>5*106Learning rate for Adam optimizer, as a function of iteration t.
exploration scheduleInterp(1,0.1) Interp(0.1,0.01) 0.01t<106 otherwise Probability of random action in e- t>5*106
greedy exploration, as a function of the iteration t . Number of uniform random ac- tions taken before learning starts.
replay start size50000
" + }, + { + "category_id": 2, + "poly": [ + 297, + 75, + 857, + 75, + 857, + 105, + 297, + 105 + ], + "score": 0.911 + }, + { + "category_id": 0, + "poly": [ + 301, + 227, + 650, + 227, + 650, + 262, + 301, + 262 + ], + "score": 0.876 + }, + { + "category_id": 2, + "poly": [ + 840, + 2088, + 859, + 2088, + 859, + 2112, + 840, + 2112 + ], + "score": 0.799 + }, + { + "category_id": 1, + "poly": [ + 297, + 317, + 1344, + 317, + 1344, + 350, + 297, + 350 + ], + "score": 0.591 + }, + { + "category_id": 6, + "poly": [ + 297, + 1934, + 1405, + 1934, + 1405, + 2028, + 297, + 2028 + ], + "score": 0.448 + }, + { + "category_id": 6, + "poly": [ + 297, + 317, + 1344, + 317, + 1344, + 350, + 297, + 350 + ], + "score": 0.367 + }, + { + "category_id": 13, + "poly": [ + 1429, + 755, + 1454, + 755, + 1454, + 780, + 1429, + 780 + ], + "score": 0.76, + "latex": "=" + }, + { + "category_id": 13, + "poly": [ + 313, + 1422, + 347, + 1422, + 347, + 1455, + 313, + 1455 + ], + "score": 0.7, + "latex": "\\beta _ { 2 }" + }, + { + "category_id": 13, + "poly": [ + 1346, + 1613, + 1359, + 1613, + 1359, + 1637, + 1346, + 1637 + ], + "score": 0.68, + "latex": "t" + }, + { + "category_id": 13, + "poly": [ + 297, + 1995, + 431, + 1995, + 431, + 2029, + 297, + 2029 + ], + "score": 0.62, + "latex": "I n t e r p ( \\cdot , \\cdot )" + }, + { + "category_id": 13, + "poly": [ + 1238, + 1773, + 1254, + 1773, + 1254, + 1798, + 1238, + 1798 + ], + "score": 0.62, + "latex": "t" + }, + { + "category_id": 13, + "poly": [ + 313, + 1359, + 346, + 1359, + 346, + 1394, + 313, + 1394 + ], + "score": 0.59, + "latex": "\\beta _ { 1 }" + }, + { + "category_id": 13, + "poly": [ + 1430, + 1716, + 1446, + 1716, + 1446, + 1737, + 1430, + 1737 + ], + "score": 0.52, + "latex": "\\epsilon" + }, + { + "category_id": 13, + "poly": [ + 582, + 1480, + 645, + 1480, + 645, + 1514, + 582, + 1514 + ], + "score": 0.39, + "latex": "1 0 ^ { - 4 }" + }, + { + "category_id": 13, + "poly": [ + 904, + 1609, + 1035, + 1609, + 1035, + 1640, + 904, + 1640 + ], + "score": 0.35, + "latex": "t > 5 * 1 0 ^ { 6 }" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 297.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 227.0, + 653.0, + 227.0, + 653.0, + 267.0, + 295.0, + 267.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2086.0, + 861.0, + 2086.0, + 861.0, + 2117.0, + 839.0, + 2117.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1932.0, + 1405.0, + 1932.0, + 1405.0, + 1971.0, + 291.0, + 1971.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1966.0, + 1408.0, + 1966.0, + 1408.0, + 2000.0, + 295.0, + 2000.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 432.0, + 1996.0, + 910.0, + 1996.0, + 910.0, + 2029.0, + 432.0, + 2029.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 314.0, + 1347.0, + 314.0, + 1347.0, + 353.0, + 295.0, + 353.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 314.0, + 1347.0, + 314.0, + 1347.0, + 353.0, + 295.0, + 353.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 8, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 5, + "poly": [ + 308, + 344, + 1454, + 344, + 1454, + 1937, + 308, + 1937 + ], + "score": 0.978, + "html": "
UCB-Exploration
Bootstrapped DQNDoubleDQNEnsemble Voting 2282.8
Alien1445.12059.72817.6
Amidar430.58667.5683.72663.8
Assault2519.062820.613213.583702.76
Asterix3829.07639.58740.08732.0
Asteroids1009.51002.31149.31007.8
Atlantis1314058.01982677.01786305.02016145.0
Bank Heist795.1789.9869.4906.9
Battle Zone26230.024880.027430.026770.0
Beam Rider8006.587743.747991.99188.26
Bowling28.6230.9232.9238.06
Boxing85.9194.0794.4798.08
Breakout400.22467.45426.78411.31
Centipede5328.775177.516153.286237.18
Chopper Command2153.03260.03544.03677.0
Crazy Climber110926.0124456.0126677.0127754.0
Demon Attack9811.4523562.5530004.459861.9
Double Dunk-10.82-14.58-11.94-4.08
Enduro1314.311439.591999.882752.55
Fishing Derby21.8923.6930.0229.71
Freeway33.5732.9333.9233.96
Frostbite1284.8529.21196.01903.0
Gopher7652.212030.010993.212910.8
Gravitar227.5279.5371.5318.0
Ice Hockey-4.62-4.63-1.73-4.71
Jamesbond594.5594.0602.0710.0
Kangaroo8186.07787.08174.014196.0
Krull8537.528517.918669.179171.61
Kung Fu Master24153.032896.030988.031291.0
Montezuma Revenge2.04.01.04.0
Ms Pacman2508.72498.13039.73425.4
Name This Game8212.49806.99255.19570.5
Pitfall-5.99-7.57-3.37-1.47
Pong21.020.6721.020.95
Private Eye1815.19788.631845.281252.01
Qbert10557.256529.512036.514198.25
Riverraid11528.011834.712785.815622.2
Road Runner52489.049039.054768.053596.0
Robotank21.0329.831.8341.04
Seaquest9320.718056.420458.624001.6
Space Invaders1549.91917.51890.82626.55
Star Gunner20115.052283.041684.047367.0
Tennis-15.11-14.04-11.63-7.8
Time Pilot5088.05548.06153.06490.0
Tutankham167.47223.43208.61200.76
Up N Down9049.111815.319528.319827.3
Venture115.096.078.067.0
Video Pinball364600.85374686.89343380.29372564.11
Wizard Of Wor2860.03877.05451.05873.0
Zaxxon592.08903.03901.03695.0
Times best17930
" + }, + { + "category_id": 2, + "poly": [ + 298, + 75, + 857, + 75, + 857, + 105, + 298, + 105 + ], + "score": 0.915 + }, + { + "category_id": 2, + "poly": [ + 836, + 2088, + 864, + 2088, + 864, + 2113, + 836, + 2113 + ], + "score": 0.852 + }, + { + "category_id": 0, + "poly": [ + 300, + 228, + 601, + 228, + 601, + 261, + 300, + 261 + ], + "score": 0.373 + }, + { + "category_id": 7, + "poly": [ + 299, + 1966, + 1405, + 1966, + 1405, + 2029, + 299, + 2029 + ], + "score": 0.294 + }, + { + "category_id": 15, + "poly": [ + 298.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 298.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 869.0, + 2085.0, + 869.0, + 2123.0, + 831.0, + 2123.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 223.0, + 607.0, + 223.0, + 607.0, + 267.0, + 292.0, + 267.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1963.0, + 1405.0, + 1963.0, + 1405.0, + 2001.0, + 293.0, + 2001.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1994.0, + 1372.0, + 1994.0, + 1372.0, + 2032.0, + 293.0, + 2032.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 9, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 5, + "poly": [ + 416, + 210, + 1289, + 210, + 1289, + 1802, + 416, + 1802 + ], + "score": 0.973, + "html": "
Ensemble VotingUCB-ExplorationA3C+
Alien2282.82817.61848.33
Amidar683.72663.8964.77
Assault3213.583702.762607.28
Asterix8740.08732.07262.77
Asteroids1149.31007.82257.92
Atlantis1786305.02016145.01733528.71
Bank Heist869.4906.9991.96
Battle Zone27430.026770.07428.99
Beam Rider7991.99188.265992.08
Bowling32.9238.0668.72
Boxing94.4798.0813.82
Breakout426.78411.31323.21
Centipede6153.286237.185338.24
Chopper Command3544.03677.05388.22
Crazy Climber126677.0127754.0104083.51
Demon Attack30004.459861.919589.95
Double Dunk-11.94-4.08-8.88
Enduro1999.882752.55749.11
Fishing Derby30.0229.7129.46
Freeway33.9233.9627.33
Frostbite1196.01903.0506.61
Gopher10993.212910.85948.40
Gravitar371.5318.0246.02
Ice Hockey-1.73-4.71-7.05
Jamesbond602.0710.01024.16
Kangaroo8174.014196.05475.73
Krull8669.179171.617587.58
Kung Fu Master30988.031291.026593.67
Montezuma Revenge1.04.0142.50
Ms Pacman3039.73425.42380.58
Name This Game9255.19570.56427.51
Pitfall-3.37-1.47-155.97
Pong21.020.9517.33
Private Eye1845.281252.01100.0
Qbert12036.514198.2515804.72
Riverraid12785.815622.210331.56
Road Runner54768.053596.049029.74
Robotank31.8341.046.68
Seaquest20458.624001.62274.06
Space Invaders1890.82626.551466.01
Star Gunner41684.047367.052466.84
Tennis-11.63-7.8-20.49
Time Pilot6153.06490.03816.38
Tutankham208.61200.76132.67
Up N Down19528.319827.38705.64
Venture78.067.00.00
Video Pinball343380.29372564.1135515.92
Wizard Of Wor5451.05873.03657.65
Zaxxon3901.03695.07956.05
Times Best102810
" + }, + { + "category_id": 2, + "poly": [ + 298, + 75, + 857, + 75, + 857, + 105, + 298, + 105 + ], + "score": 0.858 + }, + { + "category_id": 2, + "poly": [ + 835, + 2088, + 861, + 2088, + 861, + 2113, + 835, + 2113 + ], + "score": 0.849 + }, + { + "category_id": 7, + "poly": [ + 294, + 1827, + 1403, + 1827, + 1403, + 1891, + 294, + 1891 + ], + "score": 0.421 + }, + { + "category_id": 13, + "poly": [ + 296, + 1859, + 369, + 1859, + 369, + 1888, + 296, + 1888 + ], + "score": 0.85, + "latex": "{ \\bf A } 3 { \\bf C } +" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 298.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 867.0, + 2085.0, + 867.0, + 2125.0, + 831.0, + 2125.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1827.0, + 1405.0, + 1827.0, + 1405.0, + 1863.0, + 294.0, + 1863.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 370.0, + 1855.0, + 1029.0, + 1855.0, + 1029.0, + 1894.0, + 370.0, + 1894.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 10, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 296, + 911, + 1405, + 911, + 1405, + 1126, + 296, + 1126 + ], + "score": 0.983 + }, + { + "category_id": 5, + "poly": [ + 297, + 222, + 1485, + 222, + 1485, + 385, + 297, + 385 + ], + "score": 0.981, + "html": "
CategoryTotalBootstrapped DQNDouble DQNEnsemble VotingUCB-Exploration
Human Optimal2303515
Score Explicit80215
Dense Reward80116
Sparse Reward51022
" + }, + { + "category_id": 1, + "poly": [ + 297, + 694, + 1405, + 694, + 1405, + 820, + 297, + 820 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 300, + 1707, + 1405, + 1707, + 1405, + 1800, + 300, + 1800 + ], + "score": 0.975 + }, + { + "category_id": 8, + "poly": [ + 420, + 1258, + 1274, + 1258, + 1274, + 1409, + 420, + 1409 + ], + "score": 0.97 + }, + { + "category_id": 8, + "poly": [ + 478, + 1576, + 1218, + 1576, + 1218, + 1699, + 478, + 1699 + ], + "score": 0.967 + }, + { + "category_id": 8, + "poly": [ + 585, + 1127, + 1109, + 1127, + 1109, + 1182, + 585, + 1182 + ], + "score": 0.953 + }, + { + "category_id": 8, + "poly": [ + 528, + 1808, + 1166, + 1808, + 1166, + 1883, + 528, + 1883 + ], + "score": 0.952 + }, + { + "category_id": 1, + "poly": [ + 297, + 1971, + 1404, + 1971, + 1404, + 2036, + 297, + 2036 + ], + "score": 0.947 + }, + { + "category_id": 1, + "poly": [ + 302, + 1418, + 1400, + 1418, + 1400, + 1481, + 302, + 1481 + ], + "score": 0.944 + }, + { + "category_id": 1, + "poly": [ + 299, + 1187, + 1402, + 1187, + 1402, + 1249, + 299, + 1249 + ], + "score": 0.943 + }, + { + "category_id": 8, + "poly": [ + 568, + 1489, + 1129, + 1489, + 1129, + 1528, + 568, + 1528 + ], + "score": 0.934 + }, + { + "category_id": 0, + "poly": [ + 296, + 626, + 1265, + 626, + 1265, + 665, + 296, + 665 + ], + "score": 0.916 + }, + { + "category_id": 2, + "poly": [ + 298, + 75, + 857, + 75, + 857, + 105, + 298, + 105 + ], + "score": 0.912 + }, + { + "category_id": 1, + "poly": [ + 295, + 1535, + 1280, + 1535, + 1280, + 1568, + 295, + 1568 + ], + "score": 0.903 + }, + { + "category_id": 0, + "poly": [ + 297, + 1914, + 840, + 1914, + 840, + 1948, + 297, + 1948 + ], + "score": 0.899 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1357, + 1400, + 1357, + 1400, + 1388, + 1366, + 1388 + ], + "score": 0.892 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1665, + 1400, + 1665, + 1400, + 1695, + 1366, + 1695 + ], + "score": 0.887 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1493, + 1400, + 1493, + 1400, + 1523, + 1366, + 1523 + ], + "score": 0.884 + }, + { + "category_id": 9, + "poly": [ + 1365, + 1831, + 1400, + 1831, + 1400, + 1860, + 1365, + 1860 + ], + "score": 0.877 + }, + { + "category_id": 2, + "poly": [ + 835, + 2087, + 864, + 2087, + 864, + 2113, + 835, + 2113 + ], + "score": 0.87 + }, + { + "category_id": 1, + "poly": [ + 297, + 418, + 1406, + 418, + 1406, + 574, + 297, + 574 + ], + "score": 0.856 + }, + { + "category_id": 0, + "poly": [ + 299, + 855, + 896, + 855, + 896, + 888, + 299, + 888 + ], + "score": 0.616 + }, + { + "category_id": 0, + "poly": [ + 298, + 855, + 895, + 855, + 895, + 888, + 298, + 888 + ], + "score": 0.323 + }, + { + "category_id": 14, + "poly": [ + 421, + 1255, + 1274, + 1255, + 1274, + 1413, + 421, + 1413 + ], + "score": 0.94, + "latex": "\\begin{array} { r l } & { \\tilde { p } ( Q ^ { * } , T | \\tau ) = \\frac { p ( \\tau | Q ^ { * } , T ) \\cdot p ( Q ^ { * } , T ) } { Z ( \\tau ) } } \\\\ & { \\qquad = \\frac { p ( Q ^ { * } , T ) \\cdot p ( s ^ { \\prime } | Q ^ { * } , T , ( s , a ) ) \\cdot p ( r | Q ^ { * } , T , ( s , a , s ^ { \\prime } ) ) \\cdot p ( s , a ) } { Z ( \\tau ) } , } \\end{array}" + }, + { + "category_id": 14, + "poly": [ + 530, + 1808, + 1169, + 1808, + 1169, + 1883, + 530, + 1883 + ], + "score": 0.94, + "latex": "\\tilde { p } ( Q ^ { * } , T | \\tau ) = \\frac { p ( Q ^ { * } , T ) \\cdot T ( s ^ { \\prime } | s , a ) \\cdot p ( s , a ) } { Z ( \\tau ) } \\cdot \\mathbb { 1 } ( Q ^ { * } , T ) ." + }, + { + "category_id": 14, + "poly": [ + 481, + 1574, + 1217, + 1574, + 1217, + 1703, + 481, + 1703 + ], + "score": 0.94, + "latex": "\\begin{array} { r l } & { p ( r | Q ^ { * } , T , ( s , a , s ^ { \\prime } ) ) = p ( r | Q ^ { * } , T , ( s , a ) ) } \\\\ & { \\phantom { p s p a c e } = \\mathbb { 1 } _ { \\{ Q ^ { * } ( s , a ) = r + \\gamma \\cdot \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\cdot \\vert s , a ) } \\operatorname* { m a x } _ { a ^ { \\prime } } Q ^ { * } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) \\} } } \\\\ & { \\phantom { p s p a c e } : = \\mathbb { 1 } ( Q ^ { * } , T ) , } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 955, + 1739, + 992, + 1739, + 992, + 1770, + 955, + 1770 + ], + "score": 0.93, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 715, + 1188, + 819, + 1188, + 819, + 1221, + 715, + 1221 + ], + "score": 0.93, + "latex": "p ( Q ^ { * } , T )" + }, + { + "category_id": 14, + "poly": [ + 588, + 1123, + 1110, + 1123, + 1110, + 1182, + 588, + 1182 + ], + "score": 0.92, + "latex": "Q ( s , a ) = r + \\mathbb { E } _ { s ^ { \\prime } \\sim T ( \\cdot | s , a ) } \\left[ \\gamma \\operatorname* { m a x } _ { a ^ { \\prime } } Q ( s ^ { \\prime } , a ^ { \\prime } ) \\right] ." + }, + { + "category_id": 13, + "poly": [ + 731, + 757, + 796, + 757, + 796, + 790, + 731, + 790 + ], + "score": 0.92, + "latex": "\\{ { \\bar { Q } } _ { k } \\}" + }, + { + "category_id": 13, + "poly": [ + 357, + 974, + 446, + 974, + 446, + 1008, + 357, + 1008 + ], + "score": 0.92, + "latex": "( Q ^ { * } , T )" + }, + { + "category_id": 13, + "poly": [ + 1200, + 1707, + 1306, + 1707, + 1306, + 1741, + 1200, + 1741 + ], + "score": 0.92, + "latex": "\\mathbb { 1 } ( Q ^ { * } , T )" + }, + { + "category_id": 13, + "poly": [ + 832, + 1066, + 920, + 1066, + 920, + 1099, + 832, + 1099 + ], + "score": 0.92, + "latex": "( Q ^ { * } , T )" + }, + { + "category_id": 13, + "poly": [ + 1071, + 974, + 1160, + 974, + 1160, + 1008, + 1071, + 1008 + ], + "score": 0.92, + "latex": "( Q ^ { * } , T )" + }, + { + "category_id": 13, + "poly": [ + 376, + 1418, + 436, + 1418, + 436, + 1452, + 376, + 1452 + ], + "score": 0.92, + "latex": "Z ( \\tau )" + }, + { + "category_id": 13, + "poly": [ + 298, + 1066, + 479, + 1066, + 479, + 1099, + 298, + 1099 + ], + "score": 0.92, + "latex": "{ \\boldsymbol \\tau } = ( s , a , r , s ^ { \\prime } )" + }, + { + "category_id": 13, + "poly": [ + 296, + 2003, + 385, + 2003, + 385, + 2036, + 296, + 2036 + ], + "score": 0.92, + "latex": "( Q ^ { * } , T )" + }, + { + "category_id": 13, + "poly": [ + 1109, + 1036, + 1170, + 1036, + 1170, + 1068, + 1109, + 1068 + ], + "score": 0.91, + "latex": "( s , a )" + }, + { + "category_id": 13, + "poly": [ + 919, + 1006, + 981, + 1006, + 981, + 1037, + 919, + 1037 + ], + "score": 0.91, + "latex": "( s , a )" + }, + { + "category_id": 13, + "poly": [ + 375, + 1709, + 422, + 1709, + 422, + 1743, + 375, + 1743 + ], + "score": 0.9, + "latex": "\\mathbb { 1 } _ { \\{ \\cdot \\} }" + }, + { + "category_id": 13, + "poly": [ + 415, + 1974, + 451, + 1974, + 451, + 2005, + 415, + 2005 + ], + "score": 0.9, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 876, + 975, + 913, + 975, + 913, + 1006, + 876, + 1006 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 711, + 790, + 748, + 790, + 748, + 819, + 711, + 819 + ], + "score": 0.89, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 1117, + 1068, + 1153, + 1068, + 1153, + 1097, + 1117, + 1097 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1013, + 698, + 1049, + 698, + 1049, + 728, + 1013, + 728 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 913, + 2007, + 948, + 2007, + 948, + 2035, + 913, + 2035 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 723, + 858, + 759, + 858, + 759, + 888, + 723, + 888 + ], + "score": 0.88, + "latex": "Q ^ { * }" + }, + { + "category_id": 14, + "poly": [ + 570, + 1488, + 1127, + 1488, + 1127, + 1528, + 570, + 1528 + ], + "score": 0.87, + "latex": "p ( s ^ { \\prime } | Q ^ { * } , T , ( s , a ) ) = p ( s ^ { \\prime } | T , ( s , a ) ) = T ( s ^ { \\prime } | s , a ) ," + }, + { + "category_id": 13, + "poly": [ + 661, + 1918, + 685, + 1918, + 685, + 1948, + 661, + 1948 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 578, + 759, + 603, + 759, + 603, + 789, + 578, + 789 + ], + "score": 0.86, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1047, + 629, + 1079, + 629, + 1079, + 666, + 1047, + 666 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1045, + 1740, + 1069, + 1740, + 1069, + 1766, + 1045, + 1766 + ], + "score": 0.85, + "latex": "T" + }, + { + "category_id": 13, + "poly": [ + 740, + 1036, + 763, + 1036, + 763, + 1063, + 740, + 1063 + ], + "score": 0.84, + "latex": "s ^ { \\prime }" + }, + { + "category_id": 13, + "poly": [ + 760, + 629, + 792, + 629, + 792, + 666, + 760, + 666 + ], + "score": 0.84, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 472, + 1451, + 493, + 1451, + 493, + 1477, + 472, + 1477 + ], + "score": 0.84, + "latex": "s" + }, + { + "category_id": 13, + "poly": [ + 852, + 915, + 875, + 915, + 875, + 941, + 852, + 941 + ], + "score": 0.82, + "latex": "T" + }, + { + "category_id": 13, + "poly": [ + 580, + 976, + 605, + 976, + 605, + 1002, + 580, + 1002 + ], + "score": 0.8, + "latex": "R" + }, + { + "category_id": 13, + "poly": [ + 990, + 976, + 1013, + 976, + 1013, + 1002, + 990, + 1002 + ], + "score": 0.8, + "latex": "T" + }, + { + "category_id": 13, + "poly": [ + 1051, + 1190, + 1068, + 1190, + 1068, + 1219, + 1051, + 1219 + ], + "score": 0.79, + "latex": "\\tilde { p }" + }, + { + "category_id": 13, + "poly": [ + 544, + 1451, + 570, + 1451, + 570, + 1477, + 544, + 1477 + ], + "score": 0.79, + "latex": "\\mathcal { A }" + }, + { + "category_id": 13, + "poly": [ + 1150, + 915, + 1174, + 915, + 1174, + 941, + 1150, + 941 + ], + "score": 0.79, + "latex": "R" + }, + { + "category_id": 13, + "poly": [ + 1159, + 1425, + 1177, + 1425, + 1177, + 1446, + 1159, + 1446 + ], + "score": 0.76, + "latex": "s" + }, + { + "category_id": 13, + "poly": [ + 1234, + 1426, + 1252, + 1426, + 1252, + 1446, + 1234, + 1446 + ], + "score": 0.76, + "latex": "a" + }, + { + "category_id": 13, + "poly": [ + 621, + 1775, + 640, + 1775, + 640, + 1796, + 621, + 1796 + ], + "score": 0.76, + "latex": "\\tau" + }, + { + "category_id": 13, + "poly": [ + 552, + 1041, + 570, + 1041, + 570, + 1063, + 552, + 1063 + ], + "score": 0.75, + "latex": "r" + }, + { + "category_id": 13, + "poly": [ + 779, + 1537, + 816, + 1537, + 816, + 1569, + 779, + 1569 + ], + "score": 0.62, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 745, + 1538, + 767, + 1538, + 767, + 1564, + 745, + 1564 + ], + "score": 0.5, + "latex": "T" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 618.0, + 759.0, + 618.0, + 759.0, + 675.0, + 291.0, + 675.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 793.0, + 618.0, + 1046.0, + 618.0, + 1046.0, + 675.0, + 793.0, + 675.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1080.0, + 618.0, + 1272.0, + 618.0, + 1272.0, + 675.0, + 1080.0, + 675.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 298.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1910.0, + 660.0, + 1910.0, + 660.0, + 1953.0, + 294.0, + 1953.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 686.0, + 1910.0, + 841.0, + 1910.0, + 841.0, + 1953.0, + 686.0, + 1953.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 869.0, + 2085.0, + 869.0, + 2125.0, + 831.0, + 2125.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 850.0, + 722.0, + 850.0, + 722.0, + 895.0, + 293.0, + 895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 760.0, + 850.0, + 899.0, + 850.0, + 899.0, + 895.0, + 760.0, + 895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 850.0, + 722.0, + 850.0, + 722.0, + 895.0, + 293.0, + 895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 760.0, + 850.0, + 900.0, + 850.0, + 900.0, + 895.0, + 760.0, + 895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 913.0, + 851.0, + 913.0, + 851.0, + 947.0, + 294.0, + 947.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 876.0, + 913.0, + 1149.0, + 913.0, + 1149.0, + 947.0, + 876.0, + 947.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1175.0, + 913.0, + 1405.0, + 913.0, + 1405.0, + 947.0, + 1175.0, + 947.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 942.0, + 1406.0, + 942.0, + 1406.0, + 978.0, + 293.0, + 978.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 972.0, + 356.0, + 972.0, + 356.0, + 1011.0, + 291.0, + 1011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 447.0, + 972.0, + 579.0, + 972.0, + 579.0, + 1011.0, + 447.0, + 1011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 606.0, + 972.0, + 875.0, + 972.0, + 875.0, + 1011.0, + 606.0, + 1011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 914.0, + 972.0, + 989.0, + 972.0, + 989.0, + 1011.0, + 914.0, + 1011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1014.0, + 972.0, + 1070.0, + 972.0, + 1070.0, + 1011.0, + 1014.0, + 1011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1161.0, + 972.0, + 1406.0, + 972.0, + 1406.0, + 1011.0, + 1161.0, + 1011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1001.0, + 918.0, + 1001.0, + 918.0, + 1041.0, + 293.0, + 1041.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 982.0, + 1001.0, + 1406.0, + 1001.0, + 1406.0, + 1041.0, + 982.0, + 1041.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1035.0, + 551.0, + 1035.0, + 551.0, + 1070.0, + 291.0, + 1070.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 571.0, + 1035.0, + 739.0, + 1035.0, + 739.0, + 1070.0, + 571.0, + 1070.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 764.0, + 1035.0, + 1108.0, + 1035.0, + 1108.0, + 1070.0, + 764.0, + 1070.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1171.0, + 1035.0, + 1406.0, + 1035.0, + 1406.0, + 1070.0, + 1171.0, + 1070.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1065.0, + 297.0, + 1065.0, + 297.0, + 1101.0, + 291.0, + 1101.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 480.0, + 1065.0, + 831.0, + 1065.0, + 831.0, + 1101.0, + 480.0, + 1101.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 921.0, + 1065.0, + 1116.0, + 1065.0, + 1116.0, + 1101.0, + 921.0, + 1101.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1154.0, + 1065.0, + 1406.0, + 1065.0, + 1406.0, + 1101.0, + 1154.0, + 1101.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1093.0, + 502.0, + 1093.0, + 502.0, + 1132.0, + 293.0, + 1132.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 697.0, + 1012.0, + 697.0, + 1012.0, + 730.0, + 295.0, + 730.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1050.0, + 697.0, + 1403.0, + 697.0, + 1403.0, + 730.0, + 1050.0, + 730.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 728.0, + 1403.0, + 728.0, + 1403.0, + 762.0, + 294.0, + 762.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 757.0, + 577.0, + 757.0, + 577.0, + 793.0, + 294.0, + 793.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 604.0, + 757.0, + 730.0, + 757.0, + 730.0, + 793.0, + 604.0, + 793.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 797.0, + 757.0, + 1406.0, + 757.0, + 1406.0, + 793.0, + 797.0, + 793.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 786.0, + 710.0, + 786.0, + 710.0, + 824.0, + 292.0, + 824.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 749.0, + 786.0, + 759.0, + 786.0, + 759.0, + 824.0, + 749.0, + 824.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1705.0, + 374.0, + 1705.0, + 374.0, + 1743.0, + 294.0, + 1743.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 423.0, + 1705.0, + 1199.0, + 1705.0, + 1199.0, + 1743.0, + 423.0, + 1743.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1307.0, + 1705.0, + 1409.0, + 1705.0, + 1409.0, + 1743.0, + 1307.0, + 1743.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1739.0, + 954.0, + 1739.0, + 954.0, + 1773.0, + 296.0, + 1773.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 993.0, + 1739.0, + 1044.0, + 1739.0, + 1044.0, + 1773.0, + 993.0, + 1773.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1070.0, + 1739.0, + 1405.0, + 1739.0, + 1405.0, + 1773.0, + 1070.0, + 1773.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1769.0, + 620.0, + 1769.0, + 620.0, + 1803.0, + 295.0, + 1803.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 641.0, + 1769.0, + 644.0, + 1769.0, + 644.0, + 1803.0, + 641.0, + 1803.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1971.0, + 414.0, + 1971.0, + 414.0, + 2007.0, + 295.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 452.0, + 1971.0, + 1406.0, + 1971.0, + 1406.0, + 2007.0, + 452.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 386.0, + 2001.0, + 912.0, + 2001.0, + 912.0, + 2037.0, + 386.0, + 2037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 949.0, + 2001.0, + 1402.0, + 2001.0, + 1402.0, + 2037.0, + 949.0, + 2037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1418.0, + 375.0, + 1418.0, + 375.0, + 1453.0, + 297.0, + 1453.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 437.0, + 1418.0, + 1158.0, + 1418.0, + 1158.0, + 1453.0, + 437.0, + 1453.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1178.0, + 1418.0, + 1233.0, + 1418.0, + 1233.0, + 1453.0, + 1178.0, + 1453.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1253.0, + 1418.0, + 1404.0, + 1418.0, + 1404.0, + 1453.0, + 1253.0, + 1453.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1448.0, + 471.0, + 1448.0, + 471.0, + 1485.0, + 296.0, + 1485.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 494.0, + 1448.0, + 543.0, + 1448.0, + 543.0, + 1485.0, + 494.0, + 1485.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 571.0, + 1448.0, + 1300.0, + 1448.0, + 1300.0, + 1485.0, + 571.0, + 1485.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1183.0, + 714.0, + 1183.0, + 714.0, + 1224.0, + 293.0, + 1224.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 820.0, + 1183.0, + 1050.0, + 1183.0, + 1050.0, + 1224.0, + 820.0, + 1224.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1069.0, + 1183.0, + 1406.0, + 1183.0, + 1406.0, + 1224.0, + 1069.0, + 1224.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1218.0, + 538.0, + 1218.0, + 538.0, + 1253.0, + 294.0, + 1253.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1533.0, + 744.0, + 1533.0, + 744.0, + 1573.0, + 294.0, + 1573.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 768.0, + 1533.0, + 778.0, + 1533.0, + 778.0, + 1573.0, + 768.0, + 1573.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 817.0, + 1533.0, + 1286.0, + 1533.0, + 1286.0, + 1573.0, + 817.0, + 1573.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 416.0, + 1404.0, + 416.0, + 1404.0, + 455.0, + 292.0, + 455.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 447.0, + 1408.0, + 447.0, + 1408.0, + 487.0, + 292.0, + 487.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 480.0, + 1406.0, + 480.0, + 1406.0, + 516.0, + 292.0, + 516.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 508.0, + 1406.0, + 508.0, + 1406.0, + 548.0, + 292.0, + 548.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 541.0, + 669.0, + 541.0, + 669.0, + 575.0, + 295.0, + 575.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 11, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1618, + 1404, + 1618, + 1404, + 1743, + 297, + 1743 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 298, + 1910, + 1404, + 1910, + 1404, + 2035, + 298, + 2035 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 296, + 227, + 1404, + 227, + 1404, + 355, + 296, + 355 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 296, + 1415, + 1405, + 1415, + 1405, + 1542, + 296, + 1542 + ], + "score": 0.975 + }, + { + "category_id": 1, + "poly": [ + 299, + 1138, + 1406, + 1138, + 1406, + 1233, + 299, + 1233 + ], + "score": 0.962 + }, + { + "category_id": 8, + "poly": [ + 666, + 748, + 1031, + 748, + 1031, + 807, + 666, + 807 + ], + "score": 0.953 + }, + { + "category_id": 1, + "poly": [ + 292, + 1246, + 1402, + 1246, + 1402, + 1310, + 292, + 1310 + ], + "score": 0.952 + }, + { + "category_id": 8, + "poly": [ + 572, + 444, + 1124, + 444, + 1124, + 484, + 572, + 484 + ], + "score": 0.949 + }, + { + "category_id": 1, + "poly": [ + 298, + 366, + 1403, + 366, + 1403, + 432, + 298, + 432 + ], + "score": 0.945 + }, + { + "category_id": 8, + "poly": [ + 360, + 541, + 1096, + 541, + 1096, + 614, + 360, + 614 + ], + "score": 0.936 + }, + { + "category_id": 8, + "poly": [ + 389, + 622, + 1331, + 622, + 1331, + 691, + 389, + 691 + ], + "score": 0.927 + }, + { + "category_id": 1, + "poly": [ + 299, + 494, + 1396, + 494, + 1396, + 529, + 299, + 529 + ], + "score": 0.926 + }, + { + "category_id": 8, + "poly": [ + 411, + 1555, + 1283, + 1555, + 1283, + 1603, + 411, + 1603 + ], + "score": 0.922 + }, + { + "category_id": 2, + "poly": [ + 297, + 74, + 857, + 74, + 857, + 105, + 297, + 105 + ], + "score": 0.922 + }, + { + "category_id": 1, + "poly": [ + 295, + 822, + 995, + 822, + 995, + 857, + 295, + 857 + ], + "score": 0.921 + }, + { + "category_id": 1, + "poly": [ + 298, + 702, + 735, + 702, + 735, + 737, + 298, + 737 + ], + "score": 0.919 + }, + { + "category_id": 0, + "poly": [ + 301, + 1842, + 717, + 1842, + 717, + 1879, + 301, + 1879 + ], + "score": 0.916 + }, + { + "category_id": 8, + "poly": [ + 304, + 1326, + 1382, + 1326, + 1382, + 1386, + 304, + 1386 + ], + "score": 0.896 + }, + { + "category_id": 9, + "poly": [ + 1352, + 1388, + 1400, + 1388, + 1400, + 1416, + 1352, + 1416 + ], + "score": 0.896 + }, + { + "category_id": 8, + "poly": [ + 309, + 1754, + 1380, + 1754, + 1380, + 1803, + 309, + 1803 + ], + "score": 0.894 + }, + { + "category_id": 9, + "poly": [ + 1365, + 752, + 1401, + 752, + 1401, + 782, + 1365, + 782 + ], + "score": 0.884 + }, + { + "category_id": 9, + "poly": [ + 1365, + 448, + 1400, + 448, + 1400, + 478, + 1365, + 478 + ], + "score": 0.876 + }, + { + "category_id": 9, + "poly": [ + 1365, + 1086, + 1401, + 1086, + 1401, + 1116, + 1365, + 1116 + ], + "score": 0.87 + }, + { + "category_id": 2, + "poly": [ + 835, + 2088, + 863, + 2088, + 863, + 2113, + 835, + 2113 + ], + "score": 0.866 + }, + { + "category_id": 9, + "poly": [ + 1366, + 565, + 1400, + 565, + 1400, + 594, + 1366, + 594 + ], + "score": 0.863 + }, + { + "category_id": 9, + "poly": [ + 1366, + 639, + 1401, + 639, + 1401, + 669, + 1366, + 669 + ], + "score": 0.829 + }, + { + "category_id": 8, + "poly": [ + 297, + 867, + 1426, + 867, + 1426, + 1126, + 297, + 1126 + ], + "score": 0.608 + }, + { + "category_id": 8, + "poly": [ + 306, + 1028, + 1422, + 1028, + 1422, + 1075, + 306, + 1075 + ], + "score": 0.547 + }, + { + "category_id": 8, + "poly": [ + 307, + 970, + 1422, + 970, + 1422, + 1019, + 307, + 1019 + ], + "score": 0.414 + }, + { + "category_id": 8, + "poly": [ + 295, + 870, + 1421, + 870, + 1421, + 964, + 295, + 964 + ], + "score": 0.257 + }, + { + "category_id": 8, + "poly": [ + 308, + 1087, + 1160, + 1087, + 1160, + 1127, + 308, + 1127 + ], + "score": 0.174 + }, + { + "category_id": 8, + "poly": [ + 295, + 911, + 1420, + 911, + 1420, + 963, + 295, + 963 + ], + "score": 0.116 + }, + { + "category_id": 8, + "poly": [ + 290, + 1026, + 1251, + 1026, + 1251, + 1126, + 290, + 1126 + ], + "score": 0.115 + }, + { + "category_id": 14, + "poly": [ + 666, + 747, + 1034, + 747, + 1034, + 808, + 666, + 808 + ], + "score": 0.94, + "latex": "Q _ { k , i + 1 } \\operatorname * { a r g m a x } _ { Q _ { k , i + 1 } } \\tilde { p } ( Q _ { k , i + 1 } | \\tau ) ." + }, + { + "category_id": 13, + "poly": [ + 366, + 399, + 557, + 399, + 557, + 434, + 366, + 434 + ], + "score": 0.94, + "latex": "Q ^ { * } = Q _ { k , i } ( \\cdot ; \\theta _ { k } )" + }, + { + "category_id": 14, + "poly": [ + 310, + 866, + 1426, + 866, + 1426, + 1130, + 310, + 1130 + ], + "score": 0.93, + "latex": "\\begin{array} { r l r } { { \\operatorname { i } ( Q _ { k , i + 1 } \\lvert \\tau \\rangle = p ( s , a ) \\cdot \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } \\mathbb { 1 } ( Q _ { k , i + 1 } , T ) } } \\\\ & { = p ( s , a ) \\cdot \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } \\underset { c \\to + \\infty } { \\operatorname { i m } } \\exp ( - c [ Q _ { k , i + 1 } ( s , a ) - r - \\gamma \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\lvert s , a ) } \\underset { a ^ { \\prime } } { \\operatorname { m a x } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ] ^ { 2 } ) } \\\\ & { = p ( s , a ) \\cdot \\underset { c \\to + \\infty } { \\operatorname* { i m } } \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } \\exp ( - c [ Q _ { k , i + 1 } ( s , a ) - r - \\gamma \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\lvert s , a ) } \\underset { a ^ { \\prime } } { \\operatorname { m a x } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ] ^ { 2 } ) } \\\\ & { \\geq p ( s , a ) \\cdot \\underset { c \\to + \\infty } { \\operatorname* { i m } } \\exp ( - c \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } [ Q _ { k , i + 1 } ( s , a ) - r - \\gamma \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\lvert s , a ) } \\underset { a ^ { \\prime } } { \\operatorname { m a x } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ] ^ { 2 } ) } \\\\ & { = p ( s , a ) \\cdot \\underset { c \\to + \\infty } { \\operatorname* { l i m } } \\underset { c \\to \\tau ( \\tau \\lvert Q _ { k , i } , \\tau ) } { \\operatorname* { i m } } [ Q _ { k , i + 1 } ( s , a ) - r - \\gamma \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot \\lvert s , a ) } \\underset { a ^ { \\prime } } { \\operatorname* { m a x } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ] ^ { 2 } } \\\\ & = p ( s , a ) \\cdot \\mathbb { E } _ { T \\sim \\tilde { p } ( T \\lvert Q _ { k , i } , \\tau ) } [ Q _ k , \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 631, + 368, + 696, + 368, + 696, + 402, + 631, + 402 + ], + "score": 0.93, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 1102, + 496, + 1182, + 496, + 1182, + 529, + 1102, + 529 + ], + "score": 0.92, + "latex": "Q _ { k , i + 1 }" + }, + { + "category_id": 13, + "poly": [ + 958, + 261, + 1024, + 261, + 1024, + 295, + 958, + 295 + ], + "score": 0.92, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 14, + "poly": [ + 312, + 1323, + 1385, + 1323, + 1385, + 1392, + 312, + 1392 + ], + "score": 0.92, + "latex": "Q _ { k , i + 1 } \\underset { Q _ { k , i + 1 } } { \\mathrm { a r g m i n } } \\mathbb { E } _ { T \\sim \\tilde { p } ( T | Q _ { k , i } , \\tau ) } [ Q _ { k , i + 1 } ( s , a ) - ( r + \\gamma \\cdot \\mathbb { E } _ { s ^ { \\prime \\prime } \\sim T ( \\cdot | s , a ) } \\operatorname* { m a x } _ { a ^ { \\prime } } Q _ { k , i + 1 } ( s ^ { \\prime \\prime } , a ^ { \\prime } ) ) ] ^ { 2 } ." + }, + { + "category_id": 13, + "poly": [ + 846, + 822, + 985, + 822, + 985, + 858, + 846, + 858 + ], + "score": 0.92, + "latex": "\\tilde { p } ( Q _ { k , i + 1 } | \\tau )" + }, + { + "category_id": 13, + "poly": [ + 1039, + 400, + 1128, + 400, + 1128, + 433, + 1039, + 433 + ], + "score": 0.92, + "latex": "( Q ^ { * } , T )" + }, + { + "category_id": 13, + "poly": [ + 506, + 703, + 585, + 703, + 585, + 736, + 506, + 736 + ], + "score": 0.92, + "latex": "Q _ { k , i + 1 }" + }, + { + "category_id": 14, + "poly": [ + 361, + 540, + 1335, + 540, + 1335, + 694, + 361, + 694 + ], + "score": 0.92, + "latex": "\\begin{array} { r l r } { { \\tilde { p } ( Q _ { k , i + 1 } , T | \\tau ) = \\frac { p ( T | Q _ { k , i } ) \\cdot T ( s ^ { \\prime } | s , a ) \\cdot p ( s , a ) } { Z ( \\tau ) } \\cdot \\mathbb { 1 } ( Q _ { k , i + 1 } , T ) . } } \\\\ & { } & { \\tilde { p } ( Q _ { k , i + 1 } | \\tau ) = \\int _ { T } \\tilde { p } ( Q _ { k , i + 1 } , T | \\tau ) \\mathrm { d } T = p ( s , a ) \\cdot \\int _ { T } \\tilde { p } ( T | Q _ { k , i } , \\tau ) \\cdot \\mathbb { 1 } ( Q _ { k , i + 1 } , T ) \\mathrm { d } T . } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 988, + 1682, + 1068, + 1682, + 1068, + 1715, + 988, + 1715 + ], + "score": 0.92, + "latex": "Q _ { k , i + 1 }" + }, + { + "category_id": 13, + "poly": [ + 1174, + 228, + 1280, + 228, + 1280, + 265, + 1174, + 265 + ], + "score": 0.92, + "latex": "\\{ Q _ { k } \\} _ { k = 1 } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 420, + 703, + 472, + 703, + 472, + 736, + 420, + 736 + ], + "score": 0.92, + "latex": "Q _ { k , i }" + }, + { + "category_id": 13, + "poly": [ + 1164, + 1479, + 1244, + 1479, + 1244, + 1511, + 1164, + 1511 + ], + "score": 0.92, + "latex": "Q _ { k , i + 1 }" + }, + { + "category_id": 13, + "poly": [ + 1241, + 1650, + 1359, + 1650, + 1359, + 1683, + 1241, + 1683 + ], + "score": 0.92, + "latex": "( s , a , r , s ^ { \\prime } )" + }, + { + "category_id": 13, + "poly": [ + 622, + 231, + 724, + 231, + 724, + 263, + 622, + 263 + ], + "score": 0.92, + "latex": "K \\in \\mathbb { N } _ { + }" + }, + { + "category_id": 13, + "poly": [ + 1063, + 368, + 1235, + 368, + 1235, + 402, + 1063, + 402 + ], + "score": 0.92, + "latex": "\\tau = ( s , a , r , s ^ { \\prime } )" + }, + { + "category_id": 13, + "poly": [ + 298, + 1447, + 440, + 1447, + 440, + 1481, + 298, + 1481 + ], + "score": 0.92, + "latex": "\\tilde { p } ( T | Q _ { k , i } , \\tau )" + }, + { + "category_id": 13, + "poly": [ + 646, + 1510, + 726, + 1510, + 726, + 1543, + 646, + 1543 + ], + "score": 0.92, + "latex": "Q _ { k , i + 1 }" + }, + { + "category_id": 13, + "poly": [ + 901, + 1682, + 953, + 1682, + 953, + 1714, + 901, + 1714 + ], + "score": 0.91, + "latex": "Q _ { k , i }" + }, + { + "category_id": 14, + "poly": [ + 414, + 1554, + 1282, + 1554, + 1282, + 1606, + 414, + 1606 + ], + "score": 0.91, + "latex": "\\theta _ { k } \\theta _ { k } + \\eta \\cdot ( Q _ { k } ( s , a ; \\theta _ { k } ) - ( r + \\gamma \\cdot \\operatorname* { m a x } _ { a ^ { \\prime } } Q _ { k } ( s ^ { \\prime } , a ^ { \\prime } ; \\theta _ { k } ) ) ) \\nabla _ { \\theta _ { k } } Q _ { k } ( s , a ; \\theta _ { k } ) ," + }, + { + "category_id": 14, + "poly": [ + 313, + 1752, + 1381, + 1752, + 1381, + 1805, + 313, + 1805 + ], + "score": 0.91, + "latex": "\\theta _ { k } \\gets \\theta _ { k } + \\eta \\cdot \\mathbb { E } _ { ( s , a , r , s ^ { \\prime } ) \\in B _ { \\operatorname* { m i n } } } \\big [ \\big ( Q _ { k } \\big ( s , a ; \\theta _ { k } \\big ) - \\big ( r + \\gamma \\cdot \\operatorname* { m a x } _ { a ^ { \\prime } } Q _ { k } \\big ( s ^ { \\prime } , a ^ { \\prime } ; \\theta _ { k } \\big ) \\big ) \\big ) \\nabla _ { \\theta _ { k } } Q _ { k } \\big ( s , a ; \\theta _ { k } \\big ) \\big ] ." + }, + { + "category_id": 13, + "poly": [ + 544, + 1943, + 581, + 1943, + 581, + 1974, + 544, + 1974 + ], + "score": 0.9, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1009, + 1651, + 1078, + 1651, + 1078, + 1681, + 1009, + 1681 + ], + "score": 0.9, + "latex": "B _ { \\mathrm { m i n i } }" + }, + { + "category_id": 13, + "poly": [ + 433, + 232, + 470, + 232, + 470, + 263, + 433, + 263 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 1311, + 497, + 1371, + 497, + 1371, + 525, + 1311, + 525 + ], + "score": 0.89, + "latex": "i + 1" + }, + { + "category_id": 13, + "poly": [ + 800, + 1620, + 838, + 1620, + 838, + 1652, + 800, + 1652 + ], + "score": 0.89, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 1016, + 231, + 1054, + 231, + 1054, + 262, + 1016, + 262 + ], + "score": 0.89, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 355, + 322, + 393, + 322, + 393, + 354, + 355, + 354 + ], + "score": 0.89, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 747, + 401, + 777, + 401, + 777, + 430, + 747, + 430 + ], + "score": 0.88, + "latex": "\\theta _ { k }" + }, + { + "category_id": 13, + "poly": [ + 663, + 1478, + 686, + 1478, + 686, + 1505, + 663, + 1505 + ], + "score": 0.88, + "latex": "s ^ { \\prime }" + }, + { + "category_id": 14, + "poly": [ + 573, + 444, + 1125, + 444, + 1125, + 484, + 573, + 484 + ], + "score": 0.87, + "latex": "p ( Q _ { k , i } , T ) = p ( Q ^ { * } , T | Q ^ { * } = Q _ { k , i } ) = p ( T | Q _ { k , i } ) ." + }, + { + "category_id": 13, + "poly": [ + 373, + 1653, + 399, + 1653, + 399, + 1678, + 373, + 1678 + ], + "score": 0.83, + "latex": "B" + }, + { + "category_id": 13, + "poly": [ + 372, + 1626, + 391, + 1626, + 391, + 1651, + 372, + 1651 + ], + "score": 0.8, + "latex": "\\eta" + }, + { + "category_id": 13, + "poly": [ + 623, + 1449, + 646, + 1449, + 646, + 1475, + 623, + 1475 + ], + "score": 0.79, + "latex": "T" + }, + { + "category_id": 13, + "poly": [ + 749, + 1483, + 767, + 1483, + 767, + 1505, + 749, + 1505 + ], + "score": 0.73, + "latex": "\\tau" + }, + { + "category_id": 13, + "poly": [ + 1385, + 372, + 1396, + 372, + 1396, + 396, + 1385, + 396 + ], + "score": 0.6, + "latex": "i" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 72.0, + 859.0, + 72.0, + 859.0, + 108.0, + 297.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1839.0, + 722.0, + 1839.0, + 722.0, + 1885.0, + 292.0, + 1885.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 868.0, + 2085.0, + 868.0, + 2123.0, + 831.0, + 2123.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1614.0, + 371.0, + 1614.0, + 371.0, + 1657.0, + 291.0, + 1657.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 1614.0, + 799.0, + 1614.0, + 799.0, + 1657.0, + 392.0, + 1657.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 1614.0, + 1406.0, + 1614.0, + 1406.0, + 1657.0, + 839.0, + 1657.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1649.0, + 372.0, + 1649.0, + 372.0, + 1684.0, + 294.0, + 1684.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 400.0, + 1649.0, + 1008.0, + 1649.0, + 1008.0, + 1684.0, + 400.0, + 1684.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1079.0, + 1649.0, + 1240.0, + 1649.0, + 1240.0, + 1684.0, + 1079.0, + 1684.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1360.0, + 1649.0, + 1406.0, + 1649.0, + 1406.0, + 1684.0, + 1360.0, + 1684.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1679.0, + 900.0, + 1679.0, + 900.0, + 1716.0, + 294.0, + 1716.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 954.0, + 1679.0, + 987.0, + 1679.0, + 987.0, + 1716.0, + 954.0, + 1716.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1069.0, + 1679.0, + 1405.0, + 1679.0, + 1405.0, + 1716.0, + 1069.0, + 1716.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1710.0, + 382.0, + 1710.0, + 382.0, + 1747.0, + 291.0, + 1747.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1909.0, + 1404.0, + 1909.0, + 1404.0, + 1946.0, + 293.0, + 1946.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1943.0, + 543.0, + 1943.0, + 543.0, + 1976.0, + 294.0, + 1976.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 582.0, + 1943.0, + 1405.0, + 1943.0, + 1405.0, + 1976.0, + 582.0, + 1976.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1972.0, + 1408.0, + 1972.0, + 1408.0, + 2009.0, + 294.0, + 2009.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1999.0, + 1096.0, + 1999.0, + 1096.0, + 2041.0, + 293.0, + 2041.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 290.0, + 222.0, + 432.0, + 222.0, + 432.0, + 274.0, + 290.0, + 274.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 471.0, + 222.0, + 621.0, + 222.0, + 621.0, + 274.0, + 471.0, + 274.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 725.0, + 222.0, + 1015.0, + 222.0, + 1015.0, + 274.0, + 725.0, + 274.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1055.0, + 222.0, + 1173.0, + 222.0, + 1173.0, + 274.0, + 1055.0, + 274.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1281.0, + 222.0, + 1411.0, + 222.0, + 1411.0, + 274.0, + 1281.0, + 274.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 260.0, + 957.0, + 260.0, + 957.0, + 297.0, + 293.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1025.0, + 260.0, + 1405.0, + 260.0, + 1405.0, + 297.0, + 1025.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 291.0, + 1405.0, + 291.0, + 1405.0, + 328.0, + 294.0, + 328.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 315.0, + 354.0, + 315.0, + 354.0, + 360.0, + 293.0, + 360.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 315.0, + 408.0, + 315.0, + 408.0, + 360.0, + 394.0, + 360.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1415.0, + 1405.0, + 1415.0, + 1405.0, + 1450.0, + 294.0, + 1450.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 441.0, + 1447.0, + 622.0, + 1447.0, + 622.0, + 1483.0, + 441.0, + 1483.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 647.0, + 1447.0, + 1406.0, + 1447.0, + 1406.0, + 1483.0, + 647.0, + 1483.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1476.0, + 662.0, + 1476.0, + 662.0, + 1512.0, + 291.0, + 1512.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 687.0, + 1476.0, + 748.0, + 1476.0, + 748.0, + 1512.0, + 687.0, + 1512.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 768.0, + 1476.0, + 1163.0, + 1476.0, + 1163.0, + 1512.0, + 768.0, + 1512.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1245.0, + 1476.0, + 1408.0, + 1476.0, + 1408.0, + 1512.0, + 1245.0, + 1512.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1508.0, + 645.0, + 1508.0, + 645.0, + 1545.0, + 294.0, + 1545.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 727.0, + 1508.0, + 1129.0, + 1508.0, + 1129.0, + 1545.0, + 727.0, + 1545.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1139.0, + 1404.0, + 1139.0, + 1404.0, + 1173.0, + 294.0, + 1173.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1170.0, + 1406.0, + 1170.0, + 1406.0, + 1204.0, + 294.0, + 1204.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1200.0, + 937.0, + 1200.0, + 937.0, + 1235.0, + 294.0, + 1235.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1242.0, + 1405.0, + 1242.0, + 1405.0, + 1284.0, + 292.0, + 1284.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1277.0, + 1311.0, + 1277.0, + 1311.0, + 1313.0, + 294.0, + 1313.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 363.0, + 630.0, + 363.0, + 630.0, + 406.0, + 293.0, + 406.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 697.0, + 363.0, + 1062.0, + 363.0, + 1062.0, + 406.0, + 697.0, + 406.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1236.0, + 363.0, + 1384.0, + 363.0, + 1384.0, + 406.0, + 1236.0, + 406.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1397.0, + 363.0, + 1410.0, + 363.0, + 1410.0, + 406.0, + 1397.0, + 406.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 398.0, + 365.0, + 398.0, + 365.0, + 435.0, + 292.0, + 435.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 558.0, + 398.0, + 746.0, + 398.0, + 746.0, + 435.0, + 558.0, + 435.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 778.0, + 398.0, + 1038.0, + 398.0, + 1038.0, + 435.0, + 778.0, + 435.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1129.0, + 398.0, + 1267.0, + 398.0, + 1267.0, + 435.0, + 1129.0, + 435.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 490.0, + 1101.0, + 490.0, + 1101.0, + 533.0, + 294.0, + 533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1183.0, + 490.0, + 1310.0, + 490.0, + 1310.0, + 533.0, + 1183.0, + 533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1372.0, + 490.0, + 1405.0, + 490.0, + 1405.0, + 533.0, + 1372.0, + 533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 814.0, + 845.0, + 814.0, + 845.0, + 863.0, + 293.0, + 863.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 986.0, + 814.0, + 997.0, + 814.0, + 997.0, + 863.0, + 986.0, + 863.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 695.0, + 419.0, + 695.0, + 419.0, + 744.0, + 293.0, + 744.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 473.0, + 695.0, + 505.0, + 695.0, + 505.0, + 744.0, + 473.0, + 744.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 586.0, + 695.0, + 736.0, + 695.0, + 736.0, + 744.0, + 586.0, + 744.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 12, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 675, + 1405, + 675, + 1405, + 832, + 297, + 832 + ], + "score": 0.978 + }, + { + "category_id": 1, + "poly": [ + 297, + 1353, + 1405, + 1353, + 1405, + 1479, + 297, + 1479 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 297, + 508, + 1405, + 508, + 1405, + 603, + 297, + 603 + ], + "score": 0.976 + }, + { + "category_id": 8, + "poly": [ + 649, + 431, + 1047, + 431, + 1047, + 495, + 649, + 495 + ], + "score": 0.963 + }, + { + "category_id": 8, + "poly": [ + 616, + 1141, + 1080, + 1141, + 1080, + 1208, + 616, + 1208 + ], + "score": 0.959 + }, + { + "category_id": 8, + "poly": [ + 634, + 850, + 1065, + 850, + 1065, + 935, + 634, + 935 + ], + "score": 0.957 + }, + { + "category_id": 8, + "poly": [ + 665, + 1001, + 1033, + 1001, + 1033, + 1069, + 665, + 1069 + ], + "score": 0.955 + }, + { + "category_id": 1, + "poly": [ + 294, + 1491, + 1402, + 1491, + 1402, + 1556, + 294, + 1556 + ], + "score": 0.955 + }, + { + "category_id": 1, + "poly": [ + 295, + 347, + 1397, + 347, + 1397, + 411, + 295, + 411 + ], + "score": 0.954 + }, + { + "category_id": 8, + "poly": [ + 493, + 296, + 1202, + 296, + 1202, + 335, + 493, + 335 + ], + "score": 0.946 + }, + { + "category_id": 8, + "poly": [ + 668, + 1257, + 1033, + 1257, + 1033, + 1296, + 668, + 1296 + ], + "score": 0.942 + }, + { + "category_id": 1, + "poly": [ + 292, + 228, + 1402, + 228, + 1402, + 294, + 292, + 294 + ], + "score": 0.942 + }, + { + "category_id": 8, + "poly": [ + 579, + 620, + 1120, + 620, + 1120, + 661, + 579, + 661 + ], + "score": 0.939 + }, + { + "category_id": 1, + "poly": [ + 290, + 950, + 1320, + 950, + 1320, + 983, + 290, + 983 + ], + "score": 0.932 + }, + { + "category_id": 1, + "poly": [ + 299, + 1223, + 558, + 1223, + 558, + 1254, + 299, + 1254 + ], + "score": 0.932 + }, + { + "category_id": 2, + "poly": [ + 297, + 74, + 857, + 74, + 857, + 106, + 297, + 106 + ], + "score": 0.927 + }, + { + "category_id": 1, + "poly": [ + 294, + 1086, + 1284, + 1086, + 1284, + 1123, + 294, + 1123 + ], + "score": 0.926 + }, + { + "category_id": 1, + "poly": [ + 298, + 1307, + 1122, + 1307, + 1122, + 1340, + 298, + 1340 + ], + "score": 0.924 + }, + { + "category_id": 9, + "poly": [ + 1352, + 1260, + 1400, + 1260, + 1400, + 1291, + 1352, + 1291 + ], + "score": 0.908 + }, + { + "category_id": 9, + "poly": [ + 1352, + 447, + 1399, + 447, + 1399, + 479, + 1352, + 479 + ], + "score": 0.904 + }, + { + "category_id": 9, + "poly": [ + 1352, + 1160, + 1400, + 1160, + 1400, + 1191, + 1352, + 1191 + ], + "score": 0.897 + }, + { + "category_id": 2, + "poly": [ + 835, + 2088, + 863, + 2088, + 863, + 2112, + 835, + 2112 + ], + "score": 0.865 + }, + { + "category_id": 1, + "poly": [ + 305, + 1632, + 1377, + 1632, + 1377, + 2008, + 305, + 2008 + ], + "score": 0.759 + }, + { + "category_id": 0, + "poly": [ + 297, + 1592, + 990, + 1592, + 990, + 1625, + 297, + 1625 + ], + "score": 0.732 + }, + { + "category_id": 5, + "poly": [ + 305, + 1632, + 1377, + 1632, + 1377, + 2008, + 305, + 2008 + ], + "score": 0.224, + "html": "
1: : Input: Value function networks Q with K outputs {Qk}K=1. Hyperparameters T,X,and p.
2: Let B be a replay buffer storing experience for training. for each episode do
3:
4:Obtain initial state from environment so
5:for step t = 1,... until end of episode do
6:Pick an action according to at ∈ argmaxa{μ(st,a) + λ ·δ(st,a)}
7:Receive state St+1 and reward rt from environment, having taken action at
8:Calculate exploration bonus br(st） according to (12)
9:Add (St,at,rt + ρ : br(St), St+1) to replay buffer B
10:At learning interval, sample random minibatch and update {Qk}
11: end for 12: end for
" + }, + { + "category_id": 14, + "poly": [ + 633, + 849, + 1064, + 849, + 1064, + 934, + 633, + 934 + ], + "score": 0.94, + "latex": "P _ { \\mathsf { T } , k } ( a | s ) = \\frac { \\exp \\left( Q _ { k } ( s , a ) / \\mathsf { T } \\right) } { \\sum _ { a ^ { \\prime } } \\exp \\left( Q _ { k } ( s , a ^ { \\prime } ) / \\mathsf { T } \\right) } ," + }, + { + "category_id": 14, + "poly": [ + 647, + 429, + 1051, + 429, + 1051, + 494, + 647, + 494 + ], + "score": 0.94, + "latex": "H _ { \\tau _ { 1 } , \\dots , \\tau _ { N } } = \\sum _ { n = 1 } ^ { N } H _ { \\tau _ { n } | \\tau _ { 1 } , \\dots , \\tau _ { n - 1 } } ." + }, + { + "category_id": 14, + "poly": [ + 663, + 1000, + 1036, + 1000, + 1036, + 1068, + 663, + 1068 + ], + "score": 0.94, + "latex": "P _ { \\mathsf { T } , \\mathrm { a v g } } = \\frac { 1 } { K } \\cdot \\sum _ { k = 1 } ^ { K } P _ { \\mathsf { T } , k } ( a | s ) ." + }, + { + "category_id": 14, + "poly": [ + 667, + 1256, + 1031, + 1256, + 1031, + 1295, + 667, + 1295 + ], + "score": 0.94, + "latex": "\\hat { r } ( s , a , s ^ { \\prime } ) = r ( s , a ) + \\rho \\cdot b \\tau ( s ) ," + }, + { + "category_id": 14, + "poly": [ + 615, + 1139, + 1084, + 1139, + 1084, + 1208, + 615, + 1208 + ], + "score": 0.93, + "latex": "b _ { \\mathsf { T } } ( s ) = \\frac { 1 } { K } \\cdot \\sum _ { k = 1 } ^ { K } \\mathrm { D } _ { K L } [ P _ { \\mathsf { T } , k } | | P _ { \\mathsf { T } , \\mathrm { a v g } } ] ." + }, + { + "category_id": 13, + "poly": [ + 1052, + 1086, + 1175, + 1086, + 1175, + 1123, + 1052, + 1123 + ], + "score": 0.93, + "latex": "\\{ P _ { \\mathsf { T } , k } \\} _ { k = 1 } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 745, + 1848, + 819, + 1848, + 819, + 1880, + 745, + 1880 + ], + "score": 0.93, + "latex": "b _ { \\mathsf { T } } ( s _ { t } )" + }, + { + "category_id": 13, + "poly": [ + 380, + 1445, + 446, + 1445, + 446, + 1480, + 380, + 1480 + ], + "score": 0.93, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 371, + 1309, + 460, + 1309, + 460, + 1340, + 371, + 1340 + ], + "score": 0.93, + "latex": "\\rho \\in \\mathbb { R } _ { + }" + }, + { + "category_id": 13, + "poly": [ + 546, + 1354, + 620, + 1354, + 620, + 1388, + 546, + 1388 + ], + "score": 0.92, + "latex": "b _ { \\mathsf { T } } ( s _ { t } )" + }, + { + "category_id": 13, + "poly": [ + 831, + 738, + 895, + 738, + 895, + 772, + 831, + 772 + ], + "score": 0.92, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 483, + 1879, + 801, + 1879, + 801, + 1912, + 483, + 1912 + ], + "score": 0.92, + "latex": "( s _ { t } , a _ { t } , r _ { t } + \\rho \\cdot b _ { \\mathsf { T } } ( s _ { t } ) , s _ { t + 1 } )" + }, + { + "category_id": 13, + "poly": [ + 912, + 1632, + 1018, + 1632, + 1018, + 1669, + 912, + 1669 + ], + "score": 0.92, + "latex": "\\{ Q _ { k } \\} _ { k = 1 } ^ { K }" + }, + { + "category_id": 13, + "poly": [ + 882, + 1416, + 1008, + 1416, + 1008, + 1449, + 882, + 1449 + ], + "score": 0.91, + "latex": "b _ { \\mathsf { T } } ( s ) \\to 0" + }, + { + "category_id": 13, + "poly": [ + 655, + 1445, + 719, + 1445, + 719, + 1479, + 655, + 1479 + ], + "score": 0.91, + "latex": "b _ { \\mathsf { T } } ( s )" + }, + { + "category_id": 13, + "poly": [ + 1045, + 1354, + 1110, + 1354, + 1110, + 1387, + 1045, + 1387 + ], + "score": 0.91, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 617, + 678, + 725, + 678, + 725, + 711, + 617, + 711 + ], + "score": 0.91, + "latex": "H _ { \\tau _ { 1 } , \\dots , \\tau _ { N } }" + }, + { + "category_id": 13, + "poly": [ + 808, + 570, + 872, + 570, + 872, + 605, + 808, + 605 + ], + "score": 0.91, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 371, + 348, + 564, + 348, + 564, + 381, + 371, + 381 + ], + "score": 0.91, + "latex": "\\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { n } )" + }, + { + "category_id": 13, + "poly": [ + 915, + 769, + 1151, + 769, + 1151, + 803, + 915, + 803 + ], + "score": 0.91, + "latex": "\\mathsf { H } ( \\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { N } ) )" + }, + { + "category_id": 13, + "poly": [ + 943, + 1384, + 1008, + 1384, + 1008, + 1417, + 943, + 1417 + ], + "score": 0.9, + "latex": "\\{ Q _ { k } \\}" + }, + { + "category_id": 13, + "poly": [ + 1216, + 1384, + 1298, + 1384, + 1298, + 1419, + 1216, + 1419 + ], + "score": 0.9, + "latex": "\\{ P _ { \\top , k } \\}" + }, + { + "category_id": 14, + "poly": [ + 578, + 620, + 1117, + 620, + 1117, + 659, + 578, + 659 + ], + "score": 0.9, + "latex": "H _ { \\tau _ { 1 } , \\dots , \\tau _ { N } } + \\mathsf { H } ( \\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { N } ) ) = \\mathsf { H } ( p ( Q ^ { * } ) ) ," + }, + { + "category_id": 13, + "poly": [ + 907, + 349, + 944, + 349, + 944, + 380, + 907, + 380 + ], + "score": 0.9, + "latex": "Q ^ { * }" + }, + { + "category_id": 13, + "poly": [ + 664, + 711, + 729, + 711, + 729, + 742, + 664, + 742 + ], + "score": 0.9, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 296, + 380, + 431, + 380, + 431, + 412, + 296, + 412 + ], + "score": 0.9, + "latex": "\\left( \\tau _ { 1 } , \\dots , \\tau _ { n } \\right)" + }, + { + "category_id": 13, + "poly": [ + 372, + 678, + 423, + 678, + 423, + 711, + 372, + 711 + ], + "score": 0.9, + "latex": "\\mathsf { H } ( \\cdot )" + }, + { + "category_id": 13, + "poly": [ + 736, + 1785, + 1185, + 1785, + 1185, + 1823, + 736, + 1823 + ], + "score": 0.9, + "latex": "\\begin{array} { r } { \\grave { a _ { t } } \\in \\mathrm { a r g m a x } _ { a } \\left\\{ \\tilde { \\mu } ( s _ { t } , a ) + \\lambda \\cdot \\tilde { \\sigma } ( s _ { t } , a ) \\right\\} } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 371, + 951, + 444, + 951, + 444, + 979, + 371, + 979 + ], + "score": 0.89, + "latex": "\\mathsf T > 0" + }, + { + "category_id": 13, + "poly": [ + 657, + 801, + 695, + 801, + 695, + 832, + 657, + 832 + ], + "score": 0.89, + "latex": "Q _ { k }" + }, + { + "category_id": 13, + "poly": [ + 580, + 1823, + 633, + 1823, + 633, + 1849, + 580, + 1849 + ], + "score": 0.89, + "latex": "s _ { t + 1 }" + }, + { + "category_id": 14, + "poly": [ + 497, + 294, + 1199, + 294, + 1199, + 335, + 497, + 335 + ], + "score": 0.88, + "latex": "H _ { \\tau _ { t } | \\tau _ { 1 } , \\dots , \\tau _ { n - 1 } } = D _ { K L } ( \\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { n } ) | | \\tilde { p } ( Q ^ { * } | \\tau _ { 1 } , \\dots , \\tau _ { n - 1 } ) )" + }, + { + "category_id": 13, + "poly": [ + 486, + 1758, + 595, + 1758, + 595, + 1788, + 486, + 1788 + ], + "score": 0.88, + "latex": "t = 1 , \\dots" + }, + { + "category_id": 13, + "poly": [ + 1081, + 1908, + 1146, + 1908, + 1146, + 1942, + 1081, + 1942 + ], + "score": 0.87, + "latex": "\\left\\{ Q _ { k } \\right\\}" + }, + { + "category_id": 13, + "poly": [ + 1092, + 1386, + 1205, + 1386, + 1205, + 1415, + 1092, + 1415 + ], + "score": 0.86, + "latex": "\\mathsf { T } \\to + \\infty" + }, + { + "category_id": 13, + "poly": [ + 706, + 1635, + 731, + 1635, + 731, + 1666, + 706, + 1666 + ], + "score": 0.85, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1208, + 1089, + 1284, + 1089, + 1284, + 1123, + 1208, + 1123 + ], + "score": 0.83, + "latex": "P _ { \\mathrm { { T , a v g } } }" + }, + { + "category_id": 13, + "poly": [ + 1048, + 512, + 1065, + 512, + 1065, + 541, + 1048, + 541 + ], + "score": 0.83, + "latex": "\\tilde { p }" + }, + { + "category_id": 13, + "poly": [ + 767, + 1823, + 793, + 1823, + 793, + 1847, + 767, + 1847 + ], + "score": 0.83, + "latex": "r _ { t }" + }, + { + "category_id": 13, + "poly": [ + 791, + 1635, + 820, + 1635, + 820, + 1663, + 791, + 1663 + ], + "score": 0.82, + "latex": "K" + }, + { + "category_id": 13, + "poly": [ + 804, + 1733, + 833, + 1733, + 833, + 1757, + 804, + 1757 + ], + "score": 0.81, + "latex": "s _ { 0 }" + }, + { + "category_id": 13, + "poly": [ + 591, + 1528, + 610, + 1528, + 610, + 1556, + 591, + 1556 + ], + "score": 0.81, + "latex": "\\rho" + }, + { + "category_id": 13, + "poly": [ + 834, + 1595, + 859, + 1595, + 859, + 1626, + 834, + 1626 + ], + "score": 0.8, + "latex": "Q" + }, + { + "category_id": 13, + "poly": [ + 1233, + 1825, + 1259, + 1825, + 1259, + 1848, + 1233, + 1848 + ], + "score": 0.77, + "latex": "a _ { t }" + }, + { + "category_id": 13, + "poly": [ + 298, + 268, + 327, + 268, + 327, + 292, + 298, + 292 + ], + "score": 0.76, + "latex": "\\tau _ { n }" + }, + { + "category_id": 13, + "poly": [ + 985, + 1880, + 1011, + 1880, + 1011, + 1906, + 985, + 1906 + ], + "score": 0.75, + "latex": "B" + }, + { + "category_id": 13, + "poly": [ + 388, + 1667, + 414, + 1667, + 414, + 1693, + 388, + 1693 + ], + "score": 0.74, + "latex": "B" + }, + { + "category_id": 13, + "poly": [ + 1337, + 1641, + 1354, + 1641, + 1354, + 1667, + 1337, + 1667 + ], + "score": 0.73, + "latex": "\\rho" + }, + { + "category_id": 13, + "poly": [ + 489, + 1525, + 508, + 1525, + 508, + 1551, + 489, + 1551 + ], + "score": 0.73, + "latex": "\\lambda" + }, + { + "category_id": 13, + "poly": [ + 412, + 1385, + 434, + 1385, + 434, + 1412, + 412, + 1412 + ], + "score": 0.66, + "latex": "\\top" + }, + { + "category_id": 13, + "poly": [ + 1100, + 1417, + 1123, + 1417, + 1123, + 1443, + 1100, + 1443 + ], + "score": 0.62, + "latex": "\\top" + }, + { + "category_id": 13, + "poly": [ + 519, + 1524, + 541, + 1524, + 541, + 1551, + 519, + 1551 + ], + "score": 0.56, + "latex": "\\top" + }, + { + "category_id": 13, + "poly": [ + 1229, + 1635, + 1279, + 1635, + 1279, + 1666, + 1229, + 1666 + ], + "score": 0.55, + "latex": "\\tau , \\lambda" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 72.0, + 859.0, + 72.0, + 859.0, + 109.0, + 297.0, + 109.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2084.0, + 869.0, + 2084.0, + 869.0, + 2123.0, + 831.0, + 2123.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1583.0, + 833.0, + 1583.0, + 833.0, + 1637.0, + 292.0, + 1637.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 860.0, + 1583.0, + 994.0, + 1583.0, + 994.0, + 1637.0, + 860.0, + 1637.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 672.0, + 371.0, + 672.0, + 371.0, + 715.0, + 293.0, + 715.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 424.0, + 672.0, + 616.0, + 672.0, + 616.0, + 715.0, + 424.0, + 715.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 726.0, + 676.0, + 1406.0, + 676.0, + 1406.0, + 713.0, + 726.0, + 713.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 708.0, + 663.0, + 708.0, + 663.0, + 745.0, + 294.0, + 745.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 730.0, + 708.0, + 1406.0, + 708.0, + 1406.0, + 745.0, + 730.0, + 745.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 737.0, + 830.0, + 737.0, + 830.0, + 775.0, + 295.0, + 775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 896.0, + 737.0, + 1406.0, + 737.0, + 1406.0, + 775.0, + 896.0, + 775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 768.0, + 914.0, + 768.0, + 914.0, + 806.0, + 295.0, + 806.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1152.0, + 768.0, + 1405.0, + 768.0, + 1405.0, + 806.0, + 1152.0, + 806.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 796.0, + 656.0, + 796.0, + 656.0, + 837.0, + 292.0, + 837.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1350.0, + 545.0, + 1350.0, + 545.0, + 1392.0, + 291.0, + 1392.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 621.0, + 1350.0, + 1044.0, + 1350.0, + 1044.0, + 1392.0, + 621.0, + 1392.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1111.0, + 1350.0, + 1406.0, + 1350.0, + 1406.0, + 1392.0, + 1111.0, + 1392.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1384.0, + 411.0, + 1384.0, + 411.0, + 1421.0, + 294.0, + 1421.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 435.0, + 1384.0, + 942.0, + 1384.0, + 942.0, + 1421.0, + 435.0, + 1421.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1009.0, + 1384.0, + 1091.0, + 1384.0, + 1091.0, + 1421.0, + 1009.0, + 1421.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1206.0, + 1384.0, + 1215.0, + 1384.0, + 1215.0, + 1421.0, + 1206.0, + 1421.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1299.0, + 1384.0, + 1403.0, + 1384.0, + 1403.0, + 1421.0, + 1299.0, + 1421.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1416.0, + 881.0, + 1416.0, + 881.0, + 1449.0, + 295.0, + 1449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1009.0, + 1416.0, + 1099.0, + 1416.0, + 1099.0, + 1449.0, + 1009.0, + 1449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1124.0, + 1416.0, + 1403.0, + 1416.0, + 1403.0, + 1449.0, + 1124.0, + 1449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1444.0, + 379.0, + 1444.0, + 379.0, + 1483.0, + 295.0, + 1483.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 447.0, + 1444.0, + 654.0, + 1444.0, + 654.0, + 1483.0, + 447.0, + 1483.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 720.0, + 1444.0, + 817.0, + 1444.0, + 817.0, + 1483.0, + 720.0, + 1483.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 506.0, + 1047.0, + 506.0, + 1047.0, + 546.0, + 294.0, + 546.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1066.0, + 506.0, + 1407.0, + 506.0, + 1407.0, + 546.0, + 1066.0, + 546.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 540.0, + 1405.0, + 540.0, + 1405.0, + 574.0, + 296.0, + 574.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 569.0, + 807.0, + 569.0, + 807.0, + 607.0, + 294.0, + 607.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 873.0, + 569.0, + 996.0, + 569.0, + 996.0, + 607.0, + 873.0, + 607.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1491.0, + 1404.0, + 1491.0, + 1404.0, + 1527.0, + 294.0, + 1527.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1523.0, + 488.0, + 1523.0, + 488.0, + 1560.0, + 294.0, + 1560.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 509.0, + 1523.0, + 518.0, + 1523.0, + 518.0, + 1560.0, + 509.0, + 1560.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 542.0, + 1523.0, + 590.0, + 1523.0, + 590.0, + 1560.0, + 542.0, + 1560.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 611.0, + 1523.0, + 838.0, + 1523.0, + 838.0, + 1560.0, + 611.0, + 1560.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 343.0, + 370.0, + 343.0, + 370.0, + 385.0, + 293.0, + 385.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 565.0, + 343.0, + 906.0, + 343.0, + 906.0, + 385.0, + 565.0, + 385.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 945.0, + 343.0, + 1403.0, + 343.0, + 1403.0, + 385.0, + 945.0, + 385.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 375.0, + 295.0, + 375.0, + 295.0, + 415.0, + 291.0, + 415.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 432.0, + 375.0, + 771.0, + 375.0, + 771.0, + 415.0, + 432.0, + 415.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 229.0, + 1404.0, + 229.0, + 1404.0, + 265.0, + 295.0, + 265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 264.0, + 297.0, + 264.0, + 297.0, + 297.0, + 291.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 328.0, + 264.0, + 364.0, + 264.0, + 364.0, + 297.0, + 328.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 948.0, + 370.0, + 948.0, + 370.0, + 987.0, + 294.0, + 987.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 445.0, + 948.0, + 1324.0, + 948.0, + 1324.0, + 987.0, + 445.0, + 987.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1221.0, + 559.0, + 1221.0, + 559.0, + 1257.0, + 296.0, + 1257.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 290.0, + 1075.0, + 1051.0, + 1075.0, + 1051.0, + 1134.0, + 290.0, + 1134.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1176.0, + 1075.0, + 1207.0, + 1075.0, + 1207.0, + 1134.0, + 1176.0, + 1134.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1285.0, + 1075.0, + 1292.0, + 1075.0, + 1292.0, + 1134.0, + 1285.0, + 1134.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1303.0, + 370.0, + 1303.0, + 370.0, + 1346.0, + 293.0, + 1346.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 461.0, + 1303.0, + 1123.0, + 1303.0, + 1123.0, + 1346.0, + 461.0, + 1346.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 304.0, + 1627.0, + 705.0, + 1627.0, + 705.0, + 1677.0, + 304.0, + 1677.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 732.0, + 1627.0, + 790.0, + 1627.0, + 790.0, + 1677.0, + 732.0, + 1677.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 821.0, + 1627.0, + 911.0, + 1627.0, + 911.0, + 1677.0, + 821.0, + 1677.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1019.0, + 1627.0, + 1228.0, + 1627.0, + 1228.0, + 1677.0, + 1019.0, + 1677.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1280.0, + 1627.0, + 1336.0, + 1627.0, + 1336.0, + 1677.0, + 1280.0, + 1677.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1355.0, + 1627.0, + 1370.0, + 1627.0, + 1370.0, + 1677.0, + 1355.0, + 1677.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 311.0, + 1662.0, + 387.0, + 1662.0, + 387.0, + 1702.0, + 311.0, + 1702.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 415.0, + 1662.0, + 967.0, + 1662.0, + 967.0, + 1702.0, + 415.0, + 1702.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 310.0, + 1695.0, + 572.0, + 1695.0, + 572.0, + 1730.0, + 310.0, + 1730.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 1727.0, + 339.0, + 1727.0, + 339.0, + 1758.0, + 307.0, + 1758.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 1722.0, + 803.0, + 1722.0, + 803.0, + 1763.0, + 382.0, + 1763.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 834.0, + 1722.0, + 837.0, + 1722.0, + 837.0, + 1763.0, + 834.0, + 1763.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 1757.0, + 339.0, + 1757.0, + 339.0, + 1789.0, + 307.0, + 1789.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 383.0, + 1756.0, + 485.0, + 1756.0, + 485.0, + 1790.0, + 383.0, + 1790.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 596.0, + 1756.0, + 865.0, + 1756.0, + 865.0, + 1790.0, + 596.0, + 1790.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 1788.0, + 338.0, + 1788.0, + 338.0, + 1821.0, + 307.0, + 1821.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 421.0, + 1783.0, + 735.0, + 1783.0, + 735.0, + 1825.0, + 421.0, + 1825.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1816.0, + 340.0, + 1816.0, + 340.0, + 1852.0, + 306.0, + 1852.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 422.0, + 1813.0, + 579.0, + 1813.0, + 579.0, + 1856.0, + 422.0, + 1856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 634.0, + 1813.0, + 766.0, + 1813.0, + 766.0, + 1856.0, + 634.0, + 1856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 794.0, + 1813.0, + 1232.0, + 1813.0, + 1232.0, + 1856.0, + 794.0, + 1856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1260.0, + 1813.0, + 1265.0, + 1813.0, + 1265.0, + 1856.0, + 1260.0, + 1856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 307.0, + 1847.0, + 338.0, + 1847.0, + 338.0, + 1881.0, + 307.0, + 1881.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 424.0, + 1842.0, + 744.0, + 1842.0, + 744.0, + 1885.0, + 424.0, + 1885.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 820.0, + 1842.0, + 1022.0, + 1842.0, + 1022.0, + 1885.0, + 820.0, + 1885.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 305.0, + 1877.0, + 339.0, + 1877.0, + 339.0, + 1913.0, + 305.0, + 1913.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 422.0, + 1875.0, + 482.0, + 1875.0, + 482.0, + 1917.0, + 422.0, + 1917.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 802.0, + 1875.0, + 984.0, + 1875.0, + 984.0, + 1917.0, + 802.0, + 1917.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1012.0, + 1875.0, + 1018.0, + 1875.0, + 1018.0, + 1917.0, + 1012.0, + 1917.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 1906.0, + 340.0, + 1906.0, + 340.0, + 1943.0, + 298.0, + 1943.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 420.0, + 1903.0, + 1080.0, + 1903.0, + 1080.0, + 1948.0, + 420.0, + 1948.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1939.0, + 340.0, + 1939.0, + 340.0, + 1972.0, + 297.0, + 1972.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 1937.0, + 480.0, + 1937.0, + 480.0, + 1970.0, + 382.0, + 1970.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 334.0, + 1966.0, + 439.0, + 1966.0, + 439.0, + 2002.0, + 334.0, + 2002.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 1970.0, + 343.0, + 1970.0, + 343.0, + 2001.0, + 298.0, + 2001.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 13, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1350, + 1405, + 1350, + 1405, + 1627, + 297, + 1627 + ], + "score": 0.977 + }, + { + "category_id": 3, + "poly": [ + 318, + 227, + 1446, + 227, + 1446, + 913, + 318, + 913 + ], + "score": 0.964 + }, + { + "category_id": 1, + "poly": [ + 299, + 1165, + 1402, + 1165, + 1402, + 1259, + 299, + 1259 + ], + "score": 0.954 + }, + { + "category_id": 4, + "poly": [ + 297, + 957, + 1406, + 957, + 1406, + 1053, + 297, + 1053 + ], + "score": 0.945 + }, + { + "category_id": 1, + "poly": [ + 293, + 1272, + 1400, + 1272, + 1400, + 1335, + 293, + 1335 + ], + "score": 0.941 + }, + { + "category_id": 2, + "poly": [ + 297, + 76, + 857, + 76, + 857, + 104, + 297, + 104 + ], + "score": 0.889 + }, + { + "category_id": 0, + "poly": [ + 298, + 1108, + 977, + 1108, + 977, + 1141, + 298, + 1141 + ], + "score": 0.878 + }, + { + "category_id": 2, + "poly": [ + 835, + 2088, + 864, + 2088, + 864, + 2113, + 835, + 2113 + ], + "score": 0.849 + }, + { + "category_id": 3, + "poly": [ + 318, + 227, + 1446, + 227, + 1446, + 913, + 318, + 913 + ], + "score": 0.225 + }, + { + "category_id": 13, + "poly": [ + 1088, + 1227, + 1161, + 1227, + 1161, + 1255, + 1088, + 1255 + ], + "score": 0.89, + "latex": "{ \\mathsf T } = 1" + }, + { + "category_id": 13, + "poly": [ + 861, + 1227, + 1030, + 1227, + 1030, + 1259, + 861, + 1259 + ], + "score": 0.79, + "latex": "\\lambda = 0 . 1 , \\rho = 1" + }, + { + "category_id": 13, + "poly": [ + 658, + 1114, + 677, + 1114, + 677, + 1137, + 658, + 1137 + ], + "score": 0.48, + "latex": "^ +" + }, + { + "category_id": 13, + "poly": [ + 1311, + 1443, + 1334, + 1443, + 1334, + 1471, + 1311, + 1471 + ], + "score": 0.35, + "latex": "\\top" + }, + { + "category_id": 13, + "poly": [ + 1394, + 502, + 1432, + 502, + 1432, + 524, + 1394, + 524 + ], + "score": 0.34, + "latex": "\\mathsf { T } \\mathop { = } 1" + }, + { + "category_id": 15, + "poly": [ + 610.0, + 231.0, + 991.0, + 231.0, + 991.0, + 267.0, + 610.0, + 267.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 247.0, + 373.0, + 247.0, + 373.0, + 282.0, + 326.0, + 282.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 320.0, + 373.0, + 320.0, + 373.0, + 355.0, + 326.0, + 355.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1193.0, + 354.0, + 1207.0, + 354.0, + 1207.0, + 364.0, + 1193.0, + 364.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 392.0, + 374.0, + 392.0, + 374.0, + 430.0, + 324.0, + 430.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1121.0, + 430.0, + 1147.0, + 430.0, + 1147.0, + 443.0, + 1121.0, + 443.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1024.0, + 442.0, + 1038.0, + 442.0, + 1038.0, + 452.0, + 1024.0, + 452.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 464.0, + 375.0, + 464.0, + 375.0, + 504.0, + 324.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1282.0, + 475.0, + 1414.0, + 475.0, + 1414.0, + 510.0, + 1282.0, + 510.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1281.0, + 496.0, + 1393.0, + 496.0, + 1393.0, + 532.0, + 1281.0, + 532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1433.0, + 496.0, + 1441.0, + 496.0, + 1441.0, + 532.0, + 1433.0, + 532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 536.0, + 375.0, + 536.0, + 375.0, + 575.0, + 324.0, + 575.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1282.0, + 519.0, + 1451.0, + 519.0, + 1451.0, + 590.0, + 1282.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1284.0, + 575.0, + 1435.0, + 575.0, + 1435.0, + 613.0, + 1284.0, + 613.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 610.0, + 375.0, + 610.0, + 375.0, + 649.0, + 324.0, + 649.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 562.0, + 635.0, + 572.0, + 635.0, + 572.0, + 645.0, + 562.0, + 645.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1282.0, + 598.0, + 1399.0, + 598.0, + 1399.0, + 642.0, + 1282.0, + 642.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 537.0, + 648.0, + 547.0, + 648.0, + 547.0, + 658.0, + 537.0, + 658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 682.0, + 374.0, + 682.0, + 374.0, + 721.0, + 324.0, + 721.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 455.0, + 701.0, + 482.0, + 701.0, + 482.0, + 717.0, + 455.0, + 717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 445.0, + 723.0, + 455.0, + 723.0, + 455.0, + 735.0, + 445.0, + 735.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 427.0, + 737.0, + 447.0, + 737.0, + 447.0, + 757.0, + 427.0, + 757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 409.0, + 739.0, + 427.0, + 739.0, + 427.0, + 755.0, + 409.0, + 755.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 424.0, + 752.0, + 439.0, + 752.0, + 439.0, + 769.0, + 424.0, + 769.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 755.0, + 372.0, + 755.0, + 372.0, + 794.0, + 324.0, + 794.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 366.0, + 774.0, + 409.0, + 774.0, + 409.0, + 810.0, + 366.0, + 810.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 827.0, + 375.0, + 827.0, + 375.0, + 866.0, + 324.0, + 866.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 352.0, + 852.0, + 398.0, + 852.0, + 398.0, + 887.0, + 352.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 459.0, + 852.0, + 505.0, + 852.0, + 505.0, + 887.0, + 459.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 566.0, + 852.0, + 610.0, + 852.0, + 610.0, + 887.0, + 566.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 672.0, + 852.0, + 717.0, + 852.0, + 717.0, + 887.0, + 672.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 775.0, + 849.0, + 824.0, + 849.0, + 824.0, + 888.0, + 775.0, + 888.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 883.0, + 852.0, + 928.0, + 852.0, + 928.0, + 887.0, + 883.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 989.0, + 852.0, + 1034.0, + 852.0, + 1034.0, + 887.0, + 989.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1094.0, + 852.0, + 1140.0, + 852.0, + 1140.0, + 887.0, + 1094.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1201.0, + 852.0, + 1245.0, + 852.0, + 1245.0, + 887.0, + 1201.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 754.0, + 877.0, + 844.0, + 877.0, + 844.0, + 917.0, + 754.0, + 917.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1184.0, + 876.0, + 1234.0, + 876.0, + 1234.0, + 910.0, + 1184.0, + 910.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 378.0, + 762.0, + 413.0, + 762.0, + 413.0, + 775.0, + 378.0, + 775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 400.0, + 771.0, + 426.0, + 771.0, + 426.0, + 785.0, + 400.0, + 785.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 958.0, + 1404.0, + 958.0, + 1404.0, + 993.0, + 296.0, + 993.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 989.0, + 1404.0, + 989.0, + 1404.0, + 1023.0, + 295.0, + 1023.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1021.0, + 1338.0, + 1021.0, + 1338.0, + 1055.0, + 295.0, + 1055.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 297.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1108.0, + 657.0, + 1108.0, + 657.0, + 1144.0, + 296.0, + 1144.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 1108.0, + 980.0, + 1108.0, + 980.0, + 1144.0, + 678.0, + 1144.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 868.0, + 2085.0, + 868.0, + 2123.0, + 831.0, + 2123.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 610.0, + 231.0, + 991.0, + 231.0, + 991.0, + 267.0, + 610.0, + 267.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 247.0, + 373.0, + 247.0, + 373.0, + 282.0, + 326.0, + 282.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 320.0, + 373.0, + 320.0, + 373.0, + 355.0, + 326.0, + 355.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1193.0, + 354.0, + 1207.0, + 354.0, + 1207.0, + 364.0, + 1193.0, + 364.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 392.0, + 374.0, + 392.0, + 374.0, + 430.0, + 324.0, + 430.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1121.0, + 430.0, + 1147.0, + 430.0, + 1147.0, + 443.0, + 1121.0, + 443.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1024.0, + 442.0, + 1038.0, + 442.0, + 1038.0, + 452.0, + 1024.0, + 452.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 464.0, + 375.0, + 464.0, + 375.0, + 504.0, + 324.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1282.0, + 475.0, + 1414.0, + 475.0, + 1414.0, + 510.0, + 1282.0, + 510.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1281.0, + 496.0, + 1393.0, + 496.0, + 1393.0, + 532.0, + 1281.0, + 532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1433.0, + 496.0, + 1441.0, + 496.0, + 1441.0, + 532.0, + 1433.0, + 532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 536.0, + 375.0, + 536.0, + 375.0, + 575.0, + 324.0, + 575.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1282.0, + 519.0, + 1451.0, + 519.0, + 1451.0, + 590.0, + 1282.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1284.0, + 575.0, + 1435.0, + 575.0, + 1435.0, + 613.0, + 1284.0, + 613.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 610.0, + 375.0, + 610.0, + 375.0, + 649.0, + 324.0, + 649.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 562.0, + 635.0, + 572.0, + 635.0, + 572.0, + 645.0, + 562.0, + 645.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1282.0, + 598.0, + 1399.0, + 598.0, + 1399.0, + 642.0, + 1282.0, + 642.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 537.0, + 648.0, + 547.0, + 648.0, + 547.0, + 658.0, + 537.0, + 658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 682.0, + 374.0, + 682.0, + 374.0, + 721.0, + 324.0, + 721.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 455.0, + 701.0, + 482.0, + 701.0, + 482.0, + 717.0, + 455.0, + 717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 445.0, + 723.0, + 455.0, + 723.0, + 455.0, + 735.0, + 445.0, + 735.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 427.0, + 737.0, + 447.0, + 737.0, + 447.0, + 757.0, + 427.0, + 757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 409.0, + 739.0, + 427.0, + 739.0, + 427.0, + 755.0, + 409.0, + 755.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 424.0, + 752.0, + 439.0, + 752.0, + 439.0, + 769.0, + 424.0, + 769.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 755.0, + 372.0, + 755.0, + 372.0, + 794.0, + 324.0, + 794.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 366.0, + 774.0, + 409.0, + 774.0, + 409.0, + 810.0, + 366.0, + 810.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 827.0, + 375.0, + 827.0, + 375.0, + 866.0, + 324.0, + 866.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 352.0, + 852.0, + 398.0, + 852.0, + 398.0, + 887.0, + 352.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 459.0, + 852.0, + 505.0, + 852.0, + 505.0, + 887.0, + 459.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 566.0, + 852.0, + 610.0, + 852.0, + 610.0, + 887.0, + 566.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 672.0, + 852.0, + 717.0, + 852.0, + 717.0, + 887.0, + 672.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 775.0, + 849.0, + 824.0, + 849.0, + 824.0, + 888.0, + 775.0, + 888.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 883.0, + 852.0, + 928.0, + 852.0, + 928.0, + 887.0, + 883.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 989.0, + 852.0, + 1034.0, + 852.0, + 1034.0, + 887.0, + 989.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1094.0, + 852.0, + 1140.0, + 852.0, + 1140.0, + 887.0, + 1094.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1201.0, + 852.0, + 1245.0, + 852.0, + 1245.0, + 887.0, + 1201.0, + 887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 754.0, + 877.0, + 844.0, + 877.0, + 844.0, + 917.0, + 754.0, + 917.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1184.0, + 876.0, + 1234.0, + 876.0, + 1234.0, + 910.0, + 1184.0, + 910.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 378.0, + 762.0, + 413.0, + 762.0, + 413.0, + 775.0, + 378.0, + 775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 400.0, + 771.0, + 426.0, + 771.0, + 426.0, + 785.0, + 400.0, + 785.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1350.0, + 1405.0, + 1350.0, + 1405.0, + 1385.0, + 292.0, + 1385.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1382.0, + 1405.0, + 1382.0, + 1405.0, + 1415.0, + 296.0, + 1415.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1410.0, + 1405.0, + 1410.0, + 1405.0, + 1449.0, + 292.0, + 1449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1441.0, + 1310.0, + 1441.0, + 1310.0, + 1478.0, + 294.0, + 1478.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1335.0, + 1441.0, + 1405.0, + 1441.0, + 1405.0, + 1478.0, + 1335.0, + 1478.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1473.0, + 1407.0, + 1473.0, + 1407.0, + 1509.0, + 294.0, + 1509.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1503.0, + 1405.0, + 1503.0, + 1405.0, + 1538.0, + 294.0, + 1538.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1534.0, + 1408.0, + 1534.0, + 1408.0, + 1571.0, + 295.0, + 1571.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1564.0, + 1408.0, + 1564.0, + 1408.0, + 1601.0, + 294.0, + 1601.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1592.0, + 432.0, + 1592.0, + 432.0, + 1633.0, + 294.0, + 1633.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1166.0, + 1406.0, + 1166.0, + 1406.0, + 1200.0, + 295.0, + 1200.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1196.0, + 1407.0, + 1196.0, + 1407.0, + 1230.0, + 295.0, + 1230.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1227.0, + 860.0, + 1227.0, + 860.0, + 1261.0, + 297.0, + 1261.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1031.0, + 1227.0, + 1087.0, + 1227.0, + 1087.0, + 1261.0, + 1031.0, + 1261.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1162.0, + 1227.0, + 1342.0, + 1227.0, + 1342.0, + 1261.0, + 1162.0, + 1261.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1272.0, + 1403.0, + 1272.0, + 1403.0, + 1309.0, + 293.0, + 1309.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1303.0, + 668.0, + 1303.0, + 668.0, + 1339.0, + 295.0, + 1339.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 14, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 3, + "poly": [ + 298, + 772, + 1470, + 772, + 1470, + 1412, + 298, + 1412 + ], + "score": 0.963 + }, + { + "category_id": 4, + "poly": [ + 391, + 1448, + 1308, + 1448, + 1308, + 1483, + 391, + 1483 + ], + "score": 0.924 + }, + { + "category_id": 2, + "poly": [ + 297, + 75, + 858, + 75, + 858, + 106, + 297, + 106 + ], + "score": 0.9 + }, + { + "category_id": 2, + "poly": [ + 836, + 2088, + 864, + 2088, + 864, + 2113, + 836, + 2113 + ], + "score": 0.843 + }, + { + "category_id": 13, + "poly": [ + 1432, + 1257, + 1460, + 1257, + 1460, + 1273, + 1432, + 1273 + ], + "score": 0.48, + "latex": "\\top = \\uparrow" + }, + { + "category_id": 15, + "poly": [ + 443.0, + 776.0, + 554.0, + 776.0, + 554.0, + 802.0, + 443.0, + 802.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 801.0, + 775.0, + 867.0, + 775.0, + 867.0, + 804.0, + 801.0, + 804.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1128.0, + 775.0, + 1212.0, + 775.0, + 1212.0, + 808.0, + 1128.0, + 808.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 789.0, + 355.0, + 789.0, + 355.0, + 812.0, + 306.0, + 812.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 649.0, + 789.0, + 692.0, + 789.0, + 692.0, + 813.0, + 649.0, + 813.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 979.0, + 789.0, + 1027.0, + 789.0, + 1027.0, + 812.0, + 979.0, + 812.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 305.0, + 828.0, + 358.0, + 828.0, + 358.0, + 855.0, + 305.0, + 855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 647.0, + 827.0, + 695.0, + 827.0, + 695.0, + 855.0, + 647.0, + 855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 820.0, + 1027.0, + 820.0, + 1027.0, + 843.0, + 978.0, + 843.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 849.0, + 1030.0, + 849.0, + 1030.0, + 876.0, + 978.0, + 876.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 303.0, + 872.0, + 356.0, + 872.0, + 356.0, + 893.0, + 303.0, + 893.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 648.0, + 870.0, + 693.0, + 870.0, + 693.0, + 895.0, + 648.0, + 895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 881.0, + 1028.0, + 881.0, + 1028.0, + 904.0, + 978.0, + 904.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 305.0, + 910.0, + 358.0, + 910.0, + 358.0, + 937.0, + 305.0, + 937.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 649.0, + 911.0, + 693.0, + 911.0, + 693.0, + 936.0, + 649.0, + 936.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 911.0, + 1028.0, + 911.0, + 1028.0, + 936.0, + 985.0, + 936.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 942.0, + 1028.0, + 942.0, + 1028.0, + 965.0, + 985.0, + 965.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 953.0, + 357.0, + 953.0, + 357.0, + 976.0, + 306.0, + 976.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 578.0, + 965.0, + 602.0, + 965.0, + 602.0, + 980.0, + 578.0, + 980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 649.0, + 952.0, + 692.0, + 952.0, + 692.0, + 976.0, + 649.0, + 976.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 770.0, + 985.0, + 781.0, + 985.0, + 781.0, + 995.0, + 770.0, + 995.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 881.0, + 987.0, + 892.0, + 987.0, + 892.0, + 997.0, + 881.0, + 997.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 972.0, + 1028.0, + 972.0, + 1028.0, + 995.0, + 985.0, + 995.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 305.0, + 991.0, + 358.0, + 991.0, + 358.0, + 1019.0, + 305.0, + 1019.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 656.0, + 993.0, + 693.0, + 993.0, + 693.0, + 1017.0, + 656.0, + 1017.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 762.0, + 998.0, + 769.0, + 998.0, + 769.0, + 1005.0, + 762.0, + 1005.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 494.0, + 1010.0, + 522.0, + 1010.0, + 522.0, + 1024.0, + 494.0, + 1024.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 583.0, + 1005.0, + 634.0, + 1005.0, + 634.0, + 1019.0, + 583.0, + 1019.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 745.0, + 1006.0, + 763.0, + 1006.0, + 763.0, + 1023.0, + 745.0, + 1023.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1002.0, + 1028.0, + 1002.0, + 1028.0, + 1027.0, + 985.0, + 1027.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 737.0, + 1019.0, + 754.0, + 1019.0, + 754.0, + 1036.0, + 737.0, + 1036.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 338.0, + 1036.0, + 357.0, + 1036.0, + 357.0, + 1054.0, + 338.0, + 1054.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 675.0, + 1036.0, + 692.0, + 1036.0, + 692.0, + 1055.0, + 675.0, + 1055.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 730.0, + 1035.0, + 744.0, + 1035.0, + 744.0, + 1047.0, + 730.0, + 1047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1010.0, + 1036.0, + 1028.0, + 1036.0, + 1028.0, + 1054.0, + 1010.0, + 1054.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 345.0, + 1051.0, + 377.0, + 1051.0, + 377.0, + 1072.0, + 345.0, + 1072.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 378.0, + 1051.0, + 411.0, + 1051.0, + 411.0, + 1072.0, + 378.0, + 1072.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 414.0, + 1051.0, + 445.0, + 1051.0, + 445.0, + 1072.0, + 414.0, + 1072.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 447.0, + 1050.0, + 549.0, + 1050.0, + 549.0, + 1072.0, + 447.0, + 1072.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 551.0, + 1050.0, + 652.0, + 1050.0, + 652.0, + 1073.0, + 551.0, + 1073.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 680.0, + 1050.0, + 746.0, + 1050.0, + 746.0, + 1073.0, + 680.0, + 1073.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 749.0, + 1051.0, + 781.0, + 1051.0, + 781.0, + 1072.0, + 749.0, + 1072.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 783.0, + 1050.0, + 956.0, + 1050.0, + 956.0, + 1073.0, + 783.0, + 1073.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1017.0, + 1048.0, + 1255.0, + 1048.0, + 1255.0, + 1074.0, + 1017.0, + 1074.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1256.0, + 1050.0, + 1323.0, + 1050.0, + 1323.0, + 1073.0, + 1256.0, + 1073.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 467.0, + 1070.0, + 530.0, + 1070.0, + 530.0, + 1093.0, + 467.0, + 1093.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 607.0, + 1066.0, + 643.0, + 1066.0, + 643.0, + 1091.0, + 607.0, + 1091.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 803.0, + 1070.0, + 865.0, + 1070.0, + 865.0, + 1093.0, + 803.0, + 1093.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 944.0, + 1066.0, + 978.0, + 1066.0, + 978.0, + 1091.0, + 944.0, + 1091.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1138.0, + 1070.0, + 1202.0, + 1070.0, + 1202.0, + 1093.0, + 1138.0, + 1093.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1277.0, + 1068.0, + 1313.0, + 1068.0, + 1313.0, + 1088.0, + 1277.0, + 1088.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 460.0, + 1092.0, + 536.0, + 1092.0, + 536.0, + 1115.0, + 460.0, + 1115.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 793.0, + 1089.0, + 876.0, + 1089.0, + 876.0, + 1120.0, + 793.0, + 1120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1128.0, + 1091.0, + 1213.0, + 1091.0, + 1213.0, + 1118.0, + 1128.0, + 1118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1104.0, + 355.0, + 1104.0, + 355.0, + 1127.0, + 306.0, + 1127.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 643.0, + 1104.0, + 692.0, + 1104.0, + 692.0, + 1127.0, + 643.0, + 1127.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 1104.0, + 1027.0, + 1104.0, + 1027.0, + 1127.0, + 978.0, + 1127.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1134.0, + 357.0, + 1134.0, + 357.0, + 1157.0, + 306.0, + 1157.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 642.0, + 1145.0, + 692.0, + 1145.0, + 692.0, + 1168.0, + 642.0, + 1168.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 977.0, + 1153.0, + 1028.0, + 1153.0, + 1028.0, + 1176.0, + 977.0, + 1176.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1165.0, + 355.0, + 1165.0, + 355.0, + 1189.0, + 306.0, + 1189.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 306.0, + 1195.0, + 355.0, + 1195.0, + 355.0, + 1219.0, + 306.0, + 1219.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 537.0, + 1202.0, + 581.0, + 1202.0, + 581.0, + 1225.0, + 537.0, + 1225.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 643.0, + 1186.0, + 692.0, + 1186.0, + 692.0, + 1209.0, + 643.0, + 1209.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 1202.0, + 1028.0, + 1202.0, + 1028.0, + 1225.0, + 978.0, + 1225.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1353.0, + 1180.0, + 1471.0, + 1180.0, + 1471.0, + 1223.0, + 1353.0, + 1223.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 313.0, + 1227.0, + 355.0, + 1227.0, + 355.0, + 1250.0, + 313.0, + 1250.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 642.0, + 1224.0, + 693.0, + 1224.0, + 693.0, + 1251.0, + 642.0, + 1251.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1349.0, + 1216.0, + 1467.0, + 1216.0, + 1467.0, + 1245.0, + 1349.0, + 1245.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 313.0, + 1257.0, + 357.0, + 1257.0, + 357.0, + 1281.0, + 313.0, + 1281.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 978.0, + 1251.0, + 1028.0, + 1251.0, + 1028.0, + 1274.0, + 978.0, + 1274.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1350.0, + 1239.0, + 1431.0, + 1239.0, + 1431.0, + 1277.0, + 1350.0, + 1277.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1461.0, + 1239.0, + 1464.0, + 1239.0, + 1464.0, + 1277.0, + 1461.0, + 1277.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 313.0, + 1288.0, + 357.0, + 1288.0, + 357.0, + 1311.0, + 313.0, + 1311.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 643.0, + 1267.0, + 692.0, + 1267.0, + 692.0, + 1291.0, + 643.0, + 1291.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 936.0, + 1280.0, + 973.0, + 1280.0, + 973.0, + 1311.0, + 936.0, + 1311.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1299.0, + 1028.0, + 1299.0, + 1028.0, + 1323.0, + 985.0, + 1323.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1141.0, + 1282.0, + 1157.0, + 1282.0, + 1157.0, + 1303.0, + 1141.0, + 1303.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1210.0, + 1277.0, + 1260.0, + 1277.0, + 1260.0, + 1307.0, + 1210.0, + 1307.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1279.0, + 1276.0, + 1301.0, + 1276.0, + 1301.0, + 1292.0, + 1279.0, + 1292.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1350.0, + 1273.0, + 1435.0, + 1273.0, + 1435.0, + 1299.0, + 1350.0, + 1299.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 313.0, + 1318.0, + 355.0, + 1318.0, + 355.0, + 1342.0, + 313.0, + 1342.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 370.0, + 1329.0, + 377.0, + 1329.0, + 377.0, + 1335.0, + 370.0, + 1335.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 649.0, + 1308.0, + 693.0, + 1308.0, + 693.0, + 1331.0, + 649.0, + 1331.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 816.0, + 1318.0, + 830.0, + 1318.0, + 830.0, + 1330.0, + 816.0, + 1330.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 844.0, + 1315.0, + 855.0, + 1315.0, + 855.0, + 1325.0, + 844.0, + 1325.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1080.0, + 1319.0, + 1097.0, + 1319.0, + 1097.0, + 1334.0, + 1080.0, + 1334.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 818.0, + 1340.0, + 828.0, + 1340.0, + 828.0, + 1352.0, + 818.0, + 1352.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 338.0, + 1350.0, + 355.0, + 1350.0, + 355.0, + 1370.0, + 338.0, + 1370.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 672.0, + 1349.0, + 693.0, + 1349.0, + 693.0, + 1371.0, + 672.0, + 1371.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1011.0, + 1352.0, + 1026.0, + 1352.0, + 1026.0, + 1370.0, + 1011.0, + 1370.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 343.0, + 1365.0, + 652.0, + 1365.0, + 652.0, + 1389.0, + 343.0, + 1389.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 680.0, + 1365.0, + 918.0, + 1365.0, + 918.0, + 1387.0, + 680.0, + 1387.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 920.0, + 1365.0, + 987.0, + 1365.0, + 987.0, + 1389.0, + 920.0, + 1389.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1018.0, + 1367.0, + 1083.0, + 1367.0, + 1083.0, + 1386.0, + 1018.0, + 1386.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1084.0, + 1365.0, + 1324.0, + 1365.0, + 1324.0, + 1389.0, + 1084.0, + 1389.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 466.0, + 1383.0, + 531.0, + 1383.0, + 531.0, + 1409.0, + 466.0, + 1409.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 608.0, + 1383.0, + 642.0, + 1383.0, + 642.0, + 1404.0, + 608.0, + 1404.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 944.0, + 1382.0, + 978.0, + 1382.0, + 978.0, + 1406.0, + 944.0, + 1406.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1138.0, + 1383.0, + 1201.0, + 1383.0, + 1201.0, + 1409.0, + 1138.0, + 1409.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1280.0, + 1383.0, + 1313.0, + 1383.0, + 1313.0, + 1404.0, + 1280.0, + 1404.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 768.75, + 900.0, + 784.75, + 900.0, + 784.75, + 911.0, + 768.75, + 911.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 933.0, + 982.0, + 960.0, + 982.0, + 960.0, + 1003.0, + 933.0, + 1003.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 428.0, + 1003.0, + 492.0, + 1003.0, + 492.0, + 1026.0, + 428.0, + 1026.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1091.0, + 1030.5, + 1139.0, + 1030.5, + 1139.0, + 1049.5, + 1091.0, + 1049.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 946.0, + 1050.0, + 985.0, + 1050.0, + 985.0, + 1071.5, + 946.0, + 1071.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 374.25, + 1224.5, + 575.25, + 1224.5, + 575.25, + 1283.5, + 374.25, + 1283.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 893.0, + 1280.5, + 944.0, + 1280.5, + 944.0, + 1314.5, + 893.0, + 1314.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1169.0, + 1291.0, + 1206.0, + 1291.0, + 1206.0, + 1309.0, + 1169.0, + 1309.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 377.0, + 1310.0, + 381.0, + 1310.0, + 381.0, + 1328.5, + 377.0, + 1328.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1083.25, + 1301.0, + 1147.25, + 1301.0, + 1147.25, + 1325.5, + 1083.25, + 1325.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1051.25, + 1311.5, + 1091.25, + 1311.5, + 1091.25, + 1326.5, + 1051.25, + 1326.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1025.0, + 1346.0, + 1055.0, + 1346.0, + 1055.0, + 1361.0, + 1025.0, + 1361.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 802.0, + 1385.5, + 864.0, + 1385.5, + 864.0, + 1408.5, + 802.0, + 1408.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 386.0, + 1443.0, + 1310.0, + 1443.0, + 1310.0, + 1491.0, + 386.0, + 1491.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 72.0, + 858.0, + 72.0, + 858.0, + 108.0, + 297.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 868.0, + 2085.0, + 868.0, + 2124.0, + 831.0, + 2124.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 15, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 3, + "poly": [ + 345, + 278, + 1331, + 278, + 1331, + 1747, + 345, + 1747 + ], + "score": 0.96 + }, + { + "category_id": 4, + "poly": [ + 297, + 1751, + 1401, + 1751, + 1401, + 1815, + 297, + 1815 + ], + "score": 0.934 + }, + { + "category_id": 2, + "poly": [ + 297, + 75, + 857, + 75, + 857, + 105, + 297, + 105 + ], + "score": 0.909 + }, + { + "category_id": 2, + "poly": [ + 835, + 2087, + 863, + 2087, + 863, + 2112, + 835, + 2112 + ], + "score": 0.857 + }, + { + "category_id": 4, + "poly": [ + 316, + 232, + 1145, + 232, + 1145, + 261, + 316, + 261 + ], + "score": 0.434 + }, + { + "category_id": 4, + "poly": [ + 297, + 231, + 1153, + 231, + 1153, + 262, + 297, + 262 + ], + "score": 0.348 + }, + { + "category_id": 13, + "poly": [ + 432, + 235, + 451, + 235, + 451, + 258, + 432, + 258 + ], + "score": 0.35, + "latex": "^ +" + }, + { + "category_id": 15, + "poly": [ + 503.0, + 303.0, + 557.0, + 303.0, + 557.0, + 333.0, + 503.0, + 333.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 815.0, + 303.0, + 885.0, + 303.0, + 885.0, + 331.0, + 815.0, + 331.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1120.0, + 301.0, + 1220.0, + 301.0, + 1220.0, + 333.0, + 1120.0, + 333.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 315.0, + 398.0, + 315.0, + 398.0, + 344.0, + 350.0, + 344.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 315.0, + 720.0, + 315.0, + 720.0, + 344.0, + 678.0, + 344.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 318.0, + 1037.0, + 318.0, + 1037.0, + 341.0, + 985.0, + 341.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 360.0, + 398.0, + 360.0, + 398.0, + 390.0, + 350.0, + 390.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 354.0, + 720.0, + 354.0, + 720.0, + 383.0, + 678.0, + 383.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 364.0, + 1037.0, + 364.0, + 1037.0, + 387.0, + 985.0, + 387.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 393.0, + 718.0, + 393.0, + 718.0, + 422.0, + 678.0, + 422.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 408.0, + 398.0, + 408.0, + 398.0, + 437.0, + 350.0, + 437.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 983.0, + 408.0, + 1038.0, + 408.0, + 1038.0, + 436.0, + 983.0, + 436.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 680.0, + 436.0, + 718.0, + 436.0, + 718.0, + 460.0, + 680.0, + 460.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 352.0, + 454.0, + 398.0, + 454.0, + 398.0, + 483.0, + 352.0, + 483.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 987.0, + 457.0, + 1038.0, + 457.0, + 1038.0, + 480.0, + 987.0, + 480.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 473.0, + 720.0, + 473.0, + 720.0, + 503.0, + 678.0, + 503.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 352.0, + 499.0, + 400.0, + 499.0, + 400.0, + 529.0, + 352.0, + 529.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 512.0, + 720.0, + 512.0, + 720.0, + 542.0, + 678.0, + 542.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 987.0, + 503.0, + 1038.0, + 503.0, + 1038.0, + 525.0, + 987.0, + 525.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1095.0, + 506.0, + 1115.0, + 506.0, + 1115.0, + 545.0, + 1095.0, + 545.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 545.0, + 400.0, + 545.0, + 400.0, + 574.0, + 358.0, + 574.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 552.0, + 720.0, + 552.0, + 720.0, + 581.0, + 678.0, + 581.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 992.0, + 547.0, + 1040.0, + 547.0, + 1040.0, + 574.0, + 992.0, + 574.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1079.0, + 560.0, + 1101.0, + 560.0, + 1101.0, + 579.0, + 1079.0, + 579.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 718.0, + 586.0, + 739.0, + 586.0, + 739.0, + 606.0, + 718.0, + 606.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 377.0, + 596.0, + 398.0, + 596.0, + 398.0, + 617.0, + 377.0, + 617.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 697.0, + 596.0, + 718.0, + 596.0, + 718.0, + 617.0, + 697.0, + 617.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1017.0, + 596.0, + 1037.0, + 596.0, + 1037.0, + 617.0, + 1017.0, + 617.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 610.0, + 675.0, + 610.0, + 675.0, + 637.0, + 382.0, + 637.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 702.0, + 610.0, + 996.0, + 610.0, + 996.0, + 637.0, + 702.0, + 637.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1024.0, + 610.0, + 1316.0, + 610.0, + 1316.0, + 637.0, + 1024.0, + 637.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 495.0, + 628.0, + 565.0, + 628.0, + 565.0, + 660.0, + 495.0, + 660.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 626.0, + 630.0, + 665.0, + 630.0, + 665.0, + 655.0, + 626.0, + 655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 815.0, + 630.0, + 885.0, + 630.0, + 885.0, + 660.0, + 815.0, + 660.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 948.0, + 630.0, + 985.0, + 630.0, + 985.0, + 655.0, + 948.0, + 655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1135.0, + 632.0, + 1206.0, + 632.0, + 1206.0, + 658.0, + 1135.0, + 658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1266.0, + 630.0, + 1303.0, + 630.0, + 1303.0, + 655.0, + 1266.0, + 655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 449.0, + 653.0, + 612.0, + 653.0, + 612.0, + 684.0, + 449.0, + 684.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 808.0, + 649.0, + 892.0, + 649.0, + 892.0, + 686.0, + 808.0, + 686.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1124.0, + 655.0, + 1216.0, + 655.0, + 1216.0, + 682.0, + 1124.0, + 682.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 666.0, + 400.0, + 666.0, + 400.0, + 695.0, + 350.0, + 695.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 686.0, + 668.0, + 720.0, + 668.0, + 720.0, + 694.0, + 686.0, + 694.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1009.0, + 668.0, + 1038.0, + 668.0, + 1038.0, + 694.0, + 1009.0, + 694.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 697.0, + 400.0, + 697.0, + 400.0, + 726.0, + 350.0, + 726.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 688.0, + 707.0, + 718.0, + 707.0, + 718.0, + 733.0, + 688.0, + 733.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1014.0, + 712.0, + 1037.0, + 712.0, + 1037.0, + 730.0, + 1014.0, + 730.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 728.0, + 398.0, + 728.0, + 398.0, + 758.0, + 350.0, + 758.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 686.0, + 746.0, + 718.0, + 746.0, + 718.0, + 772.0, + 686.0, + 772.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1011.0, + 748.0, + 1038.0, + 748.0, + 1038.0, + 772.0, + 1011.0, + 772.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 758.0, + 398.0, + 758.0, + 398.0, + 787.0, + 350.0, + 787.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 789.0, + 398.0, + 789.0, + 398.0, + 818.0, + 350.0, + 818.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 688.0, + 787.0, + 718.0, + 787.0, + 718.0, + 813.0, + 688.0, + 813.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1011.0, + 787.0, + 1038.0, + 787.0, + 1038.0, + 811.0, + 1011.0, + 811.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 820.0, + 398.0, + 820.0, + 398.0, + 849.0, + 350.0, + 849.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 689.0, + 826.0, + 718.0, + 826.0, + 718.0, + 854.0, + 689.0, + 854.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1001.0, + 825.0, + 1040.0, + 825.0, + 1040.0, + 854.0, + 1001.0, + 854.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 352.0, + 851.0, + 400.0, + 851.0, + 400.0, + 880.0, + 352.0, + 880.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 689.0, + 865.0, + 718.0, + 865.0, + 718.0, + 892.0, + 689.0, + 892.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1001.0, + 864.0, + 1040.0, + 864.0, + 1040.0, + 893.0, + 1001.0, + 893.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 352.0, + 882.0, + 398.0, + 882.0, + 398.0, + 911.0, + 352.0, + 911.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 913.0, + 413.0, + 913.0, + 413.0, + 942.0, + 358.0, + 942.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 697.0, + 908.0, + 718.0, + 908.0, + 718.0, + 931.0, + 697.0, + 931.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1000.0, + 905.0, + 1040.0, + 905.0, + 1040.0, + 934.0, + 1000.0, + 934.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 377.0, + 947.0, + 397.0, + 947.0, + 397.0, + 968.0, + 377.0, + 968.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 699.0, + 947.0, + 717.0, + 947.0, + 717.0, + 968.0, + 699.0, + 968.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1003.0, + 945.0, + 1038.0, + 945.0, + 1038.0, + 970.0, + 1003.0, + 970.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 962.0, + 675.0, + 962.0, + 675.0, + 988.0, + 382.0, + 988.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 702.0, + 962.0, + 996.0, + 962.0, + 996.0, + 988.0, + 702.0, + 988.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1024.0, + 962.0, + 1316.0, + 962.0, + 1316.0, + 988.0, + 1024.0, + 988.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 495.0, + 981.0, + 565.0, + 981.0, + 565.0, + 1013.0, + 495.0, + 1013.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 626.0, + 981.0, + 667.0, + 981.0, + 667.0, + 1006.0, + 626.0, + 1006.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 815.0, + 981.0, + 885.0, + 981.0, + 885.0, + 1013.0, + 815.0, + 1013.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 948.0, + 981.0, + 985.0, + 981.0, + 985.0, + 1006.0, + 948.0, + 1006.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1135.0, + 981.0, + 1205.0, + 981.0, + 1205.0, + 1013.0, + 1135.0, + 1013.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1266.0, + 981.0, + 1305.0, + 981.0, + 1305.0, + 1006.0, + 1266.0, + 1006.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 1017.0, + 400.0, + 1017.0, + 400.0, + 1047.0, + 358.0, + 1047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 479.0, + 1008.0, + 581.0, + 1008.0, + 581.0, + 1034.0, + 479.0, + 1034.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 662.0, + 1019.0, + 718.0, + 1019.0, + 718.0, + 1047.0, + 662.0, + 1047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 784.0, + 1002.0, + 916.0, + 1002.0, + 916.0, + 1039.0, + 784.0, + 1039.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1021.0, + 1037.0, + 1021.0, + 1037.0, + 1044.0, + 985.0, + 1044.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1125.0, + 1002.0, + 1217.0, + 1002.0, + 1217.0, + 1039.0, + 1125.0, + 1039.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 1057.0, + 400.0, + 1057.0, + 400.0, + 1086.0, + 358.0, + 1086.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 663.0, + 1065.0, + 718.0, + 1065.0, + 718.0, + 1093.0, + 663.0, + 1093.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 983.0, + 1063.0, + 1040.0, + 1063.0, + 1040.0, + 1091.0, + 983.0, + 1091.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 360.0, + 1099.0, + 397.0, + 1099.0, + 397.0, + 1124.0, + 360.0, + 1124.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 663.0, + 1111.0, + 718.0, + 1111.0, + 718.0, + 1138.0, + 663.0, + 1138.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 982.0, + 1111.0, + 1040.0, + 1111.0, + 1040.0, + 1138.0, + 982.0, + 1138.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 1137.0, + 400.0, + 1137.0, + 400.0, + 1166.0, + 358.0, + 1166.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 663.0, + 1156.0, + 720.0, + 1156.0, + 720.0, + 1184.0, + 663.0, + 1184.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 983.0, + 1156.0, + 1040.0, + 1156.0, + 1040.0, + 1184.0, + 983.0, + 1184.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 1176.0, + 398.0, + 1176.0, + 398.0, + 1205.0, + 358.0, + 1205.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 481.0, + 1201.0, + 497.0, + 1201.0, + 497.0, + 1233.0, + 481.0, + 1233.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 663.0, + 1202.0, + 718.0, + 1202.0, + 718.0, + 1230.0, + 663.0, + 1230.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 983.0, + 1198.0, + 1041.0, + 1198.0, + 1041.0, + 1232.0, + 983.0, + 1232.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 1215.0, + 400.0, + 1215.0, + 400.0, + 1245.0, + 358.0, + 1245.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 361.0, + 1258.0, + 397.0, + 1258.0, + 397.0, + 1282.0, + 361.0, + 1282.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 670.0, + 1249.0, + 720.0, + 1249.0, + 720.0, + 1277.0, + 670.0, + 1277.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 992.0, + 1249.0, + 1040.0, + 1249.0, + 1040.0, + 1277.0, + 992.0, + 1277.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 381.0, + 1302.0, + 395.0, + 1302.0, + 395.0, + 1316.0, + 381.0, + 1316.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 697.0, + 1300.0, + 718.0, + 1300.0, + 718.0, + 1320.0, + 697.0, + 1320.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1021.0, + 1302.0, + 1035.0, + 1302.0, + 1035.0, + 1318.0, + 1021.0, + 1318.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 1313.0, + 676.0, + 1313.0, + 676.0, + 1339.0, + 382.0, + 1339.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 702.0, + 1313.0, + 996.0, + 1313.0, + 996.0, + 1339.0, + 702.0, + 1339.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1024.0, + 1313.0, + 1316.0, + 1313.0, + 1316.0, + 1339.0, + 1024.0, + 1339.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 495.0, + 1332.0, + 565.0, + 1332.0, + 565.0, + 1364.0, + 495.0, + 1364.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 625.0, + 1330.0, + 667.0, + 1330.0, + 667.0, + 1359.0, + 625.0, + 1359.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 815.0, + 1332.0, + 885.0, + 1332.0, + 885.0, + 1364.0, + 815.0, + 1364.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 948.0, + 1333.0, + 985.0, + 1333.0, + 985.0, + 1357.0, + 948.0, + 1357.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1134.0, + 1334.0, + 1206.0, + 1334.0, + 1206.0, + 1362.0, + 1134.0, + 1362.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1268.0, + 1333.0, + 1305.0, + 1333.0, + 1305.0, + 1356.0, + 1268.0, + 1356.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 463.0, + 1356.0, + 596.0, + 1356.0, + 596.0, + 1387.0, + 463.0, + 1387.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 809.0, + 1357.0, + 895.0, + 1357.0, + 895.0, + 1385.0, + 809.0, + 1385.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1126.0, + 1357.0, + 1216.0, + 1357.0, + 1216.0, + 1385.0, + 1126.0, + 1385.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 1369.0, + 398.0, + 1369.0, + 398.0, + 1398.0, + 350.0, + 1398.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 670.0, + 1369.0, + 718.0, + 1369.0, + 718.0, + 1398.0, + 670.0, + 1398.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 983.0, + 1370.0, + 1040.0, + 1370.0, + 1040.0, + 1398.0, + 983.0, + 1398.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 670.0, + 1410.0, + 718.0, + 1410.0, + 718.0, + 1437.0, + 670.0, + 1437.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 1424.0, + 400.0, + 1424.0, + 400.0, + 1454.0, + 350.0, + 1454.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1428.0, + 1038.0, + 1428.0, + 1038.0, + 1451.0, + 985.0, + 1451.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 670.0, + 1449.0, + 718.0, + 1449.0, + 718.0, + 1477.0, + 670.0, + 1477.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 352.0, + 1480.0, + 398.0, + 1480.0, + 398.0, + 1509.0, + 352.0, + 1509.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 673.0, + 1491.0, + 717.0, + 1491.0, + 717.0, + 1514.0, + 673.0, + 1514.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1480.0, + 1040.0, + 1480.0, + 1040.0, + 1508.0, + 985.0, + 1508.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 352.0, + 1535.0, + 398.0, + 1535.0, + 398.0, + 1565.0, + 352.0, + 1565.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 670.0, + 1529.0, + 718.0, + 1529.0, + 718.0, + 1557.0, + 670.0, + 1557.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 985.0, + 1535.0, + 1040.0, + 1535.0, + 1040.0, + 1563.0, + 985.0, + 1563.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 726.0, + 1557.0, + 746.0, + 1557.0, + 746.0, + 1583.0, + 726.0, + 1583.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 670.0, + 1568.0, + 718.0, + 1568.0, + 718.0, + 1596.0, + 670.0, + 1596.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 1591.0, + 398.0, + 1591.0, + 398.0, + 1620.0, + 358.0, + 1620.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 992.0, + 1591.0, + 1040.0, + 1591.0, + 1040.0, + 1619.0, + 992.0, + 1619.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 428.0, + 1611.0, + 442.0, + 1611.0, + 442.0, + 1627.0, + 428.0, + 1627.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 672.0, + 1607.0, + 720.0, + 1607.0, + 720.0, + 1635.0, + 672.0, + 1635.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 381.0, + 1653.0, + 395.0, + 1653.0, + 395.0, + 1670.0, + 381.0, + 1670.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 697.0, + 1652.0, + 720.0, + 1652.0, + 720.0, + 1671.0, + 697.0, + 1671.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1021.0, + 1653.0, + 1035.0, + 1653.0, + 1035.0, + 1668.0, + 1021.0, + 1668.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 382.0, + 1665.0, + 675.0, + 1665.0, + 675.0, + 1691.0, + 382.0, + 1691.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 701.0, + 1665.0, + 995.0, + 1665.0, + 995.0, + 1691.0, + 701.0, + 1691.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1024.0, + 1665.0, + 1316.0, + 1665.0, + 1316.0, + 1691.0, + 1024.0, + 1691.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 495.0, + 1684.0, + 565.0, + 1684.0, + 565.0, + 1716.0, + 495.0, + 1716.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 626.0, + 1684.0, + 665.0, + 1684.0, + 665.0, + 1709.0, + 626.0, + 1709.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 815.0, + 1686.0, + 885.0, + 1686.0, + 885.0, + 1714.0, + 815.0, + 1714.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 948.0, + 1684.0, + 985.0, + 1684.0, + 985.0, + 1709.0, + 948.0, + 1709.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1135.0, + 1686.0, + 1205.0, + 1686.0, + 1205.0, + 1714.0, + 1135.0, + 1714.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1268.0, + 1684.0, + 1305.0, + 1684.0, + 1305.0, + 1709.0, + 1268.0, + 1709.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1108.0, + 363.0, + 1316.0, + 363.0, + 1316.0, + 418.0, + 1108.0, + 418.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 388.0, + 433.5, + 701.0, + 433.5, + 701.0, + 513.5, + 388.0, + 513.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 747.75, + 422.0, + 1016.75, + 422.0, + 1016.75, + 488.0, + 747.75, + 488.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 749.0, + 542.5, + 800.0, + 542.5, + 800.0, + 563.5, + 749.0, + 563.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1048.0, + 564.5, + 1066.0, + 564.5, + 1066.0, + 576.5, + 1048.0, + 576.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 769.0, + 714.0, + 835.0, + 714.0, + 835.0, + 738.0, + 769.0, + 738.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 378.0, + 765.0, + 686.0, + 765.0, + 686.0, + 844.0, + 378.0, + 844.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1139.0, + 1109.0, + 1328.0, + 1109.0, + 1328.0, + 1166.5, + 1139.0, + 1166.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1164.25, + 1412.5, + 1300.25, + 1412.5, + 1300.25, + 1471.0, + 1164.25, + 1471.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 426.0, + 1475.0, + 715.0, + 1475.0, + 715.0, + 1552.0, + 426.0, + 1552.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 706.0, + 1472.5, + 1012.0, + 1472.5, + 1012.0, + 1536.5, + 706.0, + 1536.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1053.0, + 1512.5, + 1200.0, + 1512.5, + 1200.0, + 1545.5, + 1053.0, + 1545.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1748.0, + 1405.0, + 1748.0, + 1405.0, + 1788.0, + 293.0, + 1788.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1781.0, + 437.0, + 1781.0, + 437.0, + 1819.0, + 292.0, + 1819.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 73.0, + 859.0, + 73.0, + 859.0, + 108.0, + 297.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 867.0, + 2085.0, + 867.0, + 2124.0, + 831.0, + 2124.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 308.0, + 228.0, + 431.0, + 228.0, + 431.0, + 266.0, + 308.0, + 266.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 452.0, + 228.0, + 1152.0, + 228.0, + 1152.0, + 266.0, + 452.0, + 266.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 228.0, + 431.0, + 228.0, + 431.0, + 266.0, + 295.0, + 266.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 452.0, + 228.0, + 1156.0, + 228.0, + 1156.0, + 266.0, + 452.0, + 266.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 16, + "width": 1700, + "height": 2200 + } + } +] \ No newline at end of file diff --git a/parse/train/o_V-MjyyGV_/o_V-MjyyGV__content_list.json b/parse/train/o_V-MjyyGV_/o_V-MjyyGV__content_list.json new file mode 100644 index 0000000000000000000000000000000000000000..c0a346679c6aa579f3cbe899ab07a4114779eb74 --- /dev/null +++ b/parse/train/o_V-MjyyGV_/o_V-MjyyGV__content_list.json @@ -0,0 +1,1973 @@ +[ + { + "type": "text", + "text": "SELF-SUPERVISED POLICY ADAPTATION DURING DEPLOYMENT ", + "text_level": 1, + "bbox": [ + 174, + 98, + 663, + 146 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Nicklas Hansen12, Rishabh Jangir13, Yu Sun4, Guillem Alenya\\`3, Pieter Abbeel4, Alexei A Efros4, Lerrel Pinto5, Xiaolong Wang1 1UC San Diego 2Technical University of Denmark ${ } ^ { 3 } \\mathrm { I R I }$ , CSIC-UPC 4UC Berkeley 5NYU ", + "bbox": [ + 184, + 169, + 630, + 228 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "ABSTRACT ", + "text_level": 1, + "bbox": [ + 454, + 265, + 544, + 280 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "In most real world scenarios, a policy trained by reinforcement learning in one environment needs to be deployed in another, potentially quite different environment. However, generalization across different environments is known to be hard. A natural solution would be to keep training after deployment in the new environment, but this cannot be done if the new environment offers no reward signal. Our work explores the use of self-supervision to allow the policy to continue training after deployment without using any rewards. While previous methods explicitly anticipate changes in the new environment, we assume no prior knowledge of those changes yet still obtain significant improvements. Empirical evaluations are performed on diverse simulation environments from DeepMind Control suite and ViZDoom, as well as real robotic manipulation tasks in continuously changing environments, taking observations from an uncalibrated camera. Our method improves generalization in 31 out of 36 environments across various tasks and outperforms domain randomization on a majority of environments. ", + "bbox": [ + 233, + 291, + 766, + 484 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "1 INTRODUCTION ", + "text_level": 1, + "bbox": [ + 176, + 512, + 336, + 527 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Deep reinforcement learning (RL) has achieved considerable success when combined with convolutional neural networks for deriving actions from image pixels (Mnih et al., 2013; Levine et al., 2016; Nair et al., 2018; Yan et al., 2020; Andrychowicz et al., 2020). However, one significant challenge for real-world deployment of vision-based RL remains: a policy trained in one environment might not generalize to other new environments not seen during training. Already hard for RL alone, the challenge is exacerbated when a policy faces high-dimensional visual inputs. ", + "bbox": [ + 174, + 537, + 825, + 622 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "A well explored class of solutions is to learn robust policies that are simply invariant to changes in the environment (Rajeswaran et al., 2016; Tobin et al., 2017; Sadeghi & Levine, 2016; Pinto et al., 2017b; Lee et al., 2019). For example, domain randomization (Tobin et al., 2017; Peng et al., 2018; Pinto et al., 2017a; Yang et al., 2019) applies data augmentation in a simulated environment to train a single robust policy, with the hope that the augmented environment covers enough factors of variation in the test environment. However, this hope may be difficult to realize when the test environment is truly unknown. With too much randomization, training a policy that can simultaneously fit numerous augmented environments requires much larger model and sample complexity. With too little randomization, the actual changes in the test environment might not be covered, and domain randomization may do more harm than good since the randomized factors are now irrelevant. Both phenomena have been observed in our experiments. In all cases, this class of solutions requires human experts to anticipate the changes before the test environment is seen. This cannot scale as more test environments are added with more diverse changes. ", + "bbox": [ + 174, + 630, + 825, + 809 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "Instead of learning a robust policy invariant to all possible environmental changes, we argue that it is better for a policy to keep learning during deployment and adapt to its actual new environment. A naive way to implement this in RL is to fine-tune the policy in the new environment using rewards as supervision (Rusu et al., 2016; Kalashnikov et al., 2018; Julian et al., 2020). However, while it is relatively easy to craft a dense reward function during training (Gu et al., 2017; Pinto & Gupta, 2016), during deployment it is often impractical and may require substantial engineering efforts. ", + "bbox": [ + 174, + 816, + 825, + 900 + ], + "page_idx": 0 + }, + { + "type": "text", + "text": "In this paper, we tackle an alternative problem setting in vision-based RL: adapting a pre-trained policy to an unknown environment without any reward. We do this by introducing self-supervision to obtain “free” training signal during deployment. Standard self-supervised learning employs auxiliary tasks designed to automatically create training labels using only the input data (see Section 2 for details). Inspired by this, our policy is jointly trained with two objectives: a standard RL objective and, additionally, a self-supervised objective applied on an intermediate representation of the policy network. During training, both objectives are active, maximizing expected reward and simultaneously constraining the intermediate representation through self-supervision. During testing / deployment, only the self-supervised objective (on the raw observational data) remains active, forcing the intermediate representation to adapt to the new environment. ", + "bbox": [ + 174, + 103, + 825, + 242 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "We perform experiments both in simulation and with a real robot. In simulation, we evaluate on two sets of environments: DeepMind Control suite (Tassa et al., 2018) and the CRLMaze ViZDoom (Lomonaco et al., 2019; Wydmuch et al., 2018) navigation task. We evaluate generalization by testing in new environments with visual changes unknown during training. Our method improves generalization in 19 out of 22 test environments across various tasks in DeepMind Control suite, and in all considered test environments on CRLMaze. Besides simulations, we also perform Sim2Real transfer on both reaching and pushing tasks with a Kinova Gen3 robot. After training in simulation, we successfully transfer and adapt policies to 6 different environments, including continuously changing disco lights, on a real robot operating solely from an uncalibrated camera. In both simulation and real experiments, our approach outperforms domain randomization in most environments. ", + "bbox": [ + 174, + 250, + 825, + 388 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "2 RELATED WORK ", + "text_level": 1, + "bbox": [ + 176, + 400, + 344, + 416 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "Self-supervised learning is a powerful way to learn visual representations from unlabeled data (Vincent et al., 2008; Doersch et al., 2015; Wang & Gupta, 2015; Zhang et al., 2016; Pathak et al., 2016; Noroozi & Favaro, 2016; Zhang et al., 2017; Gidaris et al., 2018). Researchers have proposed to use auxiliary data prediction tasks, such as undoing rotation (Gidaris et al., 2018), solving a jigsaw puzzle (Noroozi & Favaro, 2016), tracking (Wang et al., 2019), etc. to provide supervision in lieu of labels. In RL, the idea of learning visual representations and action at the same time has been investigated (Lange & Riedmiller, 2010; Jaderberg et al., 2016; Pathak et al., 2017; Ha & Schmidhuber, 2018; Yarats et al., 2019; Srinivas et al., 2020; Laskin et al., 2020; Yan et al., 2020). For example, Srinivas et al. (2020) use self-supervised contrastive learning techniques (Chen et al., 2020; Henaff ´ et al., 2019; Wu et al., 2018; He et al., 2020) to improve sample efficiency in RL by jointly training the self-supervised objective and RL objective. However, this has not been shown to generalize to unseen environments. Other works have applied self-supervision for better generalization across environments (Pathak et al., 2017; Ebert et al., 2018; Sekar et al., 2020). For example, Pathak et al. (2017) use a self-supervised prediction task to provide dense rewards for exploration in novel environments. While results on environment exploration from scratch are encouraging, how to transfer a trained policy (with extrinsic reward) to a novel environment remains unclear. Hence, these methods are not directly applicable to the proposed problem in our paper. ", + "bbox": [ + 174, + 424, + 826, + 660 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "Generalization across different distributions is a central challenge in machine learning. In domain adaptation, target domain data is assumed to be accessible (Geirhos et al., 2018; Tzeng et al., 2017; Ganin et al., 2016; Gong et al., 2012; Long et al., 2016; Sun et al., 2019; Julian et al., 2020). For example, Tzeng et al. (2017) use adversarial learning to align the feature representations in both the source and target domain during training. Similarly, the setting of domain generalization (Ghifary et al., 2015; Li et al., 2018; Matsuura & Harada, 2019) assumes that all domains are sampled from the same meta distribution, but the same challenge remains and now becomes generalization across meta-distributions. Our work focuses instead on the setting of generalizing to truly unseen changes in the environment which cannot be anticipated at training time. ", + "bbox": [ + 174, + 666, + 825, + 791 + ], + "page_idx": 1 + }, + { + "type": "text", + "text": "There have been several recent benchmarks in our setting for image recognition (Hendrycks & Dietterich, 2018; Recht et al., 2018; 2019; Shankar et al., 2019). For example, in Hendrycks & Dietterich (2018), a classifier trained on regular images is tested on corrupted images, with corruption types unknown during training; the method of Hendrycks et al. (2019) is proposed to improve robustness on this benchmark. Following similar spirit, in the context of RL, domain randomization (Tobin et al., 2017; Pinto et al., 2017a; Peng et al., 2018; Ramos et al., 2019; Yang et al., 2019; James et al., 2019) helps a policy trained in simulation to generalize to real robots. For example, Tobin et al. (2017); Sadeghi & Levine (2016) propose to render the simulation environment with random textures and train the policy on top. The learned policy is shown to generalize to real robot manipulation tasks. Instead of deploying a fixed policy, we train and adapt the policy to the new environment with observational data that is naturally revealed during deployment. ", + "bbox": [ + 174, + 799, + 825, + 924 + ], + "page_idx": 1 + }, + { + "type": "image", + "img_path": "images/57b2c63b42fbd5db91f946cce8e82fe594fcd240b2336a571da5151324e65062.jpg", + "image_caption": [ + "Figure 1. Left: Training before deployment. Observations are sampled from a replay buffer for off-policy methods and are collected during roll-outs for on-policy methods. We optimize the RL and self-supervised objectives jointly. Right: Policy adaptation during deployment. Observations are collected from the test environment online, and we optimize only the self-supervised objective. " + ], + "image_footnote": [], + "bbox": [ + 179, + 88, + 813, + 231 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "", + "bbox": [ + 173, + 308, + 823, + 335 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "Test-time adaptation for deep learning is starting to be used in computer vision (Shocher et al., 2017; 2018; Bau et al., 2019; Mullapudi et al., 2019; Sun et al., 2020; Wortsman et al., 2018). For example, Shocher et al. (2018) shows that image super-resolution can be learned at test time (from scratch) simply by trying to upsample a downsampled version of the input image. Bau et al. (2019) show that adapting the prior of a generative adversarial network to the statistics of the test image improves photo manipulation tasks. Our work is closely related to the test-time training method of Sun et al. (2020), which performs joint optimization of image recognition and self-supervised learning with rotation prediction (Gidaris et al., 2018), then uses the self-supervised objective to adapt the representation of individual images during testing. Instead of image recognition, we perform test-time adaptation for RL with visual inputs in an online fashion. As the agent interacts with an environment, we keep obtaining new observational data in a stream for training the visual representations. ", + "bbox": [ + 173, + 342, + 825, + 496 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "3 METHOD ", + "text_level": 1, + "bbox": [ + 174, + 510, + 282, + 526 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "In this section, we describe our proposed Policy Adaptation during Deployment (PAD) approach. It can be implemented on top of any policy network and standard RL algorithm (both on-policy and off-policy) that can be described by minimizing some RL objective $J ( \\theta )$ w.r.t. the collection of parameters $\\theta$ using stochastic gradient descent. ", + "bbox": [ + 174, + 532, + 825, + 588 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "3.1 NETWORK ARCHITECTURE ", + "text_level": 1, + "bbox": [ + 176, + 604, + 403, + 618 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "We design the network architecture to allow the policy and the self-supervised prediction to share features. For the collection of parameters $\\theta$ of a given policy network $\\pi$ , we split it sequentially into $\\theta = \\left( \\theta _ { e } , \\theta _ { a } \\right)$ , where $\\theta _ { e }$ collects the parameters of the feature extractor, and $\\theta _ { a }$ is the head that outputs a distribution over actions. We define networks $\\pi _ { e }$ with parameters $\\theta _ { e }$ and $\\pi _ { a }$ with parameters $\\theta _ { a }$ such that $\\pi ( \\mathbf { s } ; \\theta ) = \\pi _ { a } ( \\pi _ { e } ( \\mathbf { s } ) )$ , where s represents an image observation. Intuitively, one can think of $\\pi _ { e }$ as a feature extractor, and $\\pi _ { a }$ as a controller based on these features. The goal of our method is to update $\\pi _ { e }$ at test-time using gradients from a self-supervised task, such that $\\pi _ { e }$ (and consequently $\\pi _ { \\theta }$ ) can generalize. Let $\\pi _ { s }$ with parameters $\\theta _ { s }$ be the self-supervised prediction head and its collection of parameters, and the input to $\\pi _ { s }$ be the output of $\\pi _ { e }$ (as illustrated in Figure 1). In this work, the self-supervised task is inverse dynamics prediction for control, and rotation prediction for navigation. ", + "bbox": [ + 173, + 626, + 825, + 765 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "3.2 INVERSE DYNAMICS PREDICTION AND ROTATION PREDICTION ", + "text_level": 1, + "bbox": [ + 176, + 781, + 650, + 796 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "At each time step, we always observe a transition sequence in the form of $\\left( \\mathbf { s } _ { t } , \\mathbf { a } _ { t } , \\mathbf { s } _ { t + 1 } \\right)$ , during both training and testing. Naturally, self-supervision can be derived from taking parts of the sequence and predicting the rest. An inverse dynamics model takes the states before and after transition, and predicts the action in between. In this work, the inverse dynamics model $\\pi _ { s }$ operates on the feature space extracted by $\\pi _ { e }$ . We can write the inverse dynamics prediction objective formally as ", + "bbox": [ + 173, + 803, + 825, + 872 + ], + "page_idx": 2 + }, + { + "type": "equation", + "img_path": "images/7468b6ceb4f801113991097451372b16a2bf35641902476ec43690677aa145b5.jpg", + "text": "$$\nL ( \\theta _ { s } , \\theta _ { e } ) = \\ell \\big ( \\mathbf { a } _ { t } , \\pi _ { s } ( \\pi _ { e } ( \\mathbf { s } _ { t } ) , \\pi _ { e } ( \\mathbf { s } _ { t + 1 } ) ) \\big ) .\n$$", + "text_format": "latex", + "bbox": [ + 357, + 875, + 640, + 893 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "For continuous actions, $\\ell$ is the mean squared error between the ground truth and the model output. For discrete actions, the output is a soft-max distribution over the action space, and $\\ell$ is the crossentropy loss. Empirically, we find this self-supervised task to be most effective with continuous actions, possibly because inverse dynamics prediction in a small space of discrete actions is not as challenging. Note that we predict the inverse dynamics instead of the forward dynamics, because when operating in feature space, the latter can produce trivial solutions such as the constant zero feature for every state2. If we instead performed prediction with forward dynamics in pixel space, the task would be extremely challenging given the large uncertainty in pixel prediction. ", + "bbox": [ + 173, + 895, + 830, + 924 + ], + "page_idx": 2 + }, + { + "type": "text", + "text": "", + "bbox": [ + 174, + 103, + 825, + 188 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "As an alternative self-supervised task, we use rotation prediction (Gidaris et al., 2018). We rotate an image by one of 0, 90, 180 and 270 degrees as input to the network, and cast this as a four-way classification problem to determine which one of these four ways the image has been rotated. This task is shown to be effective for learning representations for object configuration and scene structure, which is beneficial for visual recognition (Hendrycks et al., 2019; Doersch & Zisserman, 2017). ", + "bbox": [ + 174, + 194, + 825, + 263 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "3.3 TRAINING AND TESTING ", + "text_level": 1, + "bbox": [ + 176, + 280, + 387, + 295 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "Before deployment of the policy, because we have signals from both the reward and self-supervised auxiliary task, we can train with both in the fashion of multi-task learning. This corresponds to the following optimization problem during training $\\begin{array} { r } { \\operatorname* { m i n } _ { \\theta _ { a } , \\theta _ { s } , \\theta _ { e } } J ( \\theta _ { a } , \\theta _ { e } ) + \\alpha L ( \\theta _ { s } , \\theta _ { e } ) } \\end{array}$ , where $\\alpha > 0$ is a trade-off hyperparameter. During deployment, we cannot optimize $J$ anymore since the reward is unavailable, but we can still optimize $L$ to update both $\\theta _ { s }$ and $\\theta _ { e }$ . Empirically, we find only negligible difference with keeping $\\theta _ { s }$ fixed at test-time, so we update both since the gradients have to be computed regardless; we ablate this decision in appendix C. As we obtain new images from the stream of visual inputs in the environment, $\\theta$ keeps being updated until the episode ends. This corresponds to, for each iteration $t = 1 . . . T$ : ", + "bbox": [ + 173, + 301, + 825, + 426 + ], + "page_idx": 3 + }, + { + "type": "equation", + "img_path": "images/09a2d43c370c6c00533a9ad2763599d0bd064ebf1e85936b3031644a0c616642.jpg", + "text": "$$\n\\begin{array} { r l } & { \\quad \\mathbf s _ { t } \\sim p ( \\mathbf s _ { t } | \\mathbf a _ { t - 1 } , \\mathbf s _ { t - 1 } ) } \\\\ & { \\quad \\theta _ { s } ( t ) = \\theta _ { s } ( t - 1 ) - \\nabla _ { \\theta _ { s } } L ( \\mathbf s _ { t } ; \\theta _ { s } ( t - 1 ) , \\theta _ { e } ( t - 1 ) ) } \\\\ & { \\quad \\theta _ { e } ( t ) = \\theta _ { e } ( t - 1 ) - \\nabla _ { \\theta _ { e } } L ( \\mathbf s _ { t } ; \\theta _ { s } ( t - 1 ) , \\theta _ { e } ( t - 1 ) ) } \\\\ & { \\quad \\mathbf a _ { t } = \\pi ( \\mathbf s _ { t } ; \\theta ( t ) ) \\mathrm { ~ w i t h ~ } \\theta ( t ) = ( \\theta _ { e } ( t ) , \\theta _ { a } ) , } \\end{array}\n$$", + "text_format": "latex", + "bbox": [ + 326, + 429, + 669, + 501 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "where $\\theta _ { s } ( 0 ) = \\theta _ { s } , \\theta _ { e } ( 0 ) = \\theta _ { e }$ , $\\mathbf { s } _ { 0 }$ is the initial condition given by the environment, $\\mathbf { a } _ { 0 } = \\pi _ { \\theta } ( \\mathbf { s } _ { 0 } )$ , $p$ is the unknown environment transition, and $L$ is the self-supervised objective as previously introduced. ", + "bbox": [ + 174, + 502, + 825, + 530 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "4 EXPERIMENTS ", + "text_level": 1, + "bbox": [ + 176, + 559, + 326, + 575 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "In this work, we investigate how well an agent trained in one environment (denoted the training environment) generalizes to unseen and diverse test environments. During evaluation, agents have no access to reward signals and are expected to generalize without trials nor prior knowledge about the test environments. In simulation, we evaluate our method (PAD) and baselines extensively on continuous control tasks from DeepMind Control (DMControl) suite (Tassa et al., 2018) as well as the CRLMaze (Lomonaco et al., 2019) navigation task, and experiment with both stationary (colors, objects, textures, lighting) and non-stationary (videos) environment changes. We further show that PAD transfers from simulation to a real robot and successfully adapts to environmental differences during deployment in two robotic manipulation tasks. Samples from DMControl and CRLMaze environments are shown in Figure 2, and samples from the robot experiments are shown in Figure 4. Implementation is available at https://nicklashansen.github.io/PAD/. ", + "bbox": [ + 173, + 580, + 825, + 733 + ], + "page_idx": 3 + }, + { + "type": "text", + "text": "Network details. For DMControl and the robotic manipulation tasks we implement PAD on top of Soft Actor-Critic (SAC) (Haarnoja et al., 2018), and adopt both network architecture and hyperparameters from Yarats et al. (2019), with minor modifications: the feature extractor $\\pi _ { e }$ has 8 convolutional layers shared between the RL head $\\pi _ { a }$ and self-supervised head $\\pi _ { s }$ , and we split the network into architecturally identical heads following $\\pi _ { e }$ . Each head consists of 3 convolutional layers followed by 4 fully connected layers. For CRLMaze, we use Advantage Actor-Critic (A2C) as base algorithm (Mnih et al., 2016) and apply the same architecture as for the other experiments, but implement $\\pi _ { e }$ with only 6 convolutional layers. Observations are stacks of $k$ colored frames $k = 3$ on DMControl and CRLMaze; $k = 1$ in robotic manipulation) of size $1 0 0 \\times 1 0 0$ and time-consistent random crop is applied as in Srinivas et al. (2020). During deployment, we optimize the self-supervised objective online w.r.t. $\\theta _ { e } , \\theta _ { s }$ for one gradient step per time iteration. See appendix F for implementation details. ", + "bbox": [ + 174, + 739, + 825, + 893 + ], + "page_idx": 3 + }, + { + "type": "image", + "img_path": "images/bc62cee880a17c8b0b2832092fcc074bd1b01dcd8fb62086e20850e09ae21b15.jpg", + "image_caption": [ + "Figure 2. Left: Training environments of DMControl (top) and CRLMaze (bottom). Right: Test environments of DMControl (top) and CRLMaze (bottom). Changes to DMControl include randomized colors, video backgrounds, and distractors; changes to CRLMaze include textures and lighting. " + ], + "image_footnote": [], + "bbox": [ + 189, + 101, + 808, + 209 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "Table 1. Episodic return in test environments with randomized colors, mean and std. dev. for 10 seeds. Best method on each task is in bold and blue compares $\\mathrm { S A C + I D M }$ with and without PAD. ", + "bbox": [ + 174, + 266, + 825, + 295 + ], + "page_idx": 4 + }, + { + "type": "table", + "img_path": "images/4c3f63f639cec5d2c3bd77b6d9bbea9f1e2c7a8f87d55cc055dbf22a706060c4.jpg", + "table_caption": [ + "10x episode length " + ], + "table_footnote": [], + "table_body": "
Random colorsSAC+DR+IDM+IDM (PAD)+IDM+IDM (PAD)
Walker, walk414±74594±104406±29468±473830±5475505±592
Walker, stand719±74715±96743±37797±467832±2098566±121
Cartpole, swingup592±50647±48585±73630±636528±5397093±592
Cartpole,balance857±60867±37835±40848±297746±5267670±293
Ball in cup,catch411±183470±252471±75563±50
Finger, spin626±163465±314757±62803±727249±6427496±655
Finger, turn_easy270±43167±26283±51304±461
Cheetah, run154±41145±29121±38159±281117±5301208±487
Reacher, easy163±45105±37201±32214±441788±4412152±506
", + "bbox": [ + 207, + 310, + 787, + 449 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "4.1 DEEPMIND CONTROL ", + "text_level": 1, + "bbox": [ + 174, + 469, + 366, + 483 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "DeepMind Control (DMControl) (Tassa et al., 2018) is a collection of continuous control tasks where agents only observe raw pixels. Generalization benchmarks on DMControl represent diverse real-world tasks for motor control, and contain distracting surroundings not correlated with the reward signals. ", + "bbox": [ + 173, + 492, + 825, + 549 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "Experimental setup. We experiment with 9 tasks from DMControl and measure generalization to four types of test environments: (i) randomized colors; (ii) natural videos as background; (iii) distracting objects placed in the scene; and (iv) the unmodified training environment. For each test environment, we evaluate methods across 10 seeds and 100 random initializations. If a given test environment is not applicable to certain tasks, e.g. if a task has no background for the video background setting, they are excluded. Tasks are selected on the basis of diversity, as well as the success of vision-based RL in prior work (Yarats et al., 2019; Srinivas et al., 2020; Laskin et al., 2020; Kostrikov et al., 2020). We implement PAD on top of SAC and use an Inverse Dynamics Model (IDM) for self-supervision, as we find that learning a model of the dynamics works well for motor control. For completeness, we ablate the choice of self-supervision. Learning curves are provided in appendix B. ", + "bbox": [ + 174, + 556, + 583, + 776 + ], + "page_idx": 4 + }, + { + "type": "image", + "img_path": "images/1d57abcc32301ede5d2b8455a03059fd30b9cd73bdc63abef544626395d0fffc.jpg", + "image_caption": [ + "Figure 3. Relative improvement in instantaneous reward over time for PAD on the random color env. " + ], + "image_footnote": [], + "bbox": [ + 593, + 563, + 820, + 720 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "We compare our method to the following baselines: (i) SAC with no changes (denoted $S A C$ ); (ii) SAC trained with domain randomization on a fixed set of 100 colors (denoted $+ D R$ ); and (iii) SAC trained jointly with an IDM but without PAD (denoted $+ I D M )$ ). Our method using an IDM with PAD is denoted by $+ I D M \\left( P A D \\right)$ . For domain randomization, colors are sampled from the same distribution as in evaluation, but with lower variance, as we find that training directly on the test distribution does not converge. ", + "bbox": [ + 173, + 777, + 825, + 861 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "Random perturbation of color. Robustness to subtle changes such as color is essential to realworld deployment of RL policies. We evaluate generalization on a fixed set of 100 colors of foreground, background and the agent itself, and report the results in Table 1 (first 4 columns). We find PAD to improve generalization in all tasks considered, outperforming SAC trained with domain randomization in 6 out of 9 tasks. Surprisingly, despite a substantial overlap between training and test domains of domain randomization, it generalizes no better than vanilla SAC on a majority of tasks. ", + "bbox": [ + 174, + 867, + 825, + 924 + ], + "page_idx": 4 + }, + { + "type": "text", + "text": "", + "bbox": [ + 173, + 103, + 823, + 132 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "Long-term stability. We find the relative improvement of PAD to improve over time, as shown in Figure 3. To examine the long-term stability of PAD, we further evaluate on $1 0 \\mathrm { x }$ episode lengths and summarize the results in the last two columns in Table 1 (goal-oriented tasks excluded). While we do not explicitly prevent the embedding from drifting away from the RL task, we find empirically that PAD does not degrade the performance of the policy, even over long horizons, and when PAD does not improve, we find it to hurt minimally. We conjecture this is because we are not learning a new task, but simply continue to optimize the same (self-supervised) objective as during joint training, where both two tasks are compatible. In this setting, PAD still improves generalization in 6 out of 7 tasks, and thus naturally extends beyond episodic deployment. For completeness, we also evaluate methods in the environment in which they were trained, and report the results in appendix A. We find that, while PAD improves generalization to novel environments, performance is virtually unchanged on the training environment. We conjecture this is because the self-supervised task is already fully learned and any continued training on the same data distribution thus has little impact. ", + "bbox": [ + 173, + 138, + 825, + 319 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "Non-stationary environments. To investigate whether PAD can adapt in non-stationary environments, we evaluate generalization to diverse video backgrounds (refer to Figure 2). We find PAD to outperform all baselines on 7 out of 8 tasks, as shown in Table 2, by as much as $104 \\%$ over domain randomization on Finger, spin. Domain randomization generalizes comparably worse to videos, which we conjecture is not because the environments are non-stationary, but rather because the image statistics of videos are not covered by its training domain of randomized colors. In fact, domain randomization is outperformed by the vanilla SAC in most tasks with video back", + "bbox": [ + 174, + 327, + 387, + 631 + ], + "page_idx": 5 + }, + { + "type": "table", + "img_path": "images/4bc66c9ffdfda4de28170ce557a14b812efeda8a00f84a592bc63390067c438f.jpg", + "table_caption": [ + "Table 2. Episodic return in test environments with video backgrounds (top) and distracting objects (bottom), mean and std. dev. for 10 seeds. Best method on each task is in bold and blue compares $\\mathrm { S A C + I D M }$ with and without PAD. " + ], + "table_footnote": [ + "grounds, which is in line with the findings of Packer et al. (2018). " + ], + "table_body": "
Video backgroundsSAC+DR+IDM+IDM (PAD)
Walker, walk616±80655±55694±85717±79
Walker, stand899±53869±60902±51935±20
Cartpole, swingup375±90485±67487±90521±76
Cartpole, balance693±109766±92691±76687±58
Ball in cup, catch393±175271±189362±69436±55
Finger, spin447±102338±207605±61691±80
Finger, turn_easy355±108223±91355±110362±101
Cheetah, run194±30150±34164±42206±34
Distracting objectsSAC+DR+IDM+IDM (PAD)
Cartpole, swingup815±60809±24776±58771±64
Cartpole,balance969±20938±35964±26960±29
Ball in cup, catch177±111331±189482±128545±173
Finger, spin652±184564±288836±62867±72
Finger, turn_easy302±68165±12326±101347±48
", + "bbox": [ + 400, + 390, + 831, + 613 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "Scene content. We hypothesize that: (i) an agent trained with an IDM is comparably less distracted by scene content since objects uncorrelated to actions yield no predictive power; and (ii) that PAD can adapt to unexpected objects in the scene. We test these hypotheses by measuring robustness to colored shapes at a variety of positions in both the foreground and background of the scene (no physical interaction). Results are summarized in Table 2. PAD outperforms all baselines in 3 out of 5 tasks, with a relative improvement of $20 \\%$ over SAC on Ball in cup, catch. In the two cartpole tasks in which PAD does not improve, all methods are already relatively unaffected by the distractors. ", + "bbox": [ + 174, + 651, + 825, + 750 + ], + "page_idx": 5 + }, + { + "type": "text", + "text": "Choice of self-supervised task. We investigate how much the choice of self-supervised task contributes to the overall success of our method, and consider the following ablations: (i) replacing inverse dynamics with the rotation prediction task described in Section 3.2; and (ii) replacing it with the recently proposed CURL (Srinivas et al., 2020) contrastive learning algorithm for RL. As shown in Table 3, PAD improves generalization of CURL in a majority of tasks on the randomized color benchmark, and in 4 out of 9 tasks using rotation prediction. However, inverse dynamics as auxiliary task produces more consistent results and offers better generalization overall. We argue that learning an IDM produces better representations for motor control since it connects observations directly to actions, whereas CURL and rotation prediction operates purely on observations. In general, we find the improvement of PAD to be bigger in tasks that benefit significantly from visual information (see appendix A), and conjecture that selecting a self-supervised task that learns features useful to the RL task is crucial to the success of PAD, which we discuss further in Section 4.2. ", + "bbox": [ + 174, + 756, + 825, + 922 + ], + "page_idx": 5 + }, + { + "type": "table", + "img_path": "images/bf344df1e165df56c80486d7f51964bb5b74eb3e2e0b8e22fc58b0af02a686ee.jpg", + "table_caption": [ + "Table 3. Ablations on the randomized color domain of DMC. All methods use SAC. CURL represents RL with a contrastive learning task (Srinivas et al., 2020) and Rot represents the rotation prediction (Gidaris et al., 2018). Offline PAD is here denoted O-PAD for brevity, whereas the default usage of PAD is in an online setting. Best method is in bold and blue compares $+ \\mathrm { I D M }$ w/ and w/o PAD. " + ], + "table_footnote": [], + "table_body": "
Random colorsCURLCURL (PAD)RotRot (PAD)IDMIDM (O-PAD)IDM (PAD)
Walker, walk445±99495±70335±7330±30406±29441±16468±47
Walker, stand662±54753±49673±4653±27743±37727±21797±46
Cartpole, swingup454±110413±67493±52477±38585±73578±69630±63
Cartpole,balance782±13763±5710±72734±81835±40796±37848±29
Ball in cup, catch231±92332±78291±54314±60471±75490±16563±50
Finger, spin691±12588±22695±36689±20757±62767±43803±72
Finger, turn_easy202±32186±2283±68230±53283±51321±10304±46
Cheetah, run202±22211±20127±3135±12121±38112±35159±28
Reacher, easy325±32378±6299±29120±7201±32241±24214±44
", + "bbox": [ + 173, + 164, + 828, + 299 + ], + "page_idx": 6 + }, + { + "type": "table", + "img_path": "images/15c39b22f8aeb5f17c614888a12e0481844a65b7a4bca0ef09518735e3dd9696.jpg", + "table_caption": [ + "Table 4. Episodic return of PAD and baselines in CRLMaze environments. PAD improves generalization in all considered environments and outperforms both A2C and domain randomization by a large margin. All methods use A2C. We report mean and std. error of 10 seeds. Best method in each environment is in bold and blue compares rotation prediction with and without PAD. " + ], + "table_footnote": [], + "table_body": "
CRLMazeRandomA2C+DR+IDM+IDM (PAD)+Rot+Rot (PAD)
Walls-870±30-380±145-260±137-302±150-428±135-206±166-74±116
Floor-868±23-320±167-438±59-47±198-530±106-294±123-209±94
Ceiling-872±30-171±175-400±74166±215-508±104128±196281±83
Lights-900±29-30±213-310±106239±270-460±114-84±53312±104
", + "bbox": [ + 173, + 372, + 821, + 452 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Offline versus online learning. Observations that arrive sequentially are highly correlated, and we thus hypothesize that our method benefits significantly from learning online. To test this hypothesis, we run an offline variant of our method in which network updates are forgotten after each step. In this setting, our method can only adapt to single observations and does not benefit from learning over time. Results are shown in Table 3. We find that our method benefits substantially from online learning, but learning offline still improves generalization on select tasks. ", + "bbox": [ + 173, + 469, + 825, + 554 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "4.2 CRLMAZE ", + "text_level": 1, + "bbox": [ + 174, + 568, + 292, + 582 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "CRLMaze (Lomonaco et al., 2019) is a time-constrained, discrete-action 3D navigation task for ViZDoom (Wydmuch et al., 2018), in which an agent is to navigate a maze and collect objects. There is a positive reward associated with green columns, and a negative reward for lanterns as well as for living. Readers are referred to the respective papers for details on the task and environment. ", + "bbox": [ + 174, + 589, + 825, + 645 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Experimental setup. We train agents on a single environment and measure generalization to environments with novel textures for walls, floor, and ceiling, as well as lighting, as shown in Figure 2. We implement PAD on top of A2C (Mnih et al., 2016) and use rotation prediction (see Section 3.2) as self-supervised task. Learning to navigate novel scenes requires a generalized scene understanding, and we find that rotation prediction facilitates that more so than an IDM. We compare to the following baselines: (i) a random agent (denoted Random); (ii) A2C with no changes (denoted $A 2 C$ ); (iii) A2C trained with domain randomization (denoted $+ D R$ ); (iv) A2C with an IDM as auxiliary task (denoted $+ I D M )$ ; and (v) A2C with rotation prediction as auxiliary task (denoted $+ R o t )$ . We denote Rot with PAD as $+ R o t \\ ( P A D )$ . Domain randomization uses 56 combinations of diverse textures, partially overlapping with the test distribution, and we find it necessary to train domain randomization for twice as many episodes in order to converge. We closely follow the evaluation procedure of (Lomonaco et al., 2019) and evaluate methods across 20 starting positions and 10 random seeds. ", + "bbox": [ + 173, + 652, + 825, + 819 + ], + "page_idx": 6 + }, + { + "type": "text", + "text": "Results. We report performance on the CRLMaze environments in Table 4. PAD improves generalization in all considered test environments, outperforming both A2C and domain randomization by a large margin. Domain randomization performs consistently across all environments but is less successful overall. We further examine the importance of selecting appropriate auxiliary tasks by a simple ablation: replacing rotation prediction with an IDM for the navigation task. We conjecture that, while an auxiliary task can enforce structure in the learned representations, its features (and consequently gradients) need to be sufficiently correlated with the primary RL task for PAD to be ", + "bbox": [ + 174, + 825, + 825, + 924 + ], + "page_idx": 6 + }, + { + "type": "image", + "img_path": "images/7a75facacf5c2ac9f0591865730aa213a149b2aae09f6d5a1610fc3b0f899d74.jpg", + "image_caption": [ + "(a) Simulation. (b) Default transfer. (c) Table cloth. (d) Disco lights. " + ], + "image_footnote": [], + "bbox": [ + 178, + 95, + 820, + 194 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "Figure 4. Samples from the push robotic manipulation task. The task is to push the yellow cube to the location of the red disc. Agents are trained in setting (a) and evaluated in settings (b-d). ", + "bbox": [ + 173, + 215, + 823, + 243 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "successful during deployment. While PAD with rotation prediction improves generalization across all test environments considered, IDM does not, which suggests that rotation prediction is more suitable for tasks that require scene understanding, whereas IDM is useful for tasks that require motor control. We leave it to future work to automate the process of selecting appropriate auxiliary tasks. ", + "bbox": [ + 174, + 257, + 825, + 314 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "4.3 ROBOTIC MANIPULATION TASKS ", + "text_level": 1, + "bbox": [ + 176, + 333, + 437, + 347 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "We deploy our method and baselines on a real Kinova Gen3 robot and evaluate on two manipulation tasks: (i) reach, a task in which the robot reaches for a goal marked by a red disc; and (ii) push, a task in which the robot pushes a cube to the location of the red disc. Both tasks use an XY action space, where the Z position of the actuator is fixed. Agents operate purely from pixel observations with no access to state information. During deployment, we make no effort to calibrate camera, lighting, or ", + "bbox": [ + 173, + 354, + 450, + 521 + ], + "page_idx": 7 + }, + { + "type": "table", + "img_path": "images/6d9f88483285c3d154ac3b567adc05a0f37c1c4a8436c19262f4605b0d8b8d8b.jpg", + "table_caption": [ + "Table 5. Success rate of PAD and baselines on a real robotic arm. Best method in each environment is in bold and blue compares $+ \\mathrm { I D M }$ with and without PAD. " + ], + "table_footnote": [], + "table_body": "
Real robotSAC+DR+IDM+IDM (PAD)
Reach (default)100%100%100%100%
Reach (cloth)48%80%56%80%
Reach (disco)72%76%88%92%
Push (default)88%88%92%100%
Push (cloth)60%64%64%88%
Push (disco)60%68%72%84%
", + "bbox": [ + 464, + 406, + 820, + 512 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "physical properties such as dimensions, mass, and friction, and policies are expected to generalize with no prior knowledge of the test environment. Samples from the push task are shown in Figure 4, and samples from reach are shown in appendix E. ", + "bbox": [ + 174, + 522, + 826, + 563 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "Experimental setup. We implement PAD on top of SAC (Haarnoja et al., 2018) and apply the same experimental setup as in Section 4.1 using an Inverse Dynamics Model (IDM) for self-supervision, but without frame-stacking (i.e. $k = 1$ ). Agents are trained in simulation with dense rewards and randomized initial configurations of arm, goal, and box, and we measure generalization to 3 novel environments in the real-world: (i) default environment with pixel observations that roughly mimic the simulation; (ii) a patterned table cloth that ", + "bbox": [ + 174, + 570, + 450, + 750 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "Table 6. Success rate of PAD and baselines for the push task on a simulated robotic arm in test environments with changes to dynamics. Changes include object mass, size, and friction, arm mount position, and end effector velocity. Best method in each environment is in bold and blue compares $+ \\mathrm { I D M }$ with and without PAD. ", + "bbox": [ + 465, + 570, + 823, + 654 + ], + "page_idx": 7 + }, + { + "type": "table", + "img_path": "images/893e85f9cde46ec61efaa4988ff881556760c85e430326592a734820a3ce89f0.jpg", + "table_caption": [], + "table_footnote": [], + "table_body": "
Simulated robotSAC+DR+IDM+IDM (PAD)
Push (object)66%64%72%82%
Push (mount)68%58%86%84%
Push (velocity)70%68%70%78%
Push (all)56%50%48%76%
", + "bbox": [ + 464, + 664, + 820, + 739 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "distracts visually and greatly increases friction; and (iii) disco, an environment with non-stationary visual disco light distractions. Notably, all 3 environments also feature subtle differences in dynamics compared to the training environment, such as object dimensions, mass, friction, and uncalibrated actions. In each setting, we evaluate the success rate across 25 test runs spanning across 5 pre-defined goal locations throughout the table. The goal locations vary between the two tasks, and the robot is reset after each run. We perform comparison against direct transfer and domain randomization baselines as in Section 4.1. We further evaluate generalization to changes in dynamics by considering a variant of the simulated environment in which object mass, size, and friction, arm mount position, and end effector velocity is modified. We consider each setting both individually and jointly, and evaluate success rate across 50 unique configurations with the robot reset after each run. ", + "bbox": [ + 173, + 750, + 825, + 888 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "Results. We report transfer results in Table 5. While all methods transfer successfully to reach (default), we observe PAD to improve generalization in all settings in which the baselines show sub-optimal performance. We find PAD to be especially powerful for the push task that involves dynamics, improving by as much as $24 \\%$ in push (cloth). While domain randomization proves highly effective in reach (cloth), we observe no significant benefit in the other settings, which suggests that PAD can be more suitable in challenging tasks like push. To isolate the effect of dynamics, we further evaluate generalization to a number of simulated changes in dynamics on the push task. Results are shown in Table 6. We find PAD to improve generalization to changes in the physical properties of the object and end effector, whereas both $S A C { + } I D M$ and PAD are relatively unaffected by changes to the mount position. Consistent with the real robot results in Section 5, PAD is found to be most effective when changes in dynamics are non-trivial, improving by as much as $28 \\%$ in the push (all) setting, where all 3 environmental changes are considered jointly. These results suggest that PAD can be a simple, yet effective method for generalization to diverse, unseen environments that vary in both visuals and dynamics. ", + "bbox": [ + 173, + 895, + 825, + 924 + ], + "page_idx": 7 + }, + { + "type": "text", + "text": "", + "bbox": [ + 174, + 103, + 825, + 270 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "5 CONCLUSION ", + "text_level": 1, + "bbox": [ + 176, + 291, + 318, + 308 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "While previous work addresses generalization in RL by learning policies that are invariant to any environment changes that can be anticipated, we formulate an alternative problem setting in visionbased RL: can we instead adapt a pretrained-policy to new environments without any reward. We propose Policy Adaptation during Deployment, a self-supervised framework for online adaptation at test-time, and show empirically that our method improves generalization of policies to diverse simulated and real-world environmental changes across a variety of tasks. We find our approach benefits greatly from learning online, and we systematically evaluate how the choice of self-supervised task impacts performance. While the current framework relies on prior knowledge on selecting selfsupervised tasks for policy adaptation, we see our work as the initial step in addressing the problem of adapting vision-based policies to unknown environments. We ultimately envision embodied agents in the future to be learning all the time, with the flexibility to learn both with and without rewards, before and during deployment. ", + "bbox": [ + 174, + 319, + 825, + 484 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "Acknowledgements. This work was supported, in part, by grants from DARPA, NSF 1730158 CI-New: Cognitive Hardware and Software Ecosystem Community Infrastructure (CHASE-CI), NSF ACI-1541349 CC\\*DNI Pacific Research Platform, and gifts from Qualcomm and TuSimple. This work was also funded, in part, by grants from Berkeley DeepDrive, SAP and European Research Council (ERC) from the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE). We would like to thank Fenglu Hong and Joey Hejna for helpful discussions. ", + "bbox": [ + 174, + 502, + 825, + 598 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "REFERENCES ", + "text_level": 1, + "bbox": [ + 176, + 621, + 285, + 636 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "OpenAI: Marcin Andrychowicz, Bowen Baker, Maciek Chociej, Rafal Jozefowicz, Bob McGrew, Jakub Pachocki, Arthur Petron, Matthias Plappert, Glenn Powell, Alex Ray, et al. Learning dexterous in-hand manipulation. The International Journal of Robotics Research, 39(1):3–20, 2020. 1 \nDavid Bau, Hendrik Strobelt, William Peebles, Jonas Wulff, Bolei Zhou, Jun-Yan Zhu, and Antonio Torralba. Semantic photo manipulation with a generative image prior. ACM Trans. Graph., 38(4), 2019. ISSN 0730-0301. 3 \nTing Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. A simple framework for contrastive learning of visual representations, 2020. 2 \nCarl Doersch and Andrew Zisserman. Multi-task self-supervised visual learning. In Proceedings of the IEEE International Conference on Computer Vision, pp. 2051–2060, 2017. 4 \nCarl Doersch, Abhinav Gupta, and Alexei A Efros. Unsupervised visual representation learning by context prediction. In Proceedings of the IEEE International Conference on Computer Vision, pp. 1422–1430, 2015. 2 \nFrederik Ebert, Chelsea Finn, Sudeep Dasari, Annie Xie, Alex Lee, and Sergey Levine. Visual foresight: Model-based deep reinforcement learning for vision-based robotic control, 2018. 2 ", + "bbox": [ + 171, + 643, + 828, + 924 + ], + "page_idx": 8 + }, + { + "type": "text", + "text": "Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pascal Germain, Hugo Larochelle, Franc¸ois Laviolette, Mario Marchand, and Victor Lempitsky. Domain-adversarial training of neural networks. The Journal of Machine Learning Research, 17(1):2096–2030, 2016. 2 ", + "bbox": [ + 178, + 103, + 825, + 146 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Robert Geirhos, Carlos R. Medina Temme, Jonas Rauber, Heiko H. Schutt, Matthias Bethge, and ¨ Felix A. Wichmann. Generalisation in humans and deep neural networks. In NeurIPS, 2018. 2 ", + "bbox": [ + 174, + 156, + 823, + 185 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Muhammad Ghifary, W. Bastiaan Kleijn, Mengjie Zhang, and David Balduzzi. Domain generalization for object recognition with multi-task autoencoders. In Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV), ICCV ’15, pp. 2551–2559. IEEE Computer Society, 2015. ISBN 9781467383912. 2 ", + "bbox": [ + 173, + 195, + 826, + 251 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Spyros Gidaris, Praveer Singh, and Nikos Komodakis. Unsupervised representation learning by predicting image rotations, 2018. 2, 3, 4, 7 ", + "bbox": [ + 173, + 262, + 823, + 291 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Boqing Gong, Yuan Shi, Fei Sha, and Kristen Grauman. Geodesic flow kernel for unsupervised domain adaptation. In 2012 IEEE Conference on Computer Vision and Pattern Recognition, pp. 2066–2073. IEEE, 2012. 2 ", + "bbox": [ + 173, + 301, + 823, + 343 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Shixiang Gu, Ethan Holly, Timothy Lillicrap, and Sergey Levine. Deep reinforcement learning for robotic manipulation with asynchronous off-policy updates. In 2017 IEEE international conference on robotics and automation (ICRA), pp. 3389–3396. IEEE, 2017. 1 ", + "bbox": [ + 174, + 354, + 823, + 397 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "David Ha and Jurgen Schmidhuber. Recurrent world models facilitate policy evolution. In ¨ Advances in Neural Information Processing Systems 31, pp. 2451–2463. Curran Associates, Inc., 2018. 2 ", + "bbox": [ + 174, + 406, + 823, + 436 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Tuomas Haarnoja, Aurick Zhou, Kristian Hartikainen, George Tucker, Sehoon Ha, Jie Tan, Vikash Kumar, Henry Zhu, Abhishek Gupta, Pieter Abbeel, and Sergey Levine. Soft actor-critic algorithms and applications, 2018. 4, 8, 17 ", + "bbox": [ + 174, + 445, + 825, + 488 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. Momentum contrast for unsupervised visual representation learning. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020. 2 ", + "bbox": [ + 174, + 500, + 826, + 541 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Dan Hendrycks and Thomas G. Dietterich. Benchmarking neural network robustness to common corruptions and surface variations. arXiv: Learning, 2018. 2 ", + "bbox": [ + 168, + 551, + 823, + 582 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Dan Hendrycks, Mantas Mazeika, Saurav Kadavath, and Dawn Song. Using self-supervised learning can improve model robustness and uncertainty. ArXiv, abs/1906.12340, 2019. 2, 4 ", + "bbox": [ + 173, + 592, + 823, + 621 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Olivier J. Henaff, Aravind Srinivas, Jeffrey De Fauw, Ali Razavi, Carl Doersch, S. M. Ali Eslami, and ´ Aaron van den Oord. Data-efficient image recognition with contrastive predictive coding, 2019. 2 ", + "bbox": [ + 171, + 631, + 825, + 660 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Max Jaderberg, Volodymyr Mnih, Wojciech Marian Czarnecki, Tom Schaul, Joel Z Leibo, David Silver, and Koray Kavukcuoglu. Reinforcement learning with unsupervised auxiliary tasks, 2016. 2 ", + "bbox": [ + 174, + 669, + 825, + 710 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Stephen James, Paul Wohlhart, Mrinal Kalakrishnan, D. Kalashnikov, A. Irpan, J. Ibarz, S. Levine, Raia Hadsell, and Konstantinos Bousmalis. Sim-to-real via sim-to-sim: Data-efficient robotic grasping via randomized-to-canonical adaptation networks. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 12619–12629, 2019. 2 ", + "bbox": [ + 173, + 722, + 825, + 780 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "R. Julian, B. Swanson, G. Sukhatme, Sergey Levine, Chelsea Finn, and Karol Hausman. Never stop learning: The effectiveness of fine-tuning in robotic reinforcement learning. arXiv: Learning, 2020. 1, 2, 15 ", + "bbox": [ + 174, + 789, + 825, + 832 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "D. Kalashnikov, A. Irpan, P. Pastor, J. Ibarz, A. Herzog, Eric Jang, Deirdre Quillen, Ethan Holly, Mrinal Kalakrishnan, V. Vanhoucke, and S. Levine. Qt-opt: Scalable deep reinforcement learning for vision-based robotic manipulation. ArXiv, abs/1806.10293, 2018. 1 ", + "bbox": [ + 176, + 842, + 825, + 885 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Ilya Kostrikov, Denis Yarats, and Rob Fergus. Image augmentation is all you need: Regularizing deep reinforcement learning from pixels. 2020. 5, 17 ", + "bbox": [ + 173, + 895, + 821, + 924 + ], + "page_idx": 9 + }, + { + "type": "text", + "text": "Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton. Imagenet classification with deep convolutional neural networks. In NIPS, 2012. 17 ", + "bbox": [ + 171, + 103, + 825, + 132 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Sascha Lange and Martin A. Riedmiller. Deep auto-encoder neural networks in reinforcement learning. In The 2010 International Joint Conference on Neural Networks (IJCNN), pp. 1–8, 2010. 2 ", + "bbox": [ + 176, + 142, + 825, + 185 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Michael Laskin, Kimin Lee, Adam Stooke, Lerrel Pinto, Pieter Abbeel, and Aravind Srinivas. Reinforcement learning with augmented data. arXiv preprint arXiv:2004.14990, 2020. 2, 5, 17 ", + "bbox": [ + 176, + 195, + 825, + 224 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Kimin Lee, Kibok Lee, Jinwoo Shin, and Honglak Lee. A simple randomization technique for generalization in deep reinforcement learning. ArXiv, abs/1910.05396, 2019. 1 ", + "bbox": [ + 174, + 234, + 821, + 263 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Sergey Levine, Chelsea Finn, Trevor Darrell, and Pieter Abbeel. End-to-end training of deep visuomotor policies. The Journal of Machine Learning Research, 17(1):1334–1373, 2016. ", + "bbox": [ + 174, + 273, + 825, + 303 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Ya Feng Li, Mingming Gong, Xinmei Tian, Tongliang Liu, and Dacheng Tao. Domain generalization via conditional invariant representations. In AAAI, 2018. 2 ", + "bbox": [ + 173, + 313, + 825, + 343 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Vincenzo Lomonaco, Karen Desai, Eugenio Culurciello, and Davide Maltoni. Continual reinforcement learning in 3d non-stationary environments. arXiv preprint arXiv:1905.10112, 2019. 2, 4, 7, 16, 17 ", + "bbox": [ + 176, + 352, + 823, + 395 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Mingsheng Long, Han Zhu, Jianmin Wang, and Michael I Jordan. Unsupervised domain adaptation with residual transfer networks. In Advances in Neural Information Processing Systems, pp. 136–144, 2016. 2 ", + "bbox": [ + 173, + 405, + 825, + 448 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Toshihiko Matsuura and Tatsuya Harada. Domain generalization using a mixture of multiple latent domains, 2019. 2 ", + "bbox": [ + 173, + 458, + 821, + 488 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Alex Graves, Ioannis Antonoglou, Daan Wierstra, and Martin Riedmiller. Playing atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602, 2013. 1 ", + "bbox": [ + 174, + 497, + 825, + 540 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Volodymyr Mnih, Adria Puigdom \\` enech Badia, Mehdi Mirza, Alex Graves, Timothy P. Lillicrap, Tim \\` Harley, David Silver, and Koray Kavukcuoglu. Asynchronous methods for deep reinforcement learning, 2016. 4, 7, 17 ", + "bbox": [ + 174, + 550, + 823, + 594 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Ravi Teja Mullapudi, Steven Chen, Keyi Zhang, Deva Ramanan, and Kayvon Fatahalian. Online model distillation for efficient video inference. 2019 IEEE/CVF International Conference on Computer Vision (ICCV), Oct 2019. doi: 10.1109/iccv.2019.00367. 3 ", + "bbox": [ + 173, + 603, + 823, + 647 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Ashvin V Nair, Vitchyr Pong, Murtaza Dalal, Shikhar Bahl, Steven Lin, and Sergey Levine. Visual reinforcement learning with imagined goals. In Advances in Neural Information Processing Systems, pp. 9191–9200, 2018. 1 ", + "bbox": [ + 173, + 656, + 825, + 700 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Mehdi Noroozi and Paolo Favaro. Unsupervised learning of visual representations by solving jigsaw puzzles. In European Conference on Computer Vision, pp. 69–84. Springer, 2016. 2 ", + "bbox": [ + 173, + 710, + 825, + 739 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Charles Packer, Katelyn Gao, Jernej Kos, Philipp Krahenb ¨ uhl, Vladlen Koltun, and Dawn Song. ¨ Assessing generalization in deep reinforcement learning, 2018. 6 ", + "bbox": [ + 173, + 750, + 825, + 779 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Deepak Pathak, Philipp Krahenbuhl, Jeff Donahue, Trevor Darrell, and Alexei A Efros. Context encoders: Feature learning by inpainting. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 2536–2544, 2016. 2 ", + "bbox": [ + 174, + 787, + 825, + 832 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Deepak Pathak, Pulkit Agrawal, Alexei A. Efros, and Trevor Darrell. Curiosity-driven exploration by self-supervised prediction. In ICML, 2017. 2 ", + "bbox": [ + 169, + 842, + 825, + 871 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Xue Bin Peng, Marcin Andrychowicz, Wojciech Zaremba, and Pieter Abbeel. Sim-to-real transfer of robotic control with dynamics randomization. 2018 IEEE International Conference on Robotics and Automation (ICRA), May 2018. 1, 2 ", + "bbox": [ + 174, + 882, + 825, + 924 + ], + "page_idx": 10 + }, + { + "type": "text", + "text": "Lerrel Pinto and Abhinav Gupta. Supersizing self-supervision: Learning to grasp from 50k tries and 700 robot hours. In 2016 IEEE international conference on robotics and automation (ICRA), pp. 3406–3413. IEEE, 2016. 1 ", + "bbox": [ + 174, + 103, + 825, + 146 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Lerrel Pinto, Marcin Andrychowicz, Peter Welinder, Wojciech Zaremba, and Pieter Abbeel. Asymmetric actor critic for image-based robot learning. arXiv preprint arXiv:1710.06542, 2017a. 1, 2 ", + "bbox": [ + 176, + 155, + 823, + 196 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Lerrel Pinto, James Davidson, Rahul Sukthankar, and Abhinav Gupta. Robust adversarial reinforcement learning. In Proceedings of the 34th International Conference on Machine Learning-Volume 70, pp. 2817–2826. JMLR. org, 2017b. 1 ", + "bbox": [ + 174, + 207, + 825, + 250 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Aravind Rajeswaran, Sarvjeet Ghotra, Balaraman Ravindran, and Sergey Levine. Epopt: Learning robust neural network policies using model ensembles. arXiv preprint arXiv:1610.01283, 2016. 1 ", + "bbox": [ + 171, + 258, + 823, + 287 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Fabio Ramos, Rafael Possas, and Dieter Fox. Bayessim: Adaptive domain randomization via probabilistic inference for robotics simulators. Robotics: Science and Systems XV, Jun 2019. 2 ", + "bbox": [ + 173, + 296, + 825, + 325 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Benjamin Recht, Rebecca Roelofs, Ludwig Schmidt, and Vaishaal Shankar. Do cifar-10 classifiers generalize to cifar-10? arXiv preprint arXiv:1806.00451, 2018. 2 ", + "bbox": [ + 174, + 333, + 823, + 363 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Benjamin Recht, Rebecca Roelofs, Ludwig Schmidt, and Vaishaal Shankar. Do imagenet classifiers generalize to imagenet? arXiv preprint arXiv:1902.10811, 2019. 2 ", + "bbox": [ + 173, + 371, + 823, + 401 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Andrei A Rusu, Neil C Rabinowitz, Guillaume Desjardins, Hubert Soyer, James Kirkpatrick, Koray Kavukcuoglu, Razvan Pascanu, and Raia Hadsell. Progressive neural networks. arXiv preprint arXiv:1606.04671, 2016. 1 ", + "bbox": [ + 173, + 409, + 825, + 452 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Fereshteh Sadeghi and Sergey Levine. Cad2rl: Real single-image flight without a single real image. arXiv preprint arXiv:1611.04201, 2016. 1, 2 ", + "bbox": [ + 173, + 462, + 823, + 489 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Ramanan Sekar, Oleh Rybkin, Kostas Daniilidis, Pieter Abbeel, Danijar Hafner, and Deepak Pathak. Planning to explore via self-supervised world models, 2020. 2 ", + "bbox": [ + 173, + 498, + 823, + 529 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Vaishaal Shankar, Achal Dave, Rebecca Roelofs, Deva Ramanan, Benjamin Recht, and Ludwig Schmidt. A systematic framework for natural perturbations from videos. arXiv preprint arXiv:1906.02168, 2019. 2 ", + "bbox": [ + 173, + 536, + 825, + 579 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Assaf Shocher, Nadav Cohen, and Michal Irani. Zero-shot super-resolution using deep internal learning, 2017. 3 ", + "bbox": [ + 171, + 588, + 825, + 617 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Assaf Shocher, Shai Bagon, Phillip Isola, and Michal Irani. Ingan: Capturing and remapping the ”dna” of a natural image, 2018. 3 ", + "bbox": [ + 174, + 626, + 825, + 656 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Aravind Srinivas, Michael Laskin, and Pieter Abbeel. Curl: Contrastive unsupervised representations for reinforcement learning. arXiv preprint arXiv:2004.04136, 2020. 2, 4, 5, 6, 7, 17 ", + "bbox": [ + 173, + 664, + 825, + 694 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Yu Sun, Eric Tzeng, Trevor Darrell, and Alexei A Efros. Unsupervised domain adaptation through self-supervision. arXiv preprint, 2019. 2 ", + "bbox": [ + 173, + 702, + 823, + 731 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Yu Sun, Xiaolong Wang, Zhuang Liu, John Miller, Alexei A. Efros, and Moritz Hardt. Test-time training with self-supervision for generalization under distribution shifts. ICML, 2020. 3 ", + "bbox": [ + 173, + 739, + 823, + 768 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. Going deeper with convolutions. 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1–9, 2015. 17 ", + "bbox": [ + 178, + 777, + 826, + 821 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Yuval Tassa, Yotam Doron, Alistair Muldal, Tom Erez, Yazhe Li, Diego de Las Casas, David Budden, Abbas Abdolmaleki, Josh Merel, Andrew Lefrancq, Timothy Lillicrap, and Martin Riedmiller. DeepMind control suite. Technical report, DeepMind, January 2018. 2, 4, 5, 16, 17 ", + "bbox": [ + 179, + 829, + 826, + 872 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Josh Tobin, Rachel Fong, Alex Ray, Jonas Schneider, Wojciech Zaremba, and Pieter Abbeel. Domain randomization for transferring deep neural networks from simulation to the real world. 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Sep 2017. 1, 2 ", + "bbox": [ + 174, + 882, + 826, + 924 + ], + "page_idx": 11 + }, + { + "type": "text", + "text": "Eric Tzeng, Judy Hoffman, Kate Saenko, and Trevor Darrell. Adversarial discriminative domain adaptation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 7167–7176, 2017. 2 ", + "bbox": [ + 176, + 103, + 825, + 146 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Pascal Vincent, Hugo Larochelle, Yoshua Bengio, and Pierre-Antoine Manzagol. Extracting and composing robust features with denoising autoencoders. In Proceedings of the 25th international conference on Machine learning, pp. 1096–1103. ACM, 2008. 2 ", + "bbox": [ + 176, + 155, + 821, + 196 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Xiaolong Wang and Abhinav Gupta. Unsupervised learning of visual representations using videos. In ICCV, 2015. 2 ", + "bbox": [ + 173, + 207, + 823, + 234 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Xiaolong Wang, Allan Jabri, and Alexei A. Efros. Learning correspondence from the cycleconsistency of time. In CVPR, 2019. 2 ", + "bbox": [ + 171, + 244, + 825, + 272 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Mitchell Wortsman, Kiana Ehsani, Mohammad Rastegari, Ali Farhadi, and Roozbeh Mottaghi. Learning to learn how to learn: Self-adaptive visual navigation using meta-learning, 2018. 3 ", + "bbox": [ + 173, + 281, + 823, + 310 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Zhirong Wu, Yuanjun Xiong, Stella X Yu, and Dahua Lin. Unsupervised feature learning via nonparametric instance discrimination. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3733–3742, 2018. 2 ", + "bbox": [ + 173, + 319, + 825, + 363 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Marek Wydmuch, Michał Kempka, and Wojciech Jaskowski. Vizdoom competitions: Playing doom ´ from pixels. IEEE Transactions on Games, 2018. 2, 7, 16 ", + "bbox": [ + 171, + 371, + 823, + 400 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Wilson Yan, Ashwin Vangipuram, Pieter Abbeel, and Lerrel Pinto. Learning predictive representations for deformable objects using contrastive estimation. arXiv preprint arXiv:2003.05436, 2020. 1, 2 ", + "bbox": [ + 173, + 409, + 823, + 438 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Jiachen Yang, Brenden Petersen, Hongyuan Zha, and Daniel Faissol. Single episode policy transfer in reinforcement learning, 2019. 1, 2 ", + "bbox": [ + 173, + 446, + 825, + 476 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Denis Yarats, Amy Zhang, Ilya Kostrikov, Brandon Amos, Joelle Pineau, and Rob Fergus. Improving sample efficiency in model-free reinforcement learning from images, 2019. 2, 4, 5, 17 ", + "bbox": [ + 173, + 484, + 823, + 513 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Richard Zhang, Phillip Isola, and Alexei A Efros. Colorful image colorization. In European conference on computer vision, pp. 649–666. Springer, 2016. 2 ", + "bbox": [ + 173, + 522, + 823, + 551 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "Richard Zhang, Phillip Isola, and Alexei A Efros. Split-brain autoencoders: Unsupervised learning by cross-channel prediction. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1058–1067, 2017. 2 ", + "bbox": [ + 174, + 560, + 825, + 603 + ], + "page_idx": 12 + }, + { + "type": "text", + "text": "A PERFORMANCE ON THE TRAINING ENVIRONMENT ", + "text_level": 1, + "bbox": [ + 176, + 103, + 624, + 117 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "Historically, agents have commonly been trained and evaluated in the same environment when benchmarking RL algorithms exclusively in simulation. Although such an evaluation procedure does not consider generalization, it is still a useful metric for comparison of sample efficiency and stability of algorithms. For completeness, we also evaluate our method and baselines in this setting on both DMControl and CRLMaze. DMControl results are reported in Table 7 and results on the CRLMaze environment are shown in Table 8. In this setting, we also compare to an additional baseline on DMControl: a blind SAC agent that operates purely on its previous actions. The performance of a blind agent indicates to which degree a given task benefits from visual information. We find that, while PAD improves generalization to novel environments, performance is virtually unchanged when evaluated on the same environment as in training. We conjecture that this is because the algorithm already is adapted to the training environment and any continued training on the same data distribution thus has little influence. We further emphasize that, even when evaluated on the training environment, PAD still outperforms baselines on most tasks. For example, we observe a $15 \\%$ relative improvement over SAC on the Finger, spin task. We hypothesize that this gain in performance is because the selfsupervised objective improves learning by constraining the intermediate representation of policies. A blind agent is no better than random on this particular task, which would suggest that agents benefit substantially from visual information in Finger, spin. Therefore, learning a good intermediate representation of that information is highly beneficial to the RL objective, which we find PAD to facilitate through its self-supervised learning framework. Likewise, the SAC baseline only achieves a $51 \\%$ improvement over the blind agent on Cartpole, balance, which indicates that extracting visual information from observations is not as crucial on this task. Consequently, both PAD and baselines achieve similar performance on this task. ", + "bbox": [ + 173, + 135, + 826, + 440 + ], + "page_idx": 13 + }, + { + "type": "table", + "img_path": "images/734c0aa38d45cf7a103d5f2800883fa0730f324456bec03a3f8898db190f4406.jpg", + "table_caption": [ + "Table 7. Episodic return on the training environment for each of the 9 tasks considered in DMControl, mean and std. dev. for 10 seeds. Best method on each task is in bold and blue compares $+ \\mathrm { I D M }$ with and without PAD. It is shown that PAD hurts minimally when the environment is unchanged. " + ], + "table_footnote": [], + "table_body": "
Training env.BlindSAC+DR+IDM+IDM (PAD)
Walker, walk235±17847±71756±71911±24895±28
Walker, stand388±10959±11928±36966±8956±20
Cartpole, swingup132±41850±28807±36849±30845±34
Cartpole,balance646±131978±22971±30982±20979±21
Ball in cup, catch150±96725±355469±339919±118910±129
Finger, spin3±2809±138686±295928±45927±45
Finger, turn_easy172±27462±146243±124462±152455±160
Cheetah, run264±75387±74195±46384±88380±91
Reacher, easy107±11264±11392±45390±126365±114
", + "bbox": [ + 253, + 511, + 740, + 650 + ], + "page_idx": 13 + }, + { + "type": "table", + "img_path": "images/19c0693ab395a83584be5ba43fa0acda3ed1d09f22e138b75faf8392ab844913.jpg", + "table_caption": [ + "Table 8. Episodic return of PAD and baselines in the CRLMaze training environment. All methods use A2C. We report mean and std. error of 10 seeds. Best method is in bold and blue compares rotation prediction with and without PAD. " + ], + "table_footnote": [], + "table_body": "
CRLMazeRandomA2C+DR+IDM+IDM (PAD)+Rot+Rot (PAD)
Training env.-868±34371±198-355±93585±246-416±135729±148681±99
", + "bbox": [ + 173, + 726, + 823, + 767 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "B LEARNING CURVES ON DEEPMIND CONTROL ", + "text_level": 1, + "bbox": [ + 174, + 792, + 586, + 809 + ], + "page_idx": 13 + }, + { + "type": "text", + "text": "All methods are trained until convergence (500,000 frames) on DMControl. While we do not consider the sample efficiency of our method and baselines in this study, we report learning curves for SAC, $\\mathrm { S A C + I D M }$ and SAC trained with domain randomization on three tasks in Figure 5 for completeness. SAC trained with and without an IDM are similar in terms of sample efficiency and final performance, whereas domain randomization consistently displays worse sample efficiency, larger variation between seeds, and converges to sub-optimal performance in two out of the three tasks shown. ", + "bbox": [ + 173, + 825, + 825, + 924 + ], + "page_idx": 13 + }, + { + "type": "image", + "img_path": "images/9988caa8b9a55fc3afee7ef0f724429fd455ed6f365e57a9ccd860a250478099.jpg", + "image_caption": [ + "Figure 5. Learning curves for SAC, SAC trained with domain randomization (denoted $S A C ( D R )$ here), and $\\mathrm { S A C + I D M }$ on three tasks from the DeepMind Control suite (DMControl). Episodic return is averaged across 10 seeds and the $9 5 \\%$ confidence intervals are visualized as shaded regions. SAC and $\\mathrm { S A C + I D M }$ exhibit similar sample efficiency and final performance, whereas domain randomization consistently displays worse sample efficiency, larger variation between seeds, and converges to sub-optimal performance in two out of the three tasks shown. " + ], + "image_footnote": [], + "bbox": [ + 194, + 102, + 797, + 257 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "C KEEPING $\\pi _ { s }$ FIXED DURING POLICY ADAPTATION ", + "text_level": 1, + "bbox": [ + 174, + 371, + 612, + 386 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "We now consider a variant of PAD where the self-supervised task head $\\pi _ { s }$ is fixed at test-time such that the self-supervised objective $L$ is optimized only wrt $\\pi _ { e }$ , as discussed in Section 3.3. We measure generalization to test environments with randomized colors and report the results in Table 9 for three tasks from the DeepMind Control suite. We empirically find the difference between updating $\\pi _ { s }$ and keeping it fixed negligible, and we choose to update $\\pi _ { s }$ by default since its gradients are computed by back-propagation regardless. ", + "bbox": [ + 173, + 396, + 825, + 479 + ], + "page_idx": 14 + }, + { + "type": "table", + "img_path": "images/c266f3001c177fdb22a13cff9448f751770c90dd9138aea3b9fd7faa826e8279.jpg", + "table_caption": [ + "Table 9. Episodic return in test environments with randomized colors, mean and std. dev. for 10 seeds. All methods use SAC. IDM (PAD, fixed $\\pi _ { s . }$ ) considers a variant of PAD where $\\pi _ { s }$ is fixed at test-time, whereas $I D M \\left( P A D \\right)$ denotes the default usage of PAD in which both $\\pi _ { e }$ and $\\pi _ { s }$ are optimized at test-time using the self-supervised objective. " + ], + "table_footnote": [], + "table_body": "
Random colorsIDMIDM (PAD, fixed π s)IDM (PAD)
Walker, walk406±29452±38468±47
Walker, stand743±37802±41797±46
Cartpole, swingup585±73623±57630±63
", + "bbox": [ + 290, + 555, + 699, + 621 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "D COMPARISON TO ADAPTATION WITH REWARDS ", + "text_level": 1, + "bbox": [ + 173, + 643, + 599, + 659 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "While our method does not require data collected prior to deployment and does not assume access to a reward signal, we additionally compare our method to a na¨ıve fine-tuning approach using transitions and rewards collected from the target environment prior to deployment. To fine-tune the pre-trained policy using rewards, we collect datasets consisting of 1, 10, and 100 episodes in each target environment using the learned policy while keeping its parameters fixed, and then subsequently fine-tune both $\\pi _ { e }$ and $\\pi _ { a }$ on the collected data, following the same training procedure as during the training phase. This fine-tuning approach is analogous to Julian et al. (2020) but does not use data from the original environment during adaptation. Results are shown in Table 10. We find that na¨ıvely fine-tuning the policy using data collected prior to deployment can improve generalization but requires comparably more data than PAD, as well as access to a reward signal in the target environment. This finding suggests that PAD may be a more suitable method for settings where data from the target environment is scarce and not easily accessible prior to deployment. ", + "bbox": [ + 173, + 667, + 825, + 835 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "E ADDITIONAL ROBOTIC MANIPULATION SAMPLES ", + "text_level": 1, + "bbox": [ + 174, + 857, + 612, + 871 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "Figure 6 provides samples from the training and test environments for the reach robotic manipulation task. Agents are trained in simulation and deployed on a real robot. Samples from the push task are shown in Figure 4. ", + "bbox": [ + 174, + 882, + 823, + 924 + ], + "page_idx": 14 + }, + { + "type": "text", + "text": "Table 10. Episodic return in test environments with randomized colors, mean and std. dev. for 10 seeds. All methods use SAC trained with an inverse dynamics model (IDM) as auxiliary task. Our method is denoted IDM (PAD), and we compare to a na¨ıve fine-tuning approach that assumes access to transitions and rewards collected from 1, 10, and 100 episodes, respectively, from target environments prior to deployment. ", + "bbox": [ + 173, + 101, + 826, + 170 + ], + "page_idx": 15 + }, + { + "type": "table", + "img_path": "images/0957234fffc0806fae66ea15a6933a2dcfc79f987f199fd31739d280c420b83f.jpg", + "table_caption": [ + "Fine-tuning w/ rewards " + ], + "table_footnote": [], + "table_body": "
Random colorsIDMIDM (PAD)1 episode10 episodes100 episodes
Walker, walk406±29468±47395±78489±104561±62
Walker, stand743±37797±46661±65728±44784±31
Cartpole, swingup585±73630±63538±53605±51650±58
", + "bbox": [ + 235, + 191, + 754, + 256 + ], + "page_idx": 15 + }, + { + "type": "image", + "img_path": "images/bb8251dc060f2862ca990a7216bf7fd39204605009caf6d0e9e8d640e1c1e5a1.jpg", + "image_caption": [ + "Figure 6. Samples from the reach robotic manipulation task. The task is to move the robot gripper to the location of the red disc. Agents are trained in setting (a) and evaluated in settings (b-d) on a real robot, taking observations from an uncalibrated camera. " + ], + "image_footnote": [], + "bbox": [ + 178, + 275, + 820, + 416 + ], + "page_idx": 15 + }, + { + "type": "text", + "text": "F IMPLEMENTATION DETAILS ", + "text_level": 1, + "bbox": [ + 176, + 493, + 436, + 510 + ], + "page_idx": 15 + }, + { + "type": "text", + "text": "In this section, we elaborate on implementation details for our experiments on DeepMind Control (DMControl) suite (Tassa et al., 2018) and CRLMaze (Lomonaco et al., 2019) for ViZDoom (Wydmuch et al., 2018). Our implementation for the robotic manipulation experiments closely follows that of DMControl. Code is available at https://nicklashansen.github.io/PAD/. ", + "bbox": [ + 173, + 526, + 826, + 583 + ], + "page_idx": 15 + }, + { + "type": "image", + "img_path": "images/9e0be3b5d460fc2992f725ba0dd7adde3d5e70bc06c843e908bf8d4d2266f9a5.jpg", + "image_caption": [ + "Figure 7. Network architecture for the DMControl, CRLMaze, and robotic manipulation experiments. $\\pi ^ { s }$ and $\\pi ^ { a }$ uses a shared feature extractor $\\pi ^ { e }$ . Observations are stacks of $1 0 0 \\times 1 0 0$ colored frames. Implementation of policy and value function depends on the learning algorithm. " + ], + "image_footnote": [], + "bbox": [ + 269, + 597, + 728, + 729 + ], + "page_idx": 15 + }, + { + "type": "text", + "text": "Architecture. Our network architecture is illustrated in Figure 7. Observations are stacked frames $k = 3 ,$ ) rendered at $1 0 0 \\times 1 0 0$ and cropped to $8 4 \\times 8 4$ , i.e. inputs to the network are of dimensions $9 \\times 8 4 \\times 8 4$ , where the first dimension indicates the channel numbers and the following ones represent spatial dimensions. The same crop is applied to all frames in a stack. The shared feature extractor $\\pi ^ { e }$ consists of 8 (DMControl, robotic manipulation) or 6 (CRLMaze) convolutional layers and outputs features of size $3 2 \\times 2 1 \\times 2 1$ in DMControl and robotic manipulation, and size $3 2 \\times 2 5 \\times 2 5$ in CRLMaze. The output from $\\pi ^ { e }$ is used as input to both the self-supervised head $\\pi ^ { s }$ and RL head $\\pi ^ { a }$ , both of which consist of 3 convolutional layers followed by 3 fully-connected layers. All convolutional layers use 32 filters and all fully connected layers use a hidden size of 1024, as in Yarats et al. (2019). ", + "bbox": [ + 173, + 811, + 825, + 924 + ], + "page_idx": 15 + }, + { + "type": "table", + "img_path": "images/11ba990557b719d83352d6a55c4c0449cee2dca489f14b541e4d7634ccc4b4a3.jpg", + "table_caption": [ + "Table 11. Hyperparameters used for the DMControl (Tassa et al., 2018) tasks. " + ], + "table_footnote": [], + "table_body": "
HyperparameterValue
Frame rendering3 ×100×100
Frame after crop3×84×84
Stacked frames3
Action repeat2 (finger)
8 (cartpole) 4(otherwise)
Discount factor y0.99
Episode length1,000
Learning algorithmSoft Actor-Critic
Self-supervised taskInverse Dynamics Model
Number of training steps500,000
Replay buffer size500,000
Optimizer(πe,πä,π)Adam (β=0.9,β=0.999)
Optimizer (α)Adam(β=0.5,β=0.999)
Learning rate (πe,π,π$)3e-4 (cheetah)
Learning rate (α)le-3 (otherwise) 1e-4
Batch size128
Batch size (test-time)32
πe,π update freq.2
πe,π update freq. (test-time)1
", + "bbox": [ + 173, + 140, + 480, + 378 + ], + "page_idx": 16 + }, + { + "type": "table", + "img_path": "images/04c0e5e01a96f837d220697dc779ce3637100e060e73c95d6ad07084f86bd618.jpg", + "table_caption": [ + "Table 12. Hyperparameters used for the CRLMaze (Lomonaco et al., 2019) navigation task. " + ], + "table_footnote": [], + "table_body": "
HyperparameterValue
Frame rendering3×100×100
Frame after crop3×84×84
Stacked frames3
Action repeat4
Discount factor y0.99
Episode length1,000
Learning algorithmAdvantage Actor-Critic
Self-supervised taskRotation Prediction
Number of training episodes1,000 (dom. rand.) 500 (otherwise)
Number of processes20
OptimizerAdam (β=0.9,β=0.999)
Learning rate1e-4
Learning rate (test-time)1e-5
Batch size20
32
Batch size (test-time) π,πloss coefficient0.5
1
πe,πloss coefficient (test-time)1
πe,π update freq. πe,π update freq.(test-time)1
", + "bbox": [ + 493, + 141, + 815, + 368 + ], + "page_idx": 16 + }, + { + "type": "text", + "text": "", + "bbox": [ + 174, + 409, + 823, + 438 + ], + "page_idx": 16 + }, + { + "type": "text", + "text": "Learning algorithm. We use Soft Actor-Critic (SAC) (Haarnoja et al., 2018) for DMControl and robotic manipulation, and Advantage Actor-Critic (A2C) for CRLMaze. Network outputs depend on the task and learning algorithm. As the action spaces of both DMControl and robotic manipulation are continuous, the policy learned by SAC outputs the mean and variance of a Gaussian distribution over actions. CRLMaze has a discrete action space and the policy learned by A2C thus learns a soft-max distribution over actions. For details on the critics learned by SAC and A2C, the reader is referred to Haarnoja et al. (2018) and Mnih et al. (2016), respectively. ", + "bbox": [ + 174, + 458, + 825, + 555 + ], + "page_idx": 16 + }, + { + "type": "text", + "text": "Hyperparameters. When applicable, we adopt our hyperparameters from Yarats et al. (2019) (DMControl, robotic manipulation) and Lomonaco et al. (2019) (CRLMaze). For the robotic manipulation experiments, our implementation closely follows that of DMControl, only differing by number of frames in an observation. We use a frame stack of $k = 3$ frames for DMControl and CRLMaze, and only $k = 1$ frame for robotic manipulation. For completeness, we detail all hyperparameters used for the DMControl and CRLMaze environments in Table 11 and Table 12. ", + "bbox": [ + 174, + 577, + 825, + 660 + ], + "page_idx": 16 + }, + { + "type": "text", + "text": "Data augmentation. Random cropping is a commonly used data augmentation used in computer vision systems (Krizhevsky et al., 2012; Szegedy et al., 2015) but has only recently gained interest as a stochastic regularization technique in the RL literature (Srinivas et al., 2020; Kostrikov et al., 2020; Laskin et al., 2020). We adopt the random crop proposed in Srinivas et al. (2020): crop rendered observations of size $1 0 0 \\times 1 0 0$ to $8 4 \\times 8 4$ , applying the same crop to all frames in a stacked observation. This has the added benefits of regularization while still preserving spatio-temporal patterns between frames. When learning an inverse dynamics model, we apply the same crop to all frames of a given observation but apply two different crops to the consecutive observations $\\left( \\mathbf { s } _ { t } , \\mathbf { s } _ { t + 1 } \\right)$ used to predict action $\\mathbf { a } _ { t }$ . ", + "bbox": [ + 174, + 680, + 825, + 806 + ], + "page_idx": 16 + }, + { + "type": "text", + "text": "Policy Adaptation during Deployment. We evaluate our method and baselines by episodic return of an agent trained in a single environment and tested in a collection of test environments, each with distinct changes from the training environment. We assume no reward signal at test-time and agents are expected to generalize without pre-training or resetting in the new environment. Therefore, we make updates to the policy using a self-supervised objective, and we train using observations from the environment in an online manner without memory, i.e. we make one update per step using the most-recent observation. ", + "bbox": [ + 174, + 825, + 825, + 922 + ], + "page_idx": 16 + }, + { + "type": "text", + "text": "Empirically, we find that: (i) the random crop data augmentation used during training helps regularize learning at test-time; and (ii) our algorithm benefits from learning from a batch of randomly cropped observations rather than single observations, even when all observations in the batch are augmented copies of the most-recent observation. As such, we apply both of these techniques when performing Policy Adaptation during Deployment and use a batch size of 32. When using the policy to take actions, however, inputs to the policy are simply center-cropped. ", + "bbox": [ + 174, + 103, + 825, + 188 + ], + "page_idx": 17 + } +] \ No newline at end of file diff --git a/parse/train/o_V-MjyyGV_/o_V-MjyyGV__model.json b/parse/train/o_V-MjyyGV_/o_V-MjyyGV__model.json new file mode 100644 index 0000000000000000000000000000000000000000..38505a993ba2864bb480d1f24be6b5af07c5492c --- /dev/null +++ b/parse/train/o_V-MjyyGV_/o_V-MjyyGV__model.json @@ -0,0 +1,20919 @@ +[ + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1387, + 1404, + 1387, + 1404, + 1782, + 298, + 1782 + ], + "score": 0.984 + }, + { + "category_id": 1, + "poly": [ + 398, + 642, + 1305, + 642, + 1305, + 1069, + 398, + 1069 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 298, + 1186, + 1405, + 1186, + 1405, + 1370, + 298, + 1370 + ], + "score": 0.979 + }, + { + "category_id": 1, + "poly": [ + 298, + 1798, + 1405, + 1798, + 1405, + 1983, + 298, + 1983 + ], + "score": 0.976 + }, + { + "category_id": 1, + "poly": [ + 315, + 374, + 1074, + 374, + 1074, + 503, + 315, + 503 + ], + "score": 0.971 + }, + { + "category_id": 0, + "poly": [ + 299, + 219, + 1129, + 219, + 1129, + 323, + 299, + 323 + ], + "score": 0.958 + }, + { + "category_id": 0, + "poly": [ + 302, + 1128, + 573, + 1128, + 573, + 1162, + 302, + 1162 + ], + "score": 0.899 + }, + { + "category_id": 2, + "poly": [ + 328, + 2006, + 1194, + 2006, + 1194, + 2033, + 328, + 2033 + ], + "score": 0.898 + }, + { + "category_id": 2, + "poly": [ + 300, + 75, + 813, + 75, + 813, + 104, + 300, + 104 + ], + "score": 0.881 + }, + { + "category_id": 0, + "poly": [ + 773, + 584, + 927, + 584, + 927, + 617, + 773, + 617 + ], + "score": 0.871 + }, + { + "category_id": 2, + "poly": [ + 841, + 2089, + 856, + 2089, + 856, + 2112, + 841, + 2112 + ], + "score": 0.75 + }, + { + "category_id": 13, + "poly": [ + 314, + 470, + 367, + 470, + 367, + 502, + 314, + 502 + ], + "score": 0.3, + "latex": "{ } ^ { 3 } \\mathrm { I R I }" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 217.0, + 1134.0, + 217.0, + 1134.0, + 272.0, + 297.0, + 272.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 277.0, + 766.0, + 277.0, + 766.0, + 327.0, + 295.0, + 327.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1124.0, + 579.0, + 1124.0, + 579.0, + 1170.0, + 294.0, + 1170.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 332.0, + 2000.0, + 1197.0, + 2000.0, + 1197.0, + 2037.0, + 332.0, + 2037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 72.0, + 815.0, + 72.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 768.0, + 580.0, + 934.0, + 580.0, + 934.0, + 623.0, + 768.0, + 623.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 840.0, + 2088.0, + 859.0, + 2088.0, + 859.0, + 2118.0, + 840.0, + 2118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1385.0, + 1404.0, + 1385.0, + 1404.0, + 1420.0, + 295.0, + 1420.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1413.0, + 1406.0, + 1413.0, + 1406.0, + 1450.0, + 292.0, + 1450.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1444.0, + 1408.0, + 1444.0, + 1408.0, + 1481.0, + 293.0, + 1481.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1472.0, + 1406.0, + 1472.0, + 1406.0, + 1515.0, + 293.0, + 1515.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1508.0, + 1405.0, + 1508.0, + 1405.0, + 1541.0, + 296.0, + 1541.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1539.0, + 1405.0, + 1539.0, + 1405.0, + 1572.0, + 296.0, + 1572.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1569.0, + 1405.0, + 1569.0, + 1405.0, + 1601.0, + 294.0, + 1601.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1597.0, + 1406.0, + 1597.0, + 1406.0, + 1635.0, + 291.0, + 1635.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1631.0, + 1405.0, + 1631.0, + 1405.0, + 1663.0, + 296.0, + 1663.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1661.0, + 1404.0, + 1661.0, + 1404.0, + 1693.0, + 296.0, + 1693.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1691.0, + 1402.0, + 1691.0, + 1402.0, + 1724.0, + 296.0, + 1724.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1722.0, + 1405.0, + 1722.0, + 1405.0, + 1755.0, + 296.0, + 1755.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1751.0, + 985.0, + 1751.0, + 985.0, + 1787.0, + 293.0, + 1787.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 642.0, + 1308.0, + 642.0, + 1308.0, + 676.0, + 393.0, + 676.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 395.0, + 674.0, + 1309.0, + 674.0, + 1309.0, + 706.0, + 395.0, + 706.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 705.0, + 1306.0, + 705.0, + 1306.0, + 734.0, + 394.0, + 734.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 734.0, + 1308.0, + 734.0, + 1308.0, + 769.0, + 393.0, + 769.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 764.0, + 1306.0, + 764.0, + 1306.0, + 798.0, + 393.0, + 798.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 797.0, + 1307.0, + 797.0, + 1307.0, + 829.0, + 394.0, + 829.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 827.0, + 1306.0, + 827.0, + 1306.0, + 862.0, + 394.0, + 862.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 858.0, + 1305.0, + 858.0, + 1305.0, + 889.0, + 394.0, + 889.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 392.0, + 887.0, + 1306.0, + 887.0, + 1306.0, + 920.0, + 392.0, + 920.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 916.0, + 1306.0, + 916.0, + 1306.0, + 950.0, + 394.0, + 950.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 947.0, + 1309.0, + 947.0, + 1309.0, + 981.0, + 393.0, + 981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 978.0, + 1309.0, + 978.0, + 1309.0, + 1013.0, + 394.0, + 1013.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 393.0, + 1008.0, + 1304.0, + 1008.0, + 1304.0, + 1042.0, + 393.0, + 1042.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 394.0, + 1041.0, + 913.0, + 1041.0, + 913.0, + 1073.0, + 394.0, + 1073.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1187.0, + 1406.0, + 1187.0, + 1406.0, + 1218.0, + 296.0, + 1218.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1215.0, + 1406.0, + 1215.0, + 1406.0, + 1250.0, + 295.0, + 1250.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1244.0, + 1405.0, + 1244.0, + 1405.0, + 1283.0, + 293.0, + 1283.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1278.0, + 1405.0, + 1278.0, + 1405.0, + 1314.0, + 295.0, + 1314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1310.0, + 1405.0, + 1310.0, + 1405.0, + 1341.0, + 296.0, + 1341.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1339.0, + 1155.0, + 1339.0, + 1155.0, + 1373.0, + 295.0, + 1373.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1799.0, + 1405.0, + 1799.0, + 1405.0, + 1831.0, + 296.0, + 1831.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1831.0, + 1406.0, + 1831.0, + 1406.0, + 1863.0, + 296.0, + 1863.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1860.0, + 1403.0, + 1860.0, + 1403.0, + 1892.0, + 296.0, + 1892.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1888.0, + 1407.0, + 1888.0, + 1407.0, + 1922.0, + 293.0, + 1922.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1917.0, + 1409.0, + 1917.0, + 1409.0, + 1955.0, + 292.0, + 1955.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1948.0, + 1370.0, + 1948.0, + 1370.0, + 1986.0, + 295.0, + 1986.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 311.0, + 373.0, + 1077.0, + 373.0, + 1077.0, + 412.0, + 311.0, + 412.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 310.0, + 403.0, + 1065.0, + 403.0, + 1065.0, + 445.0, + 310.0, + 445.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 309.0, + 432.0, + 900.0, + 432.0, + 900.0, + 480.0, + 309.0, + 480.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 310.0, + 465.0, + 313.0, + 465.0, + 313.0, + 509.0, + 310.0, + 509.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 368.0, + 465.0, + 782.0, + 465.0, + 782.0, + 509.0, + 368.0, + 509.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 0, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 551, + 1403, + 551, + 1403, + 857, + 298, + 857 + ], + "score": 0.985 + }, + { + "category_id": 1, + "poly": [ + 298, + 934, + 1406, + 934, + 1406, + 1453, + 298, + 1453 + ], + "score": 0.984 + }, + { + "category_id": 1, + "poly": [ + 298, + 1468, + 1404, + 1468, + 1404, + 1744, + 298, + 1744 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 298, + 230, + 1405, + 230, + 1405, + 536, + 298, + 536 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 298, + 1759, + 1404, + 1759, + 1404, + 2034, + 298, + 2034 + ], + "score": 0.982 + }, + { + "category_id": 0, + "poly": [ + 300, + 883, + 588, + 883, + 588, + 918, + 300, + 918 + ], + "score": 0.893 + }, + { + "category_id": 2, + "poly": [ + 300, + 76, + 812, + 76, + 812, + 104, + 300, + 104 + ], + "score": 0.891 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 858, + 2088, + 858, + 2112, + 841, + 2112 + ], + "score": 0.705 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 859, + 2088, + 859, + 2112, + 841, + 2112 + ], + "score": 0.148 + }, + { + "category_id": 15, + "poly": [ + 292.0, + 880.0, + 593.0, + 880.0, + 593.0, + 924.0, + 292.0, + 924.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2085.0, + 862.0, + 2085.0, + 862.0, + 2120.0, + 839.0, + 2120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2085.0, + 862.0, + 2085.0, + 862.0, + 2120.0, + 838.0, + 2120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 549.0, + 1405.0, + 549.0, + 1405.0, + 586.0, + 294.0, + 586.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 583.0, + 1404.0, + 583.0, + 1404.0, + 615.0, + 294.0, + 615.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 609.0, + 1405.0, + 609.0, + 1405.0, + 647.0, + 293.0, + 647.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 644.0, + 1404.0, + 644.0, + 1404.0, + 676.0, + 296.0, + 676.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 672.0, + 1407.0, + 672.0, + 1407.0, + 709.0, + 292.0, + 709.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 704.0, + 1404.0, + 704.0, + 1404.0, + 736.0, + 296.0, + 736.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 732.0, + 1405.0, + 732.0, + 1405.0, + 770.0, + 293.0, + 770.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 760.0, + 1407.0, + 760.0, + 1407.0, + 803.0, + 292.0, + 803.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 793.0, + 1405.0, + 793.0, + 1405.0, + 830.0, + 294.0, + 830.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 826.0, + 1250.0, + 826.0, + 1250.0, + 860.0, + 293.0, + 860.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 934.0, + 1408.0, + 934.0, + 1408.0, + 967.0, + 296.0, + 967.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 965.0, + 1406.0, + 965.0, + 1406.0, + 997.0, + 296.0, + 997.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 992.0, + 1406.0, + 992.0, + 1406.0, + 1030.0, + 295.0, + 1030.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1022.0, + 1407.0, + 1022.0, + 1407.0, + 1063.0, + 292.0, + 1063.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1057.0, + 1404.0, + 1057.0, + 1404.0, + 1089.0, + 296.0, + 1089.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1085.0, + 1406.0, + 1085.0, + 1406.0, + 1117.0, + 296.0, + 1117.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1116.0, + 1408.0, + 1116.0, + 1408.0, + 1151.0, + 293.0, + 1151.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1144.0, + 1406.0, + 1144.0, + 1406.0, + 1182.0, + 293.0, + 1182.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1178.0, + 1406.0, + 1178.0, + 1406.0, + 1210.0, + 297.0, + 1210.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1205.0, + 1404.0, + 1205.0, + 1404.0, + 1245.0, + 293.0, + 1245.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1240.0, + 1404.0, + 1240.0, + 1404.0, + 1272.0, + 296.0, + 1272.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1270.0, + 1408.0, + 1270.0, + 1408.0, + 1302.0, + 293.0, + 1302.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1298.0, + 1406.0, + 1298.0, + 1406.0, + 1333.0, + 292.0, + 1333.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1329.0, + 1408.0, + 1329.0, + 1408.0, + 1365.0, + 295.0, + 1365.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1357.0, + 1406.0, + 1357.0, + 1406.0, + 1395.0, + 293.0, + 1395.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1392.0, + 1404.0, + 1392.0, + 1404.0, + 1424.0, + 296.0, + 1424.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1419.0, + 933.0, + 1419.0, + 933.0, + 1458.0, + 293.0, + 1458.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1468.0, + 1404.0, + 1468.0, + 1404.0, + 1500.0, + 297.0, + 1500.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1499.0, + 1406.0, + 1499.0, + 1406.0, + 1532.0, + 296.0, + 1532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1526.0, + 1406.0, + 1526.0, + 1406.0, + 1563.0, + 296.0, + 1563.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1559.0, + 1405.0, + 1559.0, + 1405.0, + 1594.0, + 293.0, + 1594.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1588.0, + 1404.0, + 1588.0, + 1404.0, + 1626.0, + 292.0, + 1626.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1620.0, + 1402.0, + 1620.0, + 1402.0, + 1653.0, + 296.0, + 1653.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1652.0, + 1404.0, + 1652.0, + 1404.0, + 1684.0, + 296.0, + 1684.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1679.0, + 1405.0, + 1679.0, + 1405.0, + 1717.0, + 293.0, + 1717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1710.0, + 982.0, + 1710.0, + 982.0, + 1747.0, + 295.0, + 1747.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 229.0, + 1405.0, + 229.0, + 1405.0, + 265.0, + 293.0, + 265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 261.0, + 1405.0, + 261.0, + 1405.0, + 297.0, + 293.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 288.0, + 1406.0, + 288.0, + 1406.0, + 330.0, + 292.0, + 330.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 322.0, + 1406.0, + 322.0, + 1406.0, + 358.0, + 295.0, + 358.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 350.0, + 1408.0, + 350.0, + 1408.0, + 389.0, + 292.0, + 389.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 383.0, + 1407.0, + 383.0, + 1407.0, + 419.0, + 295.0, + 419.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 411.0, + 1408.0, + 411.0, + 1408.0, + 449.0, + 292.0, + 449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 443.0, + 1407.0, + 443.0, + 1407.0, + 479.0, + 293.0, + 479.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 469.0, + 1406.0, + 469.0, + 1406.0, + 514.0, + 292.0, + 514.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 506.0, + 1015.0, + 506.0, + 1015.0, + 538.0, + 296.0, + 538.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1756.0, + 1406.0, + 1756.0, + 1406.0, + 1792.0, + 294.0, + 1792.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1788.0, + 1404.0, + 1788.0, + 1404.0, + 1821.0, + 296.0, + 1821.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1817.0, + 1404.0, + 1817.0, + 1404.0, + 1855.0, + 293.0, + 1855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1848.0, + 1405.0, + 1848.0, + 1405.0, + 1885.0, + 294.0, + 1885.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1880.0, + 1405.0, + 1880.0, + 1405.0, + 1915.0, + 294.0, + 1915.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1907.0, + 1406.0, + 1907.0, + 1406.0, + 1949.0, + 293.0, + 1949.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1942.0, + 1405.0, + 1942.0, + 1405.0, + 1975.0, + 294.0, + 1975.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1971.0, + 1406.0, + 1971.0, + 1406.0, + 2007.0, + 294.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 2003.0, + 1402.0, + 2003.0, + 1402.0, + 2036.0, + 296.0, + 2036.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 1, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1378, + 1405, + 1378, + 1405, + 1686, + 297, + 1686 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 297, + 755, + 1405, + 755, + 1405, + 1093, + 297, + 1093 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 299, + 1173, + 1405, + 1173, + 1405, + 1296, + 299, + 1296 + ], + "score": 0.98 + }, + { + "category_id": 1, + "poly": [ + 297, + 1767, + 1403, + 1767, + 1403, + 1922, + 297, + 1922 + ], + "score": 0.978 + }, + { + "category_id": 3, + "poly": [ + 307, + 197, + 1386, + 197, + 1386, + 509, + 307, + 509 + ], + "score": 0.967 + }, + { + "category_id": 4, + "poly": [ + 297, + 526, + 1405, + 526, + 1405, + 650, + 297, + 650 + ], + "score": 0.951 + }, + { + "category_id": 1, + "poly": [ + 296, + 1972, + 1413, + 1972, + 1413, + 2035, + 296, + 2035 + ], + "score": 0.946 + }, + { + "category_id": 1, + "poly": [ + 297, + 678, + 1400, + 678, + 1400, + 740, + 297, + 740 + ], + "score": 0.943 + }, + { + "category_id": 8, + "poly": [ + 612, + 1928, + 1086, + 1928, + 1086, + 1967, + 612, + 1967 + ], + "score": 0.938 + }, + { + "category_id": 0, + "poly": [ + 300, + 1720, + 1108, + 1720, + 1108, + 1753, + 300, + 1753 + ], + "score": 0.908 + }, + { + "category_id": 0, + "poly": [ + 300, + 1332, + 687, + 1332, + 687, + 1363, + 300, + 1363 + ], + "score": 0.907 + }, + { + "category_id": 0, + "poly": [ + 299, + 1124, + 481, + 1124, + 481, + 1159, + 299, + 1159 + ], + "score": 0.899 + }, + { + "category_id": 2, + "poly": [ + 299, + 76, + 812, + 76, + 812, + 104, + 299, + 104 + ], + "score": 0.886 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1933, + 1399, + 1933, + 1399, + 1962, + 1366, + 1962 + ], + "score": 0.882 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 858, + 2088, + 858, + 2112, + 841, + 2112 + ], + "score": 0.651 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 859, + 2088, + 859, + 2112, + 841, + 2112 + ], + "score": 0.431 + }, + { + "category_id": 13, + "poly": [ + 403, + 1501, + 630, + 1501, + 630, + 1535, + 403, + 1535 + ], + "score": 0.93, + "latex": "\\pi ( \\mathbf { s } ; \\theta ) = \\pi _ { a } ( \\pi _ { e } ( \\mathbf { s } ) )" + }, + { + "category_id": 14, + "poly": [ + 609, + 1926, + 1090, + 1926, + 1090, + 1968, + 609, + 1968 + ], + "score": 0.93, + "latex": "L ( \\theta _ { s } , \\theta _ { e } ) = \\ell \\big ( \\mathbf { a } _ { t } , \\pi _ { s } ( \\pi _ { e } ( \\mathbf { s } _ { t } ) , \\pi _ { e } ( \\mathbf { s } _ { t + 1 } ) ) \\big ) ." + }, + { + "category_id": 13, + "poly": [ + 298, + 1440, + 435, + 1440, + 435, + 1474, + 298, + 1474 + ], + "score": 0.92, + "latex": "\\theta = \\left( \\theta _ { e } , \\theta _ { a } \\right)" + }, + { + "category_id": 13, + "poly": [ + 1095, + 1234, + 1150, + 1234, + 1150, + 1268, + 1095, + 1268 + ], + "score": 0.92, + "latex": "J ( \\theta )" + }, + { + "category_id": 13, + "poly": [ + 1114, + 1769, + 1258, + 1769, + 1258, + 1802, + 1114, + 1802 + ], + "score": 0.91, + "latex": "\\left( \\mathbf { s } _ { t } , \\mathbf { a } _ { t } , \\mathbf { s } _ { t + 1 } \\right)" + }, + { + "category_id": 13, + "poly": [ + 740, + 1594, + 769, + 1594, + 769, + 1623, + 740, + 1623 + ], + "score": 0.89, + "latex": "\\theta _ { s }" + }, + { + "category_id": 13, + "poly": [ + 1113, + 1441, + 1143, + 1441, + 1143, + 1471, + 1113, + 1471 + ], + "score": 0.88, + "latex": "\\theta _ { a }" + }, + { + "category_id": 13, + "poly": [ + 519, + 1441, + 548, + 1441, + 548, + 1471, + 519, + 1471 + ], + "score": 0.88, + "latex": "\\theta _ { e }" + }, + { + "category_id": 13, + "poly": [ + 1064, + 1472, + 1093, + 1472, + 1093, + 1501, + 1064, + 1501 + ], + "score": 0.88, + "latex": "\\theta _ { e }" + }, + { + "category_id": 13, + "poly": [ + 1371, + 1472, + 1400, + 1472, + 1400, + 1501, + 1371, + 1501 + ], + "score": 0.87, + "latex": "\\theta _ { a }" + }, + { + "category_id": 13, + "poly": [ + 619, + 1537, + 651, + 1537, + 651, + 1562, + 619, + 1562 + ], + "score": 0.87, + "latex": "\\pi _ { a }" + }, + { + "category_id": 13, + "poly": [ + 1145, + 1474, + 1178, + 1474, + 1178, + 1501, + 1145, + 1501 + ], + "score": 0.87, + "latex": "\\pi _ { a }" + }, + { + "category_id": 13, + "poly": [ + 1122, + 1566, + 1153, + 1566, + 1153, + 1593, + 1122, + 1593 + ], + "score": 0.86, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 649, + 1627, + 680, + 1627, + 680, + 1654, + 649, + 1654 + ], + "score": 0.86, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 518, + 1596, + 550, + 1596, + 550, + 1623, + 518, + 1623 + ], + "score": 0.86, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 869, + 1626, + 900, + 1626, + 900, + 1654, + 869, + 1654 + ], + "score": 0.86, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 840, + 1475, + 871, + 1475, + 871, + 1501, + 840, + 1501 + ], + "score": 0.86, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 1110, + 1863, + 1141, + 1863, + 1141, + 1890, + 1110, + 1890 + ], + "score": 0.86, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 376, + 1567, + 407, + 1567, + 407, + 1593, + 376, + 1593 + ], + "score": 0.86, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 508, + 1894, + 538, + 1894, + 538, + 1922, + 508, + 1922 + ], + "score": 0.86, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 298, + 1537, + 328, + 1537, + 328, + 1563, + 298, + 1563 + ], + "score": 0.83, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 759, + 1411, + 776, + 1411, + 776, + 1437, + 759, + 1437 + ], + "score": 0.79, + "latex": "\\theta" + }, + { + "category_id": 13, + "poly": [ + 565, + 1974, + 581, + 1974, + 581, + 2000, + 565, + 2000 + ], + "score": 0.79, + "latex": "\\ell" + }, + { + "category_id": 13, + "poly": [ + 1362, + 1566, + 1393, + 1566, + 1393, + 1593, + 1362, + 1593 + ], + "score": 0.79, + "latex": "\\pi _ { \\theta }" + }, + { + "category_id": 13, + "poly": [ + 426, + 1266, + 443, + 1266, + 443, + 1292, + 426, + 1292 + ], + "score": 0.79, + "latex": "\\theta" + }, + { + "category_id": 13, + "poly": [ + 1247, + 2004, + 1263, + 2004, + 1263, + 2030, + 1247, + 2030 + ], + "score": 0.78, + "latex": "\\ell" + }, + { + "category_id": 13, + "poly": [ + 1068, + 1415, + 1088, + 1415, + 1088, + 1438, + 1068, + 1438 + ], + "score": 0.75, + "latex": "\\pi" + }, + { + "category_id": 15, + "poly": [ + 527.0, + 194.0, + 610.0, + 194.0, + 610.0, + 233.0, + 527.0, + 233.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 962.0, + 197.0, + 1287.0, + 197.0, + 1287.0, + 228.0, + 962.0, + 228.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 246.0, + 433.0, + 246.0, + 433.0, + 274.0, + 322.0, + 274.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 883.0, + 247.0, + 988.0, + 247.0, + 988.0, + 272.0, + 883.0, + 272.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 661.0, + 289.0, + 786.0, + 289.0, + 786.0, + 313.0, + 661.0, + 313.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1217.0, + 286.0, + 1345.0, + 286.0, + 1345.0, + 314.0, + 1217.0, + 314.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 684.0, + 304.0, + 764.0, + 304.0, + 764.0, + 338.0, + 684.0, + 338.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1188.0, + 305.0, + 1198.0, + 305.0, + 1198.0, + 316.0, + 1188.0, + 316.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1243.0, + 304.0, + 1321.0, + 304.0, + 1321.0, + 338.0, + 1243.0, + 338.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 663.0, + 394.0, + 785.0, + 394.0, + 785.0, + 427.0, + 663.0, + 427.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1220.0, + 395.0, + 1343.0, + 395.0, + 1343.0, + 427.0, + 1220.0, + 427.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 451.0, + 416.0, + 462.0, + 416.0, + 462.0, + 425.0, + 451.0, + 425.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 684.0, + 414.0, + 764.0, + 414.0, + 764.0, + 449.0, + 684.0, + 449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1103.0, + 415.0, + 1129.0, + 415.0, + 1129.0, + 428.0, + 1103.0, + 428.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1131.0, + 414.0, + 1152.0, + 414.0, + 1152.0, + 435.0, + 1131.0, + 435.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1188.0, + 416.0, + 1199.0, + 416.0, + 1199.0, + 426.0, + 1188.0, + 426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1242.0, + 414.0, + 1322.0, + 414.0, + 1322.0, + 449.0, + 1242.0, + 449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1097.0, + 425.0, + 1103.0, + 425.0, + 1103.0, + 432.0, + 1097.0, + 432.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 329.0, + 471.0, + 428.0, + 471.0, + 428.0, + 496.0, + 329.0, + 496.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 887.0, + 469.0, + 986.0, + 469.0, + 986.0, + 499.0, + 887.0, + 499.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 553.75, + 294.5, + 593.75, + 294.5, + 593.75, + 326.5, + 553.75, + 326.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1112.75, + 292.5, + 1149.75, + 292.5, + 1149.75, + 326.0, + 1112.75, + 326.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 465.25, + 404.0, + 505.25, + 404.0, + 505.25, + 437.0, + 465.25, + 437.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 555.0, + 404.0, + 596.0, + 404.0, + 596.0, + 437.0, + 555.0, + 437.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1023.0, + 404.0, + 1063.0, + 404.0, + 1063.0, + 437.0, + 1023.0, + 437.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 527.0, + 1406.0, + 527.0, + 1406.0, + 560.0, + 294.0, + 560.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 556.0, + 1406.0, + 556.0, + 1406.0, + 593.0, + 295.0, + 593.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 589.0, + 1405.0, + 589.0, + 1405.0, + 622.0, + 295.0, + 622.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 618.0, + 1354.0, + 618.0, + 1354.0, + 653.0, + 294.0, + 653.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1719.0, + 1112.0, + 1719.0, + 1112.0, + 1757.0, + 294.0, + 1757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1332.0, + 690.0, + 1332.0, + 690.0, + 1365.0, + 296.0, + 1365.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 290.0, + 1118.0, + 487.0, + 1118.0, + 487.0, + 1168.0, + 290.0, + 1168.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2085.0, + 860.0, + 2085.0, + 860.0, + 2116.0, + 839.0, + 2116.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2085.0, + 862.0, + 2085.0, + 862.0, + 2118.0, + 838.0, + 2118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1376.0, + 1405.0, + 1376.0, + 1405.0, + 1413.0, + 294.0, + 1413.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1407.0, + 758.0, + 1407.0, + 758.0, + 1445.0, + 292.0, + 1445.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 777.0, + 1407.0, + 1067.0, + 1407.0, + 1067.0, + 1445.0, + 777.0, + 1445.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1089.0, + 1407.0, + 1405.0, + 1407.0, + 1405.0, + 1445.0, + 1089.0, + 1445.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1436.0, + 297.0, + 1436.0, + 297.0, + 1478.0, + 291.0, + 1478.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 436.0, + 1436.0, + 518.0, + 1436.0, + 518.0, + 1478.0, + 436.0, + 1478.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 549.0, + 1436.0, + 1112.0, + 1436.0, + 1112.0, + 1478.0, + 549.0, + 1478.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1144.0, + 1436.0, + 1406.0, + 1436.0, + 1406.0, + 1478.0, + 1144.0, + 1478.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1468.0, + 839.0, + 1468.0, + 839.0, + 1507.0, + 292.0, + 1507.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 872.0, + 1468.0, + 1063.0, + 1468.0, + 1063.0, + 1507.0, + 872.0, + 1507.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1094.0, + 1468.0, + 1144.0, + 1468.0, + 1144.0, + 1507.0, + 1094.0, + 1507.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1179.0, + 1468.0, + 1370.0, + 1468.0, + 1370.0, + 1507.0, + 1179.0, + 1507.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1401.0, + 1468.0, + 1405.0, + 1468.0, + 1405.0, + 1507.0, + 1401.0, + 1507.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1500.0, + 402.0, + 1500.0, + 402.0, + 1536.0, + 292.0, + 1536.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 631.0, + 1500.0, + 1407.0, + 1500.0, + 1407.0, + 1536.0, + 631.0, + 1536.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1532.0, + 297.0, + 1532.0, + 297.0, + 1566.0, + 294.0, + 1566.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 329.0, + 1532.0, + 618.0, + 1532.0, + 618.0, + 1566.0, + 329.0, + 1566.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 652.0, + 1532.0, + 1407.0, + 1532.0, + 1407.0, + 1566.0, + 652.0, + 1566.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1563.0, + 375.0, + 1563.0, + 375.0, + 1598.0, + 296.0, + 1598.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 408.0, + 1563.0, + 1121.0, + 1563.0, + 1121.0, + 1598.0, + 408.0, + 1598.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1154.0, + 1563.0, + 1361.0, + 1563.0, + 1361.0, + 1598.0, + 1154.0, + 1598.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1394.0, + 1563.0, + 1407.0, + 1563.0, + 1407.0, + 1598.0, + 1394.0, + 1598.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1593.0, + 517.0, + 1593.0, + 517.0, + 1625.0, + 294.0, + 1625.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 551.0, + 1593.0, + 739.0, + 1593.0, + 739.0, + 1625.0, + 551.0, + 1625.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 770.0, + 1593.0, + 1407.0, + 1593.0, + 1407.0, + 1625.0, + 770.0, + 1625.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1623.0, + 648.0, + 1623.0, + 648.0, + 1658.0, + 295.0, + 1658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 681.0, + 1623.0, + 868.0, + 1623.0, + 868.0, + 1658.0, + 681.0, + 1658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 901.0, + 1623.0, + 1406.0, + 1623.0, + 1406.0, + 1658.0, + 901.0, + 1658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1653.0, + 1408.0, + 1653.0, + 1408.0, + 1689.0, + 295.0, + 1689.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 752.0, + 1406.0, + 752.0, + 1406.0, + 788.0, + 294.0, + 788.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 785.0, + 1405.0, + 785.0, + 1405.0, + 819.0, + 295.0, + 819.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 816.0, + 1403.0, + 816.0, + 1403.0, + 851.0, + 295.0, + 851.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 847.0, + 1405.0, + 847.0, + 1405.0, + 882.0, + 294.0, + 882.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 873.0, + 1406.0, + 873.0, + 1406.0, + 914.0, + 292.0, + 914.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 909.0, + 1405.0, + 909.0, + 1405.0, + 940.0, + 295.0, + 940.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 934.0, + 1406.0, + 934.0, + 1406.0, + 976.0, + 292.0, + 976.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 970.0, + 1403.0, + 970.0, + 1403.0, + 1001.0, + 295.0, + 1001.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 997.0, + 1406.0, + 997.0, + 1406.0, + 1035.0, + 291.0, + 1035.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1031.0, + 1407.0, + 1031.0, + 1407.0, + 1062.0, + 296.0, + 1062.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1062.0, + 1322.0, + 1062.0, + 1322.0, + 1093.0, + 296.0, + 1093.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1173.0, + 1407.0, + 1173.0, + 1407.0, + 1209.0, + 293.0, + 1209.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1203.0, + 1406.0, + 1203.0, + 1406.0, + 1239.0, + 293.0, + 1239.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1236.0, + 1094.0, + 1236.0, + 1094.0, + 1268.0, + 295.0, + 1268.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1151.0, + 1236.0, + 1405.0, + 1236.0, + 1405.0, + 1268.0, + 1151.0, + 1268.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1267.0, + 425.0, + 1267.0, + 425.0, + 1298.0, + 294.0, + 1298.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 444.0, + 1267.0, + 822.0, + 1267.0, + 822.0, + 1298.0, + 444.0, + 1298.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1765.0, + 1113.0, + 1765.0, + 1113.0, + 1805.0, + 292.0, + 1805.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1259.0, + 1765.0, + 1406.0, + 1765.0, + 1406.0, + 1805.0, + 1259.0, + 1805.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1797.0, + 1405.0, + 1797.0, + 1405.0, + 1834.0, + 295.0, + 1834.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1829.0, + 1405.0, + 1829.0, + 1405.0, + 1863.0, + 293.0, + 1863.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1859.0, + 1109.0, + 1859.0, + 1109.0, + 1895.0, + 293.0, + 1895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1142.0, + 1859.0, + 1405.0, + 1859.0, + 1405.0, + 1895.0, + 1142.0, + 1895.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1888.0, + 507.0, + 1888.0, + 507.0, + 1928.0, + 292.0, + 1928.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 539.0, + 1888.0, + 1303.0, + 1888.0, + 1303.0, + 1928.0, + 539.0, + 1928.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1972.0, + 564.0, + 1972.0, + 564.0, + 2007.0, + 295.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 582.0, + 1972.0, + 1408.0, + 1972.0, + 1408.0, + 2007.0, + 582.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 2001.0, + 1246.0, + 2001.0, + 1246.0, + 2038.0, + 293.0, + 2038.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1264.0, + 2001.0, + 1410.0, + 2001.0, + 1410.0, + 2038.0, + 1264.0, + 2038.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 677.0, + 1404.0, + 677.0, + 1404.0, + 713.0, + 296.0, + 713.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 708.0, + 1262.0, + 708.0, + 1262.0, + 743.0, + 295.0, + 743.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 2, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1630, + 1405, + 1630, + 1405, + 1968, + 298, + 1968 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 297, + 665, + 1404, + 665, + 1404, + 940, + 297, + 940 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 297, + 1280, + 1405, + 1280, + 1405, + 1616, + 297, + 1616 + ], + "score": 0.982 + }, + { + "category_id": 1, + "poly": [ + 299, + 229, + 1403, + 229, + 1403, + 415, + 299, + 415 + ], + "score": 0.978 + }, + { + "category_id": 1, + "poly": [ + 299, + 429, + 1404, + 429, + 1404, + 582, + 299, + 582 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 298, + 1106, + 1403, + 1106, + 1403, + 1169, + 298, + 1169 + ], + "score": 0.945 + }, + { + "category_id": 8, + "poly": [ + 594, + 951, + 843, + 951, + 843, + 984, + 594, + 984 + ], + "score": 0.915 + }, + { + "category_id": 0, + "poly": [ + 300, + 618, + 660, + 618, + 660, + 650, + 300, + 650 + ], + "score": 0.914 + }, + { + "category_id": 0, + "poly": [ + 300, + 1233, + 558, + 1233, + 558, + 1267, + 300, + 1267 + ], + "score": 0.912 + }, + { + "category_id": 8, + "poly": [ + 559, + 989, + 1137, + 989, + 1137, + 1022, + 559, + 1022 + ], + "score": 0.91 + }, + { + "category_id": 8, + "poly": [ + 565, + 1027, + 1132, + 1027, + 1132, + 1061, + 565, + 1061 + ], + "score": 0.909 + }, + { + "category_id": 2, + "poly": [ + 300, + 76, + 812, + 76, + 812, + 104, + 300, + 104 + ], + "score": 0.896 + }, + { + "category_id": 8, + "poly": [ + 587, + 1065, + 1076, + 1065, + 1076, + 1100, + 587, + 1100 + ], + "score": 0.866 + }, + { + "category_id": 9, + "poly": [ + 1366, + 952, + 1400, + 952, + 1400, + 981, + 1366, + 981 + ], + "score": 0.864 + }, + { + "category_id": 2, + "poly": [ + 297, + 1978, + 1407, + 1978, + 1407, + 2034, + 297, + 2034 + ], + "score": 0.864 + }, + { + "category_id": 9, + "poly": [ + 1367, + 1030, + 1400, + 1030, + 1400, + 1058, + 1367, + 1058 + ], + "score": 0.855 + }, + { + "category_id": 9, + "poly": [ + 1366, + 991, + 1400, + 991, + 1400, + 1020, + 1366, + 1020 + ], + "score": 0.85 + }, + { + "category_id": 9, + "poly": [ + 1366, + 1069, + 1399, + 1069, + 1399, + 1097, + 1366, + 1097 + ], + "score": 0.844 + }, + { + "category_id": 2, + "poly": [ + 840, + 2089, + 857, + 2089, + 857, + 2111, + 840, + 2111 + ], + "score": 0.781 + }, + { + "category_id": 14, + "poly": [ + 556, + 946, + 1140, + 946, + 1140, + 1103, + 556, + 1103 + ], + "score": 0.92, + "latex": "\\begin{array} { r l } & { \\quad \\mathbf s _ { t } \\sim p ( \\mathbf s _ { t } | \\mathbf a _ { t - 1 } , \\mathbf s _ { t - 1 } ) } \\\\ & { \\quad \\theta _ { s } ( t ) = \\theta _ { s } ( t - 1 ) - \\nabla _ { \\theta _ { s } } L ( \\mathbf s _ { t } ; \\theta _ { s } ( t - 1 ) , \\theta _ { e } ( t - 1 ) ) } \\\\ & { \\quad \\theta _ { e } ( t ) = \\theta _ { e } ( t - 1 ) - \\nabla _ { \\theta _ { e } } L ( \\mathbf s _ { t } ; \\theta _ { s } ( t - 1 ) , \\theta _ { e } ( t - 1 ) ) } \\\\ & { \\quad \\mathbf a _ { t } = \\pi ( \\mathbf s _ { t } ; \\theta ( t ) ) \\mathrm { ~ w i t h ~ } \\theta ( t ) = ( \\theta _ { e } ( t ) , \\theta _ { a } ) , } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 893, + 1876, + 1013, + 1876, + 1013, + 1905, + 893, + 1905 + ], + "score": 0.91, + "latex": "1 0 0 \\times 1 0 0" + }, + { + "category_id": 13, + "poly": [ + 829, + 727, + 1218, + 727, + 1218, + 761, + 829, + 761 + ], + "score": 0.9, + "latex": "\\begin{array} { r } { \\operatorname* { m i n } _ { \\theta _ { a } , \\theta _ { s } , \\theta _ { e } } J ( \\theta _ { a } , \\theta _ { e } ) + \\alpha L ( \\theta _ { s } , \\theta _ { e } ) } \\end{array}" + }, + { + "category_id": 13, + "poly": [ + 471, + 1876, + 540, + 1876, + 540, + 1904, + 471, + 1904 + ], + "score": 0.9, + "latex": "k = 1" + }, + { + "category_id": 13, + "poly": [ + 433, + 1937, + 498, + 1937, + 498, + 1968, + 433, + 1968 + ], + "score": 0.9, + "latex": "\\theta _ { e } , \\theta _ { s }" + }, + { + "category_id": 13, + "poly": [ + 1304, + 728, + 1375, + 728, + 1375, + 756, + 1304, + 756 + ], + "score": 0.9, + "latex": "\\alpha > 0" + }, + { + "category_id": 13, + "poly": [ + 669, + 910, + 776, + 910, + 776, + 938, + 669, + 938 + ], + "score": 0.89, + "latex": "t = 1 . . . T" + }, + { + "category_id": 13, + "poly": [ + 370, + 1106, + 626, + 1106, + 626, + 1140, + 370, + 1140 + ], + "score": 0.88, + "latex": "\\theta _ { s } ( 0 ) = \\theta _ { s } , \\theta _ { e } ( 0 ) = \\theta _ { e }" + }, + { + "category_id": 13, + "poly": [ + 971, + 790, + 999, + 790, + 999, + 819, + 971, + 819 + ], + "score": 0.88, + "latex": "\\theta _ { s }" + }, + { + "category_id": 13, + "poly": [ + 1055, + 789, + 1083, + 789, + 1083, + 819, + 1055, + 819 + ], + "score": 0.88, + "latex": "\\theta _ { e }" + }, + { + "category_id": 13, + "poly": [ + 1162, + 1845, + 1232, + 1845, + 1232, + 1874, + 1162, + 1874 + ], + "score": 0.88, + "latex": "k = 3" + }, + { + "category_id": 13, + "poly": [ + 1206, + 1106, + 1348, + 1106, + 1348, + 1140, + 1206, + 1140 + ], + "score": 0.87, + "latex": "\\mathbf { a } _ { 0 } = \\pi _ { \\theta } ( \\mathbf { s } _ { 0 } )" + }, + { + "category_id": 13, + "poly": [ + 700, + 1727, + 733, + 1727, + 733, + 1754, + 700, + 1754 + ], + "score": 0.87, + "latex": "\\pi _ { a }" + }, + { + "category_id": 13, + "poly": [ + 687, + 820, + 715, + 820, + 715, + 849, + 687, + 849 + ], + "score": 0.87, + "latex": "\\theta _ { s }" + }, + { + "category_id": 13, + "poly": [ + 1151, + 1696, + 1182, + 1696, + 1182, + 1723, + 1151, + 1723 + ], + "score": 0.87, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 1024, + 1727, + 1055, + 1727, + 1055, + 1754, + 1024, + 1754 + ], + "score": 0.86, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 754, + 1758, + 784, + 1758, + 784, + 1784, + 754, + 1784 + ], + "score": 0.85, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 755, + 1139, + 777, + 1139, + 777, + 1165, + 755, + 1165 + ], + "score": 0.83, + "latex": "L" + }, + { + "category_id": 13, + "poly": [ + 1370, + 1819, + 1400, + 1819, + 1400, + 1845, + 1370, + 1845 + ], + "score": 0.83, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 960, + 1846, + 979, + 1846, + 979, + 1872, + 960, + 1872 + ], + "score": 0.82, + "latex": "k" + }, + { + "category_id": 13, + "poly": [ + 1081, + 761, + 1103, + 761, + 1103, + 786, + 1081, + 786 + ], + "score": 0.81, + "latex": "J" + }, + { + "category_id": 13, + "poly": [ + 638, + 1109, + 667, + 1109, + 667, + 1139, + 638, + 1139 + ], + "score": 0.8, + "latex": "\\mathbf { s } _ { 0 }" + }, + { + "category_id": 13, + "poly": [ + 828, + 881, + 846, + 881, + 846, + 907, + 828, + 907 + ], + "score": 0.79, + "latex": "\\theta" + }, + { + "category_id": 13, + "poly": [ + 767, + 790, + 790, + 790, + 790, + 816, + 767, + 816 + ], + "score": 0.78, + "latex": "L" + }, + { + "category_id": 13, + "poly": [ + 1358, + 1112, + 1376, + 1112, + 1376, + 1139, + 1358, + 1139 + ], + "score": 0.48, + "latex": "p" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 618.0, + 663.0, + 618.0, + 663.0, + 654.0, + 294.0, + 654.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1230.0, + 560.0, + 1230.0, + 560.0, + 1272.0, + 292.0, + 1272.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 329.0, + 1973.0, + 1405.0, + 1973.0, + 1405.0, + 2012.0, + 329.0, + 2012.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 2005.0, + 1405.0, + 2005.0, + 1405.0, + 2036.0, + 294.0, + 2036.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 2086.0, + 862.0, + 2086.0, + 862.0, + 2118.0, + 838.0, + 2118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1628.0, + 1406.0, + 1628.0, + 1406.0, + 1666.0, + 293.0, + 1666.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1661.0, + 1407.0, + 1661.0, + 1407.0, + 1697.0, + 293.0, + 1697.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1693.0, + 1150.0, + 1693.0, + 1150.0, + 1724.0, + 295.0, + 1724.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1183.0, + 1693.0, + 1402.0, + 1693.0, + 1402.0, + 1724.0, + 1183.0, + 1724.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1723.0, + 699.0, + 1723.0, + 699.0, + 1757.0, + 293.0, + 1757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 734.0, + 1723.0, + 1023.0, + 1723.0, + 1023.0, + 1757.0, + 734.0, + 1757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1056.0, + 1723.0, + 1405.0, + 1723.0, + 1405.0, + 1757.0, + 1056.0, + 1757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1753.0, + 753.0, + 1753.0, + 753.0, + 1788.0, + 295.0, + 1788.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 785.0, + 1753.0, + 1405.0, + 1753.0, + 1405.0, + 1788.0, + 785.0, + 1788.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1782.0, + 1403.0, + 1782.0, + 1403.0, + 1816.0, + 295.0, + 1816.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1811.0, + 1369.0, + 1811.0, + 1369.0, + 1851.0, + 293.0, + 1851.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1401.0, + 1811.0, + 1405.0, + 1811.0, + 1405.0, + 1851.0, + 1401.0, + 1851.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1844.0, + 959.0, + 1844.0, + 959.0, + 1877.0, + 293.0, + 1877.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 980.0, + 1844.0, + 1161.0, + 1844.0, + 1161.0, + 1877.0, + 980.0, + 1877.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1233.0, + 1844.0, + 1405.0, + 1844.0, + 1405.0, + 1877.0, + 1233.0, + 1877.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1875.0, + 470.0, + 1875.0, + 470.0, + 1910.0, + 295.0, + 1910.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 541.0, + 1875.0, + 892.0, + 1875.0, + 892.0, + 1910.0, + 541.0, + 1910.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1014.0, + 1875.0, + 1403.0, + 1875.0, + 1403.0, + 1910.0, + 1014.0, + 1910.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1904.0, + 1407.0, + 1904.0, + 1407.0, + 1942.0, + 293.0, + 1942.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1933.0, + 432.0, + 1933.0, + 432.0, + 1973.0, + 293.0, + 1973.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 499.0, + 1933.0, + 1410.0, + 1933.0, + 1410.0, + 1973.0, + 499.0, + 1973.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 666.0, + 1405.0, + 666.0, + 1405.0, + 702.0, + 295.0, + 702.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 697.0, + 1405.0, + 697.0, + 1405.0, + 730.0, + 296.0, + 730.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 724.0, + 828.0, + 724.0, + 828.0, + 769.0, + 292.0, + 769.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1219.0, + 724.0, + 1303.0, + 724.0, + 1303.0, + 769.0, + 1219.0, + 769.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1376.0, + 724.0, + 1409.0, + 724.0, + 1409.0, + 769.0, + 1376.0, + 769.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 757.0, + 1080.0, + 757.0, + 1080.0, + 793.0, + 294.0, + 793.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1104.0, + 757.0, + 1406.0, + 757.0, + 1406.0, + 793.0, + 1104.0, + 793.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 789.0, + 766.0, + 789.0, + 766.0, + 822.0, + 295.0, + 822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 791.0, + 789.0, + 970.0, + 789.0, + 970.0, + 822.0, + 791.0, + 822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1000.0, + 789.0, + 1054.0, + 789.0, + 1054.0, + 822.0, + 1000.0, + 822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1084.0, + 789.0, + 1404.0, + 789.0, + 1404.0, + 822.0, + 1084.0, + 822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 818.0, + 686.0, + 818.0, + 686.0, + 855.0, + 294.0, + 855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 716.0, + 818.0, + 1405.0, + 818.0, + 1405.0, + 855.0, + 716.0, + 855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 851.0, + 1405.0, + 851.0, + 1405.0, + 883.0, + 296.0, + 883.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 879.0, + 827.0, + 879.0, + 827.0, + 916.0, + 295.0, + 916.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 847.0, + 879.0, + 1406.0, + 879.0, + 1406.0, + 916.0, + 847.0, + 916.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 909.0, + 668.0, + 909.0, + 668.0, + 942.0, + 295.0, + 942.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 777.0, + 909.0, + 786.0, + 909.0, + 786.0, + 942.0, + 777.0, + 942.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1276.0, + 1405.0, + 1276.0, + 1405.0, + 1318.0, + 292.0, + 1318.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1311.0, + 1405.0, + 1311.0, + 1405.0, + 1345.0, + 295.0, + 1345.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1341.0, + 1405.0, + 1341.0, + 1405.0, + 1376.0, + 294.0, + 1376.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1370.0, + 1405.0, + 1370.0, + 1405.0, + 1406.0, + 294.0, + 1406.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1402.0, + 1405.0, + 1402.0, + 1405.0, + 1437.0, + 295.0, + 1437.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1432.0, + 1408.0, + 1432.0, + 1408.0, + 1467.0, + 295.0, + 1467.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1464.0, + 1405.0, + 1464.0, + 1405.0, + 1495.0, + 296.0, + 1495.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1493.0, + 1403.0, + 1493.0, + 1403.0, + 1526.0, + 295.0, + 1526.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1522.0, + 1405.0, + 1522.0, + 1405.0, + 1557.0, + 295.0, + 1557.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1555.0, + 1408.0, + 1555.0, + 1408.0, + 1589.0, + 295.0, + 1589.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1584.0, + 1249.0, + 1584.0, + 1249.0, + 1618.0, + 294.0, + 1618.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 230.0, + 1405.0, + 230.0, + 1405.0, + 265.0, + 294.0, + 265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 263.0, + 1405.0, + 263.0, + 1405.0, + 295.0, + 295.0, + 295.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 293.0, + 1404.0, + 293.0, + 1404.0, + 325.0, + 295.0, + 325.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 323.0, + 1404.0, + 323.0, + 1404.0, + 355.0, + 295.0, + 355.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 351.0, + 1405.0, + 351.0, + 1405.0, + 387.0, + 294.0, + 387.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 385.0, + 1227.0, + 385.0, + 1227.0, + 416.0, + 295.0, + 416.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 428.0, + 1406.0, + 428.0, + 1406.0, + 465.0, + 295.0, + 465.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 457.0, + 1404.0, + 457.0, + 1404.0, + 496.0, + 293.0, + 496.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 491.0, + 1404.0, + 491.0, + 1404.0, + 524.0, + 297.0, + 524.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 519.0, + 1410.0, + 519.0, + 1410.0, + 558.0, + 293.0, + 558.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 552.0, + 1362.0, + 552.0, + 1362.0, + 585.0, + 294.0, + 585.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1102.0, + 369.0, + 1102.0, + 369.0, + 1144.0, + 293.0, + 1144.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 627.0, + 1102.0, + 637.0, + 1102.0, + 637.0, + 1144.0, + 627.0, + 1144.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 668.0, + 1102.0, + 1205.0, + 1102.0, + 1205.0, + 1144.0, + 668.0, + 1144.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1349.0, + 1102.0, + 1357.0, + 1102.0, + 1357.0, + 1144.0, + 1349.0, + 1144.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1377.0, + 1102.0, + 1407.0, + 1102.0, + 1407.0, + 1144.0, + 1377.0, + 1144.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1135.0, + 754.0, + 1135.0, + 754.0, + 1172.0, + 293.0, + 1172.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 778.0, + 1135.0, + 1408.0, + 1135.0, + 1408.0, + 1172.0, + 778.0, + 1172.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 3, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 5, + "poly": [ + 353, + 684, + 1341, + 684, + 1341, + 990, + 353, + 990 + ], + "score": 0.98, + "html": "
Random colorsSAC+DR+IDM+IDM (PAD)+IDM+IDM (PAD)
Walker, walk414±74594±104406±29468±473830±5475505±592
Walker, stand719±74715±96743±37797±467832±2098566±121
Cartpole, swingup592±50647±48585±73630±636528±5397093±592
Cartpole,balance857±60867±37835±40848±297746±5267670±293
Ball in cup,catch411±183470±252471±75563±50
Finger, spin626±163465±314757±62803±727249±6427496±655
Finger, turn_easy270±43167±26283±51304±461
Cheetah, run154±41145±29121±38159±281117±5301208±487
Reacher, easy163±45105±37201±32214±441788±4412152±506
" + }, + { + "category_id": 1, + "poly": [ + 299, + 1226, + 993, + 1226, + 993, + 1710, + 299, + 1710 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 297, + 1086, + 1403, + 1086, + 1403, + 1209, + 297, + 1209 + ], + "score": 0.973 + }, + { + "category_id": 1, + "poly": [ + 299, + 1911, + 1404, + 1911, + 1404, + 2034, + 299, + 2034 + ], + "score": 0.971 + }, + { + "category_id": 1, + "poly": [ + 297, + 1713, + 1404, + 1713, + 1404, + 1897, + 297, + 1897 + ], + "score": 0.97 + }, + { + "category_id": 3, + "poly": [ + 1010, + 1240, + 1396, + 1240, + 1396, + 1588, + 1010, + 1588 + ], + "score": 0.96 + }, + { + "category_id": 4, + "poly": [ + 1010, + 1602, + 1403, + 1602, + 1403, + 1692, + 1010, + 1692 + ], + "score": 0.953 + }, + { + "category_id": 3, + "poly": [ + 323, + 224, + 1376, + 224, + 1376, + 462, + 323, + 462 + ], + "score": 0.953 + }, + { + "category_id": 4, + "poly": [ + 301, + 474, + 1400, + 474, + 1400, + 569, + 301, + 569 + ], + "score": 0.944 + }, + { + "category_id": 0, + "poly": [ + 298, + 1034, + 623, + 1034, + 623, + 1066, + 298, + 1066 + ], + "score": 0.901 + }, + { + "category_id": 2, + "poly": [ + 299, + 76, + 812, + 76, + 812, + 105, + 299, + 105 + ], + "score": 0.893 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 858, + 2088, + 858, + 2112, + 841, + 2112 + ], + "score": 0.734 + }, + { + "category_id": 1, + "poly": [ + 299, + 588, + 1403, + 588, + 1403, + 652, + 299, + 652 + ], + "score": 0.614 + }, + { + "category_id": 6, + "poly": [ + 299, + 588, + 1403, + 588, + 1403, + 652, + 299, + 652 + ], + "score": 0.437 + }, + { + "category_id": 6, + "poly": [ + 1099, + 668, + 1286, + 668, + 1286, + 695, + 1099, + 695 + ], + "score": 0.277 + }, + { + "category_id": 13, + "poly": [ + 1091, + 1744, + 1151, + 1744, + 1151, + 1772, + 1091, + 1772 + ], + "score": 0.86, + "latex": "+ D R" + }, + { + "category_id": 13, + "poly": [ + 846, + 1775, + 926, + 1775, + 926, + 1803, + 846, + 1803 + ], + "score": 0.85, + "latex": "+ I D M )" + }, + { + "category_id": 13, + "poly": [ + 426, + 1804, + 575, + 1804, + 575, + 1835, + 426, + 1835 + ], + "score": 0.76, + "latex": "+ I D M \\left( P A D \\right)" + }, + { + "category_id": 13, + "poly": [ + 987, + 620, + 1114, + 620, + 1114, + 649, + 987, + 649 + ], + "score": 0.64, + "latex": "\\mathrm { S A C + I D M }" + }, + { + "category_id": 13, + "poly": [ + 1235, + 1713, + 1288, + 1713, + 1288, + 1742, + 1235, + 1742 + ], + "score": 0.41, + "latex": "S A C" + }, + { + "category_id": 15, + "poly": [ + 1033.0, + 1266.0, + 1074.0, + 1266.0, + 1074.0, + 1290.0, + 1033.0, + 1290.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1008.0, + 1295.0, + 1037.0, + 1295.0, + 1037.0, + 1492.0, + 1008.0, + 1492.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1035.0, + 1320.0, + 1072.0, + 1320.0, + 1072.0, + 1342.0, + 1035.0, + 1342.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1041.0, + 1373.0, + 1073.0, + 1373.0, + 1073.0, + 1395.0, + 1041.0, + 1395.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1041.0, + 1426.0, + 1073.0, + 1426.0, + 1073.0, + 1448.0, + 1041.0, + 1448.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1253.0, + 1457.0, + 1346.0, + 1457.0, + 1346.0, + 1479.0, + 1253.0, + 1479.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1037.0, + 1478.0, + 1074.0, + 1478.0, + 1074.0, + 1503.0, + 1037.0, + 1503.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1253.0, + 1475.0, + 1352.0, + 1475.0, + 1352.0, + 1497.0, + 1253.0, + 1497.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1252.0, + 1493.0, + 1378.0, + 1493.0, + 1378.0, + 1518.0, + 1252.0, + 1518.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1252.0, + 1513.0, + 1376.0, + 1513.0, + 1376.0, + 1536.0, + 1252.0, + 1536.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1030.0, + 1531.0, + 1074.0, + 1531.0, + 1074.0, + 1555.0, + 1030.0, + 1555.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1070.0, + 1549.0, + 1087.0, + 1549.0, + 1087.0, + 1568.0, + 1070.0, + 1568.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1121.0, + 1546.0, + 1159.0, + 1546.0, + 1159.0, + 1571.0, + 1121.0, + 1571.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1182.0, + 1546.0, + 1220.0, + 1546.0, + 1220.0, + 1571.0, + 1182.0, + 1571.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1242.0, + 1546.0, + 1281.0, + 1546.0, + 1281.0, + 1571.0, + 1242.0, + 1571.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1304.0, + 1546.0, + 1343.0, + 1546.0, + 1343.0, + 1571.0, + 1304.0, + 1571.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1362.0, + 1546.0, + 1403.0, + 1546.0, + 1403.0, + 1571.0, + 1362.0, + 1571.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1171.0, + 1565.0, + 1293.0, + 1565.0, + 1293.0, + 1591.0, + 1171.0, + 1591.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1009.0, + 1598.0, + 1405.0, + 1598.0, + 1405.0, + 1635.0, + 1009.0, + 1635.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1010.0, + 1632.0, + 1405.0, + 1632.0, + 1405.0, + 1663.0, + 1010.0, + 1663.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1010.0, + 1659.0, + 1353.0, + 1659.0, + 1353.0, + 1694.0, + 1010.0, + 1694.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 266.0, + 369.0, + 266.0, + 369.0, + 275.0, + 358.0, + 275.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 494.0, + 266.0, + 505.0, + 266.0, + 505.0, + 275.0, + 494.0, + 275.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 608.0, + 266.0, + 618.0, + 266.0, + 618.0, + 275.0, + 608.0, + 275.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 721.0, + 266.0, + 732.0, + 266.0, + 732.0, + 275.0, + 721.0, + 275.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1176.0, + 266.0, + 1186.0, + 266.0, + 1186.0, + 275.0, + 1176.0, + 275.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1288.0, + 266.0, + 1299.0, + 266.0, + 1299.0, + 275.0, + 1288.0, + 275.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1168.0, + 309.0, + 1187.0, + 309.0, + 1187.0, + 328.0, + 1168.0, + 328.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 340.0, + 414.0, + 340.0, + 414.0, + 390.0, + 327.0, + 390.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 810.0, + 344.0, + 1116.0, + 344.0, + 1116.0, + 392.0, + 810.0, + 392.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1128.0, + 343.0, + 1318.0, + 343.0, + 1318.0, + 385.0, + 1128.0, + 385.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 327.0, + 403.0, + 409.0, + 403.0, + 409.0, + 431.0, + 327.0, + 431.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 460.0, + 405.0, + 544.0, + 405.0, + 544.0, + 430.0, + 460.0, + 430.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 699.0, + 404.0, + 769.0, + 404.0, + 769.0, + 429.0, + 699.0, + 429.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 905.0, + 404.0, + 1005.0, + 404.0, + 1005.0, + 431.0, + 905.0, + 431.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1140.0, + 403.0, + 1234.0, + 403.0, + 1234.0, + 429.0, + 1140.0, + 429.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1256.0, + 406.0, + 1337.0, + 406.0, + 1337.0, + 428.0, + 1256.0, + 428.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 590.0, + 404.5, + 656.0, + 404.5, + 656.0, + 428.0, + 590.0, + 428.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 815.0, + 405.0, + 887.0, + 405.0, + 887.0, + 429.0, + 815.0, + 429.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1022.0, + 401.0, + 1118.0, + 401.0, + 1118.0, + 431.5, + 1022.0, + 431.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 474.0, + 1405.0, + 474.0, + 1405.0, + 512.0, + 295.0, + 512.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 507.0, + 1405.0, + 507.0, + 1405.0, + 541.0, + 296.0, + 541.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 534.0, + 1354.0, + 534.0, + 1354.0, + 575.0, + 295.0, + 575.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1030.0, + 629.0, + 1030.0, + 629.0, + 1072.0, + 293.0, + 1072.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 70.0, + 816.0, + 70.0, + 816.0, + 110.0, + 293.0, + 110.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2085.0, + 861.0, + 2085.0, + 861.0, + 2120.0, + 839.0, + 2120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 588.0, + 1405.0, + 588.0, + 1405.0, + 624.0, + 294.0, + 624.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 620.0, + 986.0, + 620.0, + 986.0, + 652.0, + 295.0, + 652.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1115.0, + 620.0, + 1375.0, + 620.0, + 1375.0, + 652.0, + 1115.0, + 652.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1097.0, + 665.0, + 1286.0, + 665.0, + 1286.0, + 699.0, + 1097.0, + 699.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1226.0, + 994.0, + 1226.0, + 994.0, + 1257.0, + 297.0, + 1257.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1253.0, + 996.0, + 1253.0, + 996.0, + 1290.0, + 295.0, + 1290.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1284.0, + 994.0, + 1284.0, + 994.0, + 1321.0, + 295.0, + 1321.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1316.0, + 995.0, + 1316.0, + 995.0, + 1350.0, + 296.0, + 1350.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1348.0, + 991.0, + 1348.0, + 991.0, + 1380.0, + 297.0, + 1380.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1378.0, + 994.0, + 1378.0, + 994.0, + 1409.0, + 297.0, + 1409.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1408.0, + 991.0, + 1408.0, + 991.0, + 1440.0, + 297.0, + 1440.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1440.0, + 992.0, + 1440.0, + 992.0, + 1471.0, + 296.0, + 1471.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1469.0, + 995.0, + 1469.0, + 995.0, + 1499.0, + 296.0, + 1499.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1500.0, + 993.0, + 1500.0, + 993.0, + 1532.0, + 296.0, + 1532.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1530.0, + 993.0, + 1530.0, + 993.0, + 1561.0, + 297.0, + 1561.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1559.0, + 991.0, + 1559.0, + 991.0, + 1594.0, + 297.0, + 1594.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1590.0, + 994.0, + 1590.0, + 994.0, + 1624.0, + 295.0, + 1624.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1621.0, + 991.0, + 1621.0, + 991.0, + 1653.0, + 295.0, + 1653.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1652.0, + 992.0, + 1652.0, + 992.0, + 1683.0, + 295.0, + 1683.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1682.0, + 994.0, + 1682.0, + 994.0, + 1713.0, + 296.0, + 1713.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1086.0, + 1405.0, + 1086.0, + 1405.0, + 1118.0, + 296.0, + 1118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1115.0, + 1405.0, + 1115.0, + 1405.0, + 1151.0, + 293.0, + 1151.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1148.0, + 1405.0, + 1148.0, + 1405.0, + 1180.0, + 292.0, + 1180.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1177.0, + 390.0, + 1177.0, + 390.0, + 1216.0, + 291.0, + 1216.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1911.0, + 1408.0, + 1911.0, + 1408.0, + 1942.0, + 294.0, + 1942.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1942.0, + 1406.0, + 1942.0, + 1406.0, + 1975.0, + 294.0, + 1975.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1974.0, + 1404.0, + 1974.0, + 1404.0, + 2006.0, + 295.0, + 2006.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 2004.0, + 1404.0, + 2004.0, + 1404.0, + 2036.0, + 295.0, + 2036.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1711.0, + 1234.0, + 1711.0, + 1234.0, + 1747.0, + 295.0, + 1747.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1289.0, + 1711.0, + 1406.0, + 1711.0, + 1406.0, + 1747.0, + 1289.0, + 1747.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1743.0, + 1090.0, + 1743.0, + 1090.0, + 1775.0, + 295.0, + 1775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1152.0, + 1743.0, + 1405.0, + 1743.0, + 1405.0, + 1775.0, + 1152.0, + 1775.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1773.0, + 845.0, + 1773.0, + 845.0, + 1808.0, + 292.0, + 1808.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 927.0, + 1773.0, + 1406.0, + 1773.0, + 1406.0, + 1808.0, + 927.0, + 1808.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1805.0, + 425.0, + 1805.0, + 425.0, + 1837.0, + 295.0, + 1837.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 576.0, + 1805.0, + 1405.0, + 1805.0, + 1405.0, + 1837.0, + 576.0, + 1837.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1834.0, + 1406.0, + 1834.0, + 1406.0, + 1870.0, + 294.0, + 1870.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1864.0, + 452.0, + 1864.0, + 452.0, + 1903.0, + 293.0, + 1903.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 588.0, + 1405.0, + 588.0, + 1405.0, + 624.0, + 294.0, + 624.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 620.0, + 986.0, + 620.0, + 986.0, + 652.0, + 295.0, + 652.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1115.0, + 620.0, + 1375.0, + 620.0, + 1375.0, + 652.0, + 1115.0, + 652.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 4, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1666, + 1404, + 1666, + 1404, + 2032, + 298, + 2032 + ], + "score": 0.983 + }, + { + "category_id": 5, + "poly": [ + 681, + 860, + 1414, + 860, + 1414, + 1352, + 681, + 1352 + ], + "score": 0.981, + "html": "
Video backgroundsSAC+DR+IDM+IDM (PAD)
Walker, walk616±80655±55694±85717±79
Walker, stand899±53869±60902±51935±20
Cartpole, swingup375±90485±67487±90521±76
Cartpole, balance693±109766±92691±76687±58
Ball in cup, catch393±175271±189362±69436±55
Finger, spin447±102338±207605±61691±80
Finger, turn_easy355±108223±91355±110362±101
Cheetah, run194±30150±34164±42206±34
Distracting objectsSAC+DR+IDM+IDM (PAD)
Cartpole, swingup815±60809±24776±58771±64
Cartpole,balance969±20938±35964±26960±29
Ball in cup, catch177±111331±189482±128545±173
Finger, spin652±184564±288836±62867±72
Finger, turn_easy302±68165±12326±101347±48
" + }, + { + "category_id": 1, + "poly": [ + 297, + 308, + 1405, + 308, + 1405, + 704, + 297, + 704 + ], + "score": 0.981 + }, + { + "category_id": 1, + "poly": [ + 298, + 1436, + 1404, + 1436, + 1404, + 1651, + 298, + 1651 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 299, + 721, + 661, + 721, + 661, + 1390, + 299, + 1390 + ], + "score": 0.968 + }, + { + "category_id": 1, + "poly": [ + 295, + 229, + 1401, + 229, + 1401, + 293, + 295, + 293 + ], + "score": 0.951 + }, + { + "category_id": 2, + "poly": [ + 300, + 76, + 813, + 76, + 813, + 104, + 300, + 104 + ], + "score": 0.896 + }, + { + "category_id": 6, + "poly": [ + 681, + 722, + 1402, + 722, + 1402, + 844, + 681, + 844 + ], + "score": 0.883 + }, + { + "category_id": 2, + "poly": [ + 840, + 2089, + 858, + 2089, + 858, + 2112, + 840, + 2112 + ], + "score": 0.779 + }, + { + "category_id": 7, + "poly": [ + 300, + 1391, + 1025, + 1391, + 1025, + 1421, + 300, + 1421 + ], + "score": 0.598 + }, + { + "category_id": 1, + "poly": [ + 300, + 1391, + 1025, + 1391, + 1025, + 1421, + 300, + 1421 + ], + "score": 0.097 + }, + { + "category_id": 13, + "poly": [ + 705, + 1590, + 777, + 1590, + 777, + 1619, + 705, + 1619 + ], + "score": 0.87, + "latex": "20 \\%" + }, + { + "category_id": 13, + "poly": [ + 330, + 994, + 402, + 994, + 402, + 1024, + 330, + 1024 + ], + "score": 0.85, + "latex": "104 \\%" + }, + { + "category_id": 13, + "poly": [ + 795, + 814, + 922, + 814, + 922, + 843, + 795, + 843 + ], + "score": 0.69, + "latex": "\\mathrm { S A C + I D M }" + }, + { + "category_id": 13, + "poly": [ + 1135, + 340, + 1180, + 340, + 1180, + 368, + 1135, + 368 + ], + "score": 0.33, + "latex": "1 0 \\mathrm { x }" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 72.0, + 816.0, + 72.0, + 816.0, + 110.0, + 294.0, + 110.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 679.0, + 721.0, + 1408.0, + 721.0, + 1408.0, + 754.0, + 679.0, + 754.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 678.0, + 751.0, + 1409.0, + 751.0, + 1409.0, + 786.0, + 678.0, + 786.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 679.0, + 781.0, + 1406.0, + 781.0, + 1406.0, + 816.0, + 679.0, + 816.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 679.0, + 814.0, + 794.0, + 814.0, + 794.0, + 845.0, + 679.0, + 845.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 923.0, + 814.0, + 1185.0, + 814.0, + 1185.0, + 845.0, + 923.0, + 845.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 840.0, + 2087.0, + 861.0, + 2087.0, + 861.0, + 2118.0, + 840.0, + 2118.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1388.0, + 1031.0, + 1388.0, + 1031.0, + 1425.0, + 294.0, + 1425.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1667.0, + 1408.0, + 1667.0, + 1408.0, + 1701.0, + 297.0, + 1701.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1694.0, + 1405.0, + 1694.0, + 1405.0, + 1735.0, + 292.0, + 1735.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1728.0, + 1404.0, + 1728.0, + 1404.0, + 1762.0, + 294.0, + 1762.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1759.0, + 1405.0, + 1759.0, + 1405.0, + 1792.0, + 293.0, + 1792.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1787.0, + 1405.0, + 1787.0, + 1405.0, + 1822.0, + 293.0, + 1822.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1819.0, + 1404.0, + 1819.0, + 1404.0, + 1855.0, + 293.0, + 1855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1845.0, + 1406.0, + 1845.0, + 1406.0, + 1888.0, + 292.0, + 1888.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1879.0, + 1406.0, + 1879.0, + 1406.0, + 1913.0, + 294.0, + 1913.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1911.0, + 1406.0, + 1911.0, + 1406.0, + 1945.0, + 294.0, + 1945.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1941.0, + 1406.0, + 1941.0, + 1406.0, + 1975.0, + 294.0, + 1975.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1973.0, + 1407.0, + 1973.0, + 1407.0, + 2003.0, + 293.0, + 2003.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 2003.0, + 1160.0, + 2003.0, + 1160.0, + 2033.0, + 296.0, + 2033.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 310.0, + 1403.0, + 310.0, + 1403.0, + 342.0, + 295.0, + 342.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 339.0, + 1134.0, + 339.0, + 1134.0, + 372.0, + 296.0, + 372.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1181.0, + 339.0, + 1403.0, + 339.0, + 1403.0, + 372.0, + 1181.0, + 372.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 368.0, + 1406.0, + 368.0, + 1406.0, + 404.0, + 294.0, + 404.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 399.0, + 1405.0, + 399.0, + 1405.0, + 435.0, + 294.0, + 435.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 430.0, + 1406.0, + 430.0, + 1406.0, + 466.0, + 295.0, + 466.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 459.0, + 1406.0, + 459.0, + 1406.0, + 497.0, + 292.0, + 497.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 488.0, + 1407.0, + 488.0, + 1407.0, + 528.0, + 294.0, + 528.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 519.0, + 1408.0, + 519.0, + 1408.0, + 558.0, + 291.0, + 558.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 550.0, + 1408.0, + 550.0, + 1408.0, + 590.0, + 294.0, + 590.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 581.0, + 1406.0, + 581.0, + 1406.0, + 617.0, + 292.0, + 617.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 612.0, + 1406.0, + 612.0, + 1406.0, + 648.0, + 295.0, + 648.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 639.0, + 1406.0, + 639.0, + 1406.0, + 680.0, + 292.0, + 680.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 675.0, + 1256.0, + 675.0, + 1256.0, + 708.0, + 296.0, + 708.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1438.0, + 1405.0, + 1438.0, + 1405.0, + 1472.0, + 294.0, + 1472.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1467.0, + 1404.0, + 1467.0, + 1404.0, + 1503.0, + 294.0, + 1503.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1499.0, + 1406.0, + 1499.0, + 1406.0, + 1534.0, + 294.0, + 1534.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1529.0, + 1406.0, + 1529.0, + 1406.0, + 1563.0, + 294.0, + 1563.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1557.0, + 1407.0, + 1557.0, + 1407.0, + 1593.0, + 292.0, + 1593.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1591.0, + 704.0, + 1591.0, + 704.0, + 1622.0, + 296.0, + 1622.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 778.0, + 1591.0, + 1404.0, + 1591.0, + 1404.0, + 1622.0, + 778.0, + 1622.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1617.0, + 1369.0, + 1617.0, + 1369.0, + 1657.0, + 292.0, + 1657.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 720.0, + 666.0, + 720.0, + 666.0, + 753.0, + 295.0, + 753.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 749.0, + 661.0, + 749.0, + 661.0, + 781.0, + 295.0, + 781.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 782.0, + 666.0, + 782.0, + 666.0, + 814.0, + 296.0, + 814.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 812.0, + 666.0, + 812.0, + 666.0, + 844.0, + 292.0, + 844.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 841.0, + 666.0, + 841.0, + 666.0, + 873.0, + 295.0, + 873.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 872.0, + 663.0, + 872.0, + 663.0, + 904.0, + 294.0, + 904.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 902.0, + 665.0, + 902.0, + 665.0, + 935.0, + 294.0, + 935.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 931.0, + 662.0, + 931.0, + 662.0, + 965.0, + 295.0, + 965.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 960.0, + 663.0, + 960.0, + 663.0, + 999.0, + 294.0, + 999.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 994.0, + 329.0, + 994.0, + 329.0, + 1025.0, + 295.0, + 1025.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 403.0, + 994.0, + 666.0, + 994.0, + 666.0, + 1025.0, + 403.0, + 1025.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1023.0, + 663.0, + 1023.0, + 663.0, + 1057.0, + 293.0, + 1057.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1057.0, + 665.0, + 1057.0, + 665.0, + 1088.0, + 294.0, + 1088.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1086.0, + 663.0, + 1086.0, + 663.0, + 1116.0, + 292.0, + 1116.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1118.0, + 661.0, + 1118.0, + 661.0, + 1147.0, + 293.0, + 1147.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1144.0, + 665.0, + 1144.0, + 665.0, + 1181.0, + 294.0, + 1181.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1174.0, + 663.0, + 1174.0, + 663.0, + 1212.0, + 294.0, + 1212.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1208.0, + 665.0, + 1208.0, + 665.0, + 1238.0, + 296.0, + 1238.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1239.0, + 665.0, + 1239.0, + 665.0, + 1270.0, + 295.0, + 1270.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1268.0, + 665.0, + 1268.0, + 665.0, + 1299.0, + 294.0, + 1299.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1298.0, + 666.0, + 1298.0, + 666.0, + 1331.0, + 294.0, + 1331.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1330.0, + 662.0, + 1330.0, + 662.0, + 1360.0, + 294.0, + 1360.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1360.0, + 663.0, + 1360.0, + 663.0, + 1390.0, + 293.0, + 1390.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 229.0, + 1405.0, + 229.0, + 1405.0, + 265.0, + 295.0, + 265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 260.0, + 1400.0, + 260.0, + 1400.0, + 297.0, + 294.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1388.0, + 1031.0, + 1388.0, + 1031.0, + 1425.0, + 294.0, + 1425.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 5, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1438, + 1405, + 1438, + 1405, + 1804, + 297, + 1804 + ], + "score": 0.983 + }, + { + "category_id": 5, + "poly": [ + 295, + 362, + 1409, + 362, + 1409, + 661, + 295, + 661 + ], + "score": 0.983, + "html": "
Random colorsCURLCURL (PAD)RotRot (PAD)IDMIDM (O-PAD)IDM (PAD)
Walker, walk445±99495±70335±7330±30406±29441±16468±47
Walker, stand662±54753±49673±4653±27743±37727±21797±46
Cartpole, swingup454±110413±67493±52477±38585±73578±69630±63
Cartpole,balance782±13763±5710±72734±81835±40796±37848±29
Ball in cup, catch231±92332±78291±54314±60471±75490±16563±50
Finger, spin691±12588±22695±36689±20757±62767±43803±72
Finger, turn_easy202±32186±2283±68230±53283±51321±10304±46
Cheetah, run202±22211±20127±3135±12121±38112±35159±28
Reacher, easy325±32378±6299±29120±7201±32241±24214±44
" + }, + { + "category_id": 1, + "poly": [ + 298, + 1819, + 1403, + 1819, + 1403, + 2034, + 298, + 2034 + ], + "score": 0.979 + }, + { + "category_id": 1, + "poly": [ + 298, + 1298, + 1403, + 1298, + 1403, + 1422, + 298, + 1422 + ], + "score": 0.977 + }, + { + "category_id": 5, + "poly": [ + 295, + 822, + 1399, + 822, + 1399, + 995, + 295, + 995 + ], + "score": 0.977, + "html": "
CRLMazeRandomA2C+DR+IDM+IDM (PAD)+Rot+Rot (PAD)
Walls-870±30-380±145-260±137-302±150-428±135-206±166-74±116
Floor-868±23-320±167-438±59-47±198-530±106-294±123-209±94
Ceiling-872±30-171±175-400±74166±215-508±104128±196281±83
Lights-900±29-30±213-310±106239±270-460±114-84±53312±104
" + }, + { + "category_id": 1, + "poly": [ + 297, + 1035, + 1404, + 1035, + 1404, + 1220, + 297, + 1220 + ], + "score": 0.973 + }, + { + "category_id": 6, + "poly": [ + 295, + 223, + 1406, + 223, + 1406, + 348, + 295, + 348 + ], + "score": 0.952 + }, + { + "category_id": 2, + "poly": [ + 300, + 76, + 812, + 76, + 812, + 104, + 300, + 104 + ], + "score": 0.897 + }, + { + "category_id": 0, + "poly": [ + 298, + 1250, + 498, + 1250, + 498, + 1282, + 298, + 1282 + ], + "score": 0.892 + }, + { + "category_id": 6, + "poly": [ + 296, + 683, + 1405, + 683, + 1405, + 807, + 296, + 807 + ], + "score": 0.872 + }, + { + "category_id": 2, + "poly": [ + 841, + 2088, + 858, + 2088, + 858, + 2111, + 841, + 2111 + ], + "score": 0.751 + }, + { + "category_id": 1, + "poly": [ + 296, + 683, + 1405, + 683, + 1405, + 807, + 296, + 807 + ], + "score": 0.208 + }, + { + "category_id": 13, + "poly": [ + 870, + 1622, + 932, + 1622, + 932, + 1650, + 870, + 1650 + ], + "score": 0.87, + "latex": "+ D R" + }, + { + "category_id": 13, + "poly": [ + 1109, + 316, + 1184, + 316, + 1184, + 345, + 1109, + 345 + ], + "score": 0.86, + "latex": "+ \\mathrm { I D M }" + }, + { + "category_id": 13, + "poly": [ + 401, + 1652, + 480, + 1652, + 480, + 1682, + 401, + 1682 + ], + "score": 0.86, + "latex": "+ I D M )" + }, + { + "category_id": 13, + "poly": [ + 1199, + 1653, + 1263, + 1653, + 1263, + 1681, + 1199, + 1681 + ], + "score": 0.83, + "latex": "+ R o t )" + }, + { + "category_id": 13, + "poly": [ + 502, + 1683, + 642, + 1683, + 642, + 1713, + 502, + 1713 + ], + "score": 0.63, + "latex": "+ R o t \\ ( P A D )" + }, + { + "category_id": 13, + "poly": [ + 1278, + 1592, + 1335, + 1592, + 1335, + 1620, + 1278, + 1620 + ], + "score": 0.5, + "latex": "A 2 C" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 220.0, + 1406.0, + 220.0, + 1406.0, + 259.0, + 292.0, + 259.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 253.0, + 1408.0, + 253.0, + 1408.0, + 288.0, + 293.0, + 288.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 282.0, + 1406.0, + 282.0, + 1406.0, + 322.0, + 293.0, + 322.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 312.0, + 1108.0, + 312.0, + 1108.0, + 351.0, + 293.0, + 351.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1185.0, + 312.0, + 1384.0, + 312.0, + 1384.0, + 351.0, + 1185.0, + 351.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1246.0, + 503.0, + 1246.0, + 503.0, + 1289.0, + 292.0, + 1289.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 682.0, + 1408.0, + 682.0, + 1408.0, + 717.0, + 293.0, + 717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 715.0, + 1407.0, + 715.0, + 1407.0, + 747.0, + 294.0, + 747.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 744.0, + 1405.0, + 744.0, + 1405.0, + 776.0, + 294.0, + 776.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 775.0, + 1240.0, + 775.0, + 1240.0, + 808.0, + 294.0, + 808.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2085.0, + 861.0, + 2085.0, + 861.0, + 2120.0, + 839.0, + 2120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1439.0, + 1405.0, + 1439.0, + 1405.0, + 1473.0, + 296.0, + 1473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1469.0, + 1408.0, + 1469.0, + 1408.0, + 1503.0, + 294.0, + 1503.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1500.0, + 1406.0, + 1500.0, + 1406.0, + 1534.0, + 295.0, + 1534.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1526.0, + 1410.0, + 1526.0, + 1410.0, + 1566.0, + 292.0, + 1566.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1558.0, + 1406.0, + 1558.0, + 1406.0, + 1596.0, + 294.0, + 1596.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1591.0, + 1277.0, + 1591.0, + 1277.0, + 1625.0, + 295.0, + 1625.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1336.0, + 1591.0, + 1406.0, + 1591.0, + 1406.0, + 1625.0, + 1336.0, + 1625.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1619.0, + 869.0, + 1619.0, + 869.0, + 1654.0, + 292.0, + 1654.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 933.0, + 1619.0, + 1406.0, + 1619.0, + 1406.0, + 1654.0, + 933.0, + 1654.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1651.0, + 400.0, + 1651.0, + 400.0, + 1685.0, + 295.0, + 1685.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 481.0, + 1651.0, + 1198.0, + 1651.0, + 1198.0, + 1685.0, + 481.0, + 1685.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1264.0, + 1651.0, + 1406.0, + 1651.0, + 1406.0, + 1685.0, + 1264.0, + 1685.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1679.0, + 501.0, + 1679.0, + 501.0, + 1717.0, + 294.0, + 1717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 643.0, + 1679.0, + 1408.0, + 1679.0, + 1408.0, + 1717.0, + 643.0, + 1717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1713.0, + 1406.0, + 1713.0, + 1406.0, + 1746.0, + 294.0, + 1746.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1742.0, + 1407.0, + 1742.0, + 1407.0, + 1776.0, + 292.0, + 1776.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1774.0, + 1369.0, + 1774.0, + 1369.0, + 1808.0, + 295.0, + 1808.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1821.0, + 1408.0, + 1821.0, + 1408.0, + 1855.0, + 296.0, + 1855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1849.0, + 1405.0, + 1849.0, + 1405.0, + 1886.0, + 293.0, + 1886.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1882.0, + 1405.0, + 1882.0, + 1405.0, + 1917.0, + 294.0, + 1917.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1910.0, + 1407.0, + 1910.0, + 1407.0, + 1947.0, + 292.0, + 1947.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1942.0, + 1405.0, + 1942.0, + 1405.0, + 1977.0, + 294.0, + 1977.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1973.0, + 1405.0, + 1973.0, + 1405.0, + 2008.0, + 294.0, + 2008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 2004.0, + 1403.0, + 2004.0, + 1403.0, + 2038.0, + 294.0, + 2038.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1296.0, + 1405.0, + 1296.0, + 1405.0, + 1333.0, + 294.0, + 1333.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1327.0, + 1407.0, + 1327.0, + 1407.0, + 1366.0, + 292.0, + 1366.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1360.0, + 1407.0, + 1360.0, + 1407.0, + 1396.0, + 293.0, + 1396.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1389.0, + 1319.0, + 1389.0, + 1319.0, + 1427.0, + 292.0, + 1427.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1037.0, + 1404.0, + 1037.0, + 1404.0, + 1069.0, + 297.0, + 1069.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1065.0, + 1408.0, + 1065.0, + 1408.0, + 1103.0, + 294.0, + 1103.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1097.0, + 1406.0, + 1097.0, + 1406.0, + 1131.0, + 292.0, + 1131.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1124.0, + 1406.0, + 1124.0, + 1406.0, + 1166.0, + 292.0, + 1166.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1158.0, + 1405.0, + 1158.0, + 1405.0, + 1192.0, + 294.0, + 1192.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1189.0, + 1115.0, + 1189.0, + 1115.0, + 1224.0, + 294.0, + 1224.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 682.0, + 1408.0, + 682.0, + 1408.0, + 717.0, + 293.0, + 717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 715.0, + 1407.0, + 715.0, + 1407.0, + 747.0, + 294.0, + 747.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 744.0, + 1405.0, + 744.0, + 1405.0, + 776.0, + 294.0, + 776.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 775.0, + 1240.0, + 775.0, + 1240.0, + 808.0, + 294.0, + 808.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 6, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 5, + "poly": [ + 790, + 895, + 1397, + 895, + 1397, + 1128, + 790, + 1128 + ], + "score": 0.979, + "html": "
Real robotSAC+DR+IDM+IDM (PAD)
Reach (default)100%100%100%100%
Reach (cloth)48%80%56%80%
Reach (disco)72%76%88%92%
Push (default)88%88%92%100%
Push (cloth)60%64%64%88%
Push (disco)60%68%72%84%
" + }, + { + "category_id": 5, + "poly": [ + 790, + 1462, + 1396, + 1462, + 1396, + 1629, + 790, + 1629 + ], + "score": 0.977, + "html": "
Simulated robotSAC+DR+IDM+IDM (PAD)
Push (object)66%64%72%82%
Push (mount)68%58%86%84%
Push (velocity)70%68%70%78%
Push (all)56%50%48%76%
" + }, + { + "category_id": 1, + "poly": [ + 297, + 1651, + 1404, + 1651, + 1404, + 1957, + 297, + 1957 + ], + "score": 0.977 + }, + { + "category_id": 1, + "poly": [ + 297, + 783, + 769, + 783, + 769, + 1148, + 297, + 1148 + ], + "score": 0.974 + }, + { + "category_id": 1, + "poly": [ + 299, + 569, + 1405, + 569, + 1405, + 693, + 299, + 693 + ], + "score": 0.974 + }, + { + "category_id": 1, + "poly": [ + 298, + 1256, + 769, + 1256, + 769, + 1650, + 298, + 1650 + ], + "score": 0.969 + }, + { + "category_id": 1, + "poly": [ + 792, + 1257, + 1402, + 1257, + 1402, + 1441, + 792, + 1441 + ], + "score": 0.95 + }, + { + "category_id": 6, + "poly": [ + 790, + 784, + 1402, + 784, + 1402, + 876, + 790, + 876 + ], + "score": 0.95 + }, + { + "category_id": 1, + "poly": [ + 295, + 1971, + 1403, + 1971, + 1403, + 2035, + 295, + 2035 + ], + "score": 0.931 + }, + { + "category_id": 1, + "poly": [ + 298, + 1150, + 1407, + 1150, + 1407, + 1240, + 298, + 1240 + ], + "score": 0.907 + }, + { + "category_id": 3, + "poly": [ + 303, + 213, + 1395, + 213, + 1395, + 430, + 303, + 430 + ], + "score": 0.901 + }, + { + "category_id": 2, + "poly": [ + 299, + 76, + 813, + 76, + 813, + 104, + 299, + 104 + ], + "score": 0.888 + }, + { + "category_id": 0, + "poly": [ + 302, + 735, + 747, + 735, + 747, + 765, + 302, + 765 + ], + "score": 0.87 + }, + { + "category_id": 1, + "poly": [ + 297, + 476, + 1400, + 476, + 1400, + 538, + 297, + 538 + ], + "score": 0.866 + }, + { + "category_id": 2, + "poly": [ + 840, + 2088, + 858, + 2088, + 858, + 2112, + 840, + 2112 + ], + "score": 0.791 + }, + { + "category_id": 4, + "poly": [ + 310, + 438, + 1267, + 438, + 1267, + 469, + 310, + 469 + ], + "score": 0.494 + }, + { + "category_id": 1, + "poly": [ + 310, + 438, + 1267, + 438, + 1267, + 469, + 310, + 469 + ], + "score": 0.105 + }, + { + "category_id": 13, + "poly": [ + 557, + 1408, + 628, + 1408, + 628, + 1437, + 557, + 1437 + ], + "score": 0.89, + "latex": "k = 1" + }, + { + "category_id": 13, + "poly": [ + 1006, + 845, + 1081, + 845, + 1081, + 874, + 1006, + 874 + ], + "score": 0.86, + "latex": "+ \\mathrm { I D M }" + }, + { + "category_id": 13, + "poly": [ + 1006, + 1410, + 1081, + 1410, + 1081, + 1439, + 1006, + 1439 + ], + "score": 0.83, + "latex": "+ \\mathrm { I D M }" + }, + { + "category_id": 15, + "poly": [ + 789.0, + 783.0, + 1404.0, + 783.0, + 1404.0, + 816.0, + 789.0, + 816.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 790.0, + 814.0, + 1405.0, + 814.0, + 1405.0, + 845.0, + 790.0, + 845.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 789.0, + 845.0, + 1005.0, + 845.0, + 1005.0, + 876.0, + 789.0, + 876.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1082.0, + 845.0, + 1345.0, + 845.0, + 1345.0, + 876.0, + 1082.0, + 876.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 371.0, + 329.0, + 401.0, + 329.0, + 401.0, + 364.0, + 371.0, + 364.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 413.0, + 331.0, + 431.0, + 331.0, + 431.0, + 345.0, + 413.0, + 345.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 624.0, + 343.0, + 655.0, + 343.0, + 655.0, + 366.0, + 624.0, + 366.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 733.0, + 752.0, + 733.0, + 752.0, + 769.0, + 294.0, + 769.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 836.0, + 2085.0, + 859.0, + 2085.0, + 859.0, + 2116.0, + 836.0, + 2116.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 435.0, + 929.0, + 435.0, + 929.0, + 474.0, + 326.0, + 474.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1090.0, + 435.0, + 1265.0, + 435.0, + 1265.0, + 473.0, + 1090.0, + 473.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1649.0, + 1404.0, + 1649.0, + 1404.0, + 1687.0, + 294.0, + 1687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1681.0, + 1406.0, + 1681.0, + 1406.0, + 1716.0, + 294.0, + 1716.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1713.0, + 1406.0, + 1713.0, + 1406.0, + 1747.0, + 292.0, + 1747.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1742.0, + 1406.0, + 1742.0, + 1406.0, + 1776.0, + 292.0, + 1776.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1774.0, + 1406.0, + 1774.0, + 1406.0, + 1808.0, + 292.0, + 1808.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1805.0, + 1405.0, + 1805.0, + 1405.0, + 1837.0, + 294.0, + 1837.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1828.0, + 1408.0, + 1828.0, + 1408.0, + 1874.0, + 291.0, + 1874.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1864.0, + 1407.0, + 1864.0, + 1407.0, + 1900.0, + 291.0, + 1900.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1894.0, + 1405.0, + 1894.0, + 1405.0, + 1930.0, + 294.0, + 1930.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1927.0, + 1277.0, + 1927.0, + 1277.0, + 1959.0, + 295.0, + 1959.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 782.0, + 772.0, + 782.0, + 772.0, + 815.0, + 297.0, + 815.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 811.0, + 771.0, + 811.0, + 771.0, + 844.0, + 294.0, + 844.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 844.0, + 771.0, + 844.0, + 771.0, + 874.0, + 294.0, + 874.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 873.0, + 772.0, + 873.0, + 772.0, + 905.0, + 295.0, + 905.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 903.0, + 772.0, + 903.0, + 772.0, + 935.0, + 294.0, + 935.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 935.0, + 772.0, + 935.0, + 772.0, + 965.0, + 295.0, + 965.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 964.0, + 771.0, + 964.0, + 771.0, + 997.0, + 295.0, + 997.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 995.0, + 774.0, + 995.0, + 774.0, + 1028.0, + 293.0, + 1028.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1027.0, + 772.0, + 1027.0, + 772.0, + 1059.0, + 294.0, + 1059.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1057.0, + 772.0, + 1057.0, + 772.0, + 1089.0, + 295.0, + 1089.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1087.0, + 771.0, + 1087.0, + 771.0, + 1120.0, + 295.0, + 1120.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1116.0, + 772.0, + 1116.0, + 772.0, + 1153.0, + 294.0, + 1153.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 570.0, + 1405.0, + 570.0, + 1405.0, + 603.0, + 293.0, + 603.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 602.0, + 1405.0, + 602.0, + 1405.0, + 635.0, + 294.0, + 635.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 630.0, + 1410.0, + 630.0, + 1410.0, + 667.0, + 292.0, + 667.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 660.0, + 1301.0, + 660.0, + 1301.0, + 696.0, + 293.0, + 696.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1254.0, + 770.0, + 1254.0, + 770.0, + 1289.0, + 296.0, + 1289.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1285.0, + 771.0, + 1285.0, + 771.0, + 1319.0, + 295.0, + 1319.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1316.0, + 771.0, + 1316.0, + 771.0, + 1350.0, + 294.0, + 1350.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1346.0, + 771.0, + 1346.0, + 771.0, + 1381.0, + 295.0, + 1381.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1378.0, + 774.0, + 1378.0, + 774.0, + 1410.0, + 296.0, + 1410.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1408.0, + 556.0, + 1408.0, + 556.0, + 1441.0, + 295.0, + 1441.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 629.0, + 1408.0, + 772.0, + 1408.0, + 772.0, + 1441.0, + 629.0, + 1441.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1440.0, + 770.0, + 1440.0, + 770.0, + 1468.0, + 296.0, + 1468.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1468.0, + 773.0, + 1468.0, + 773.0, + 1500.0, + 295.0, + 1500.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1499.0, + 773.0, + 1499.0, + 773.0, + 1533.0, + 295.0, + 1533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1531.0, + 770.0, + 1531.0, + 770.0, + 1560.0, + 297.0, + 1560.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1560.0, + 771.0, + 1560.0, + 771.0, + 1592.0, + 296.0, + 1592.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1591.0, + 770.0, + 1591.0, + 770.0, + 1623.0, + 295.0, + 1623.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1621.0, + 771.0, + 1621.0, + 771.0, + 1653.0, + 296.0, + 1653.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 789.0, + 1254.0, + 1405.0, + 1254.0, + 1405.0, + 1291.0, + 789.0, + 1291.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 790.0, + 1287.0, + 1405.0, + 1287.0, + 1405.0, + 1320.0, + 790.0, + 1320.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 788.0, + 1316.0, + 1409.0, + 1316.0, + 1409.0, + 1353.0, + 788.0, + 1353.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 790.0, + 1350.0, + 1405.0, + 1350.0, + 1405.0, + 1381.0, + 790.0, + 1381.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 789.0, + 1380.0, + 1406.0, + 1380.0, + 1406.0, + 1411.0, + 789.0, + 1411.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 789.0, + 1409.0, + 1005.0, + 1409.0, + 1005.0, + 1442.0, + 789.0, + 1442.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1082.0, + 1409.0, + 1345.0, + 1409.0, + 1345.0, + 1442.0, + 1082.0, + 1442.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1970.0, + 1405.0, + 1970.0, + 1405.0, + 2007.0, + 294.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 2002.0, + 1405.0, + 2002.0, + 1405.0, + 2036.0, + 292.0, + 2036.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1146.0, + 1405.0, + 1146.0, + 1405.0, + 1182.0, + 292.0, + 1182.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1180.0, + 1409.0, + 1180.0, + 1409.0, + 1213.0, + 295.0, + 1213.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1208.0, + 855.0, + 1208.0, + 855.0, + 1241.0, + 293.0, + 1241.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 474.0, + 1405.0, + 474.0, + 1405.0, + 510.0, + 295.0, + 510.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 503.0, + 1311.0, + 503.0, + 1311.0, + 540.0, + 293.0, + 540.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 435.0, + 929.0, + 435.0, + 929.0, + 474.0, + 326.0, + 474.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1090.0, + 435.0, + 1265.0, + 435.0, + 1265.0, + 473.0, + 1090.0, + 473.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 7, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 230, + 1404, + 230, + 1404, + 595, + 298, + 595 + ], + "score": 0.982 + }, + { + "category_id": 1, + "poly": [ + 298, + 704, + 1405, + 704, + 1405, + 1069, + 298, + 1069 + ], + "score": 0.981 + }, + { + "category_id": 1, + "poly": [ + 298, + 1106, + 1405, + 1106, + 1405, + 1319, + 298, + 1319 + ], + "score": 0.975 + }, + { + "category_id": 0, + "poly": [ + 300, + 643, + 544, + 643, + 544, + 678, + 300, + 678 + ], + "score": 0.895 + }, + { + "category_id": 2, + "poly": [ + 300, + 76, + 812, + 76, + 812, + 104, + 300, + 104 + ], + "score": 0.89 + }, + { + "category_id": 1, + "poly": [ + 293, + 1419, + 1409, + 1419, + 1409, + 2035, + 293, + 2035 + ], + "score": 0.885 + }, + { + "category_id": 0, + "poly": [ + 300, + 1369, + 488, + 1369, + 488, + 1401, + 300, + 1401 + ], + "score": 0.882 + }, + { + "category_id": 2, + "poly": [ + 840, + 2088, + 858, + 2088, + 858, + 2111, + 840, + 2111 + ], + "score": 0.774 + }, + { + "category_id": 13, + "poly": [ + 1160, + 475, + 1218, + 475, + 1218, + 504, + 1160, + 504 + ], + "score": 0.87, + "latex": "28 \\%" + }, + { + "category_id": 13, + "poly": [ + 692, + 262, + 750, + 262, + 750, + 291, + 692, + 291 + ], + "score": 0.87, + "latex": "24 \\%" + }, + { + "category_id": 13, + "poly": [ + 761, + 414, + 887, + 414, + 887, + 443, + 761, + 443 + ], + "score": 0.83, + "latex": "S A C { + } I D M" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 639.0, + 548.0, + 639.0, + 548.0, + 687.0, + 291.0, + 687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1367.0, + 492.0, + 1367.0, + 492.0, + 1407.0, + 296.0, + 1407.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 2087.0, + 860.0, + 2087.0, + 860.0, + 2117.0, + 839.0, + 2117.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 231.0, + 1405.0, + 231.0, + 1405.0, + 265.0, + 296.0, + 265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 261.0, + 691.0, + 261.0, + 691.0, + 297.0, + 295.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 751.0, + 261.0, + 1404.0, + 261.0, + 1404.0, + 297.0, + 751.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 290.0, + 1405.0, + 290.0, + 1405.0, + 327.0, + 293.0, + 327.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 322.0, + 1406.0, + 322.0, + 1406.0, + 356.0, + 294.0, + 356.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 353.0, + 1405.0, + 353.0, + 1405.0, + 387.0, + 294.0, + 387.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 380.0, + 1406.0, + 380.0, + 1406.0, + 419.0, + 292.0, + 419.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 411.0, + 760.0, + 411.0, + 760.0, + 449.0, + 293.0, + 449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 888.0, + 411.0, + 1404.0, + 411.0, + 1404.0, + 449.0, + 888.0, + 449.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 445.0, + 1405.0, + 445.0, + 1405.0, + 475.0, + 296.0, + 475.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 473.0, + 1159.0, + 473.0, + 1159.0, + 511.0, + 293.0, + 511.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1219.0, + 473.0, + 1406.0, + 473.0, + 1406.0, + 511.0, + 1219.0, + 511.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 502.0, + 1405.0, + 502.0, + 1405.0, + 541.0, + 292.0, + 541.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 535.0, + 1404.0, + 535.0, + 1404.0, + 569.0, + 294.0, + 569.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 566.0, + 547.0, + 566.0, + 547.0, + 600.0, + 294.0, + 600.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 701.0, + 1406.0, + 701.0, + 1406.0, + 742.0, + 292.0, + 742.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 735.0, + 1408.0, + 735.0, + 1408.0, + 769.0, + 293.0, + 769.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 765.0, + 1405.0, + 765.0, + 1405.0, + 799.0, + 295.0, + 799.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 796.0, + 1406.0, + 796.0, + 1406.0, + 831.0, + 292.0, + 831.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 827.0, + 1405.0, + 827.0, + 1405.0, + 861.0, + 295.0, + 861.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 857.0, + 1405.0, + 857.0, + 1405.0, + 891.0, + 295.0, + 891.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 887.0, + 1406.0, + 887.0, + 1406.0, + 920.0, + 293.0, + 920.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 918.0, + 1407.0, + 918.0, + 1407.0, + 952.0, + 295.0, + 952.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 947.0, + 1406.0, + 947.0, + 1406.0, + 983.0, + 292.0, + 983.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 979.0, + 1405.0, + 979.0, + 1405.0, + 1012.0, + 295.0, + 1012.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1008.0, + 1407.0, + 1008.0, + 1407.0, + 1041.0, + 293.0, + 1041.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1037.0, + 644.0, + 1037.0, + 644.0, + 1076.0, + 294.0, + 1076.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1105.0, + 1406.0, + 1105.0, + 1406.0, + 1143.0, + 293.0, + 1143.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1137.0, + 1405.0, + 1137.0, + 1405.0, + 1171.0, + 295.0, + 1171.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1164.0, + 1406.0, + 1164.0, + 1406.0, + 1203.0, + 292.0, + 1203.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1195.0, + 1406.0, + 1195.0, + 1406.0, + 1235.0, + 292.0, + 1235.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1225.0, + 1409.0, + 1225.0, + 1409.0, + 1266.0, + 292.0, + 1266.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1255.0, + 1408.0, + 1255.0, + 1408.0, + 1296.0, + 292.0, + 1296.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1290.0, + 370.0, + 1290.0, + 370.0, + 1323.0, + 293.0, + 1323.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1419.0, + 1408.0, + 1419.0, + 1408.0, + 1460.0, + 295.0, + 1460.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 319.0, + 1450.0, + 1406.0, + 1450.0, + 1406.0, + 1491.0, + 319.0, + 1491.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1483.0, + 1407.0, + 1483.0, + 1407.0, + 1518.0, + 322.0, + 1518.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1513.0, + 419.0, + 1513.0, + 419.0, + 1547.0, + 323.0, + 1547.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1567.0, + 1407.0, + 1567.0, + 1407.0, + 1605.0, + 294.0, + 1605.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 319.0, + 1596.0, + 1408.0, + 1596.0, + 1408.0, + 1638.0, + 319.0, + 1638.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1629.0, + 624.0, + 1629.0, + 624.0, + 1664.0, + 323.0, + 1664.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1681.0, + 1406.0, + 1681.0, + 1406.0, + 1723.0, + 294.0, + 1723.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 1717.0, + 929.0, + 1717.0, + 929.0, + 1751.0, + 320.0, + 1751.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1768.0, + 1407.0, + 1768.0, + 1407.0, + 1809.0, + 292.0, + 1809.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 1800.0, + 1231.0, + 1800.0, + 1231.0, + 1838.0, + 320.0, + 1838.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1854.0, + 1406.0, + 1854.0, + 1406.0, + 1896.0, + 292.0, + 1896.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 1885.0, + 1408.0, + 1885.0, + 1408.0, + 1928.0, + 320.0, + 1928.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1918.0, + 556.0, + 1918.0, + 556.0, + 1951.0, + 324.0, + 1951.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1972.0, + 1406.0, + 1972.0, + 1406.0, + 2010.0, + 292.0, + 2010.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 2002.0, + 1369.0, + 2002.0, + 1369.0, + 2040.0, + 320.0, + 2040.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 8, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 2, + "poly": [ + 300, + 75, + 813, + 75, + 813, + 105, + 300, + 105 + ], + "score": 0.883 + }, + { + "category_id": 2, + "poly": [ + 836, + 2088, + 864, + 2088, + 864, + 2113, + 836, + 2113 + ], + "score": 0.829 + }, + { + "category_id": 1, + "poly": [ + 297, + 664, + 1401, + 664, + 1401, + 757, + 297, + 757 + ], + "score": 0.826 + }, + { + "category_id": 1, + "poly": [ + 288, + 1216, + 1402, + 1216, + 1402, + 1281, + 288, + 1281 + ], + "score": 0.823 + }, + { + "category_id": 1, + "poly": [ + 297, + 579, + 1402, + 579, + 1402, + 643, + 297, + 643 + ], + "score": 0.823 + }, + { + "category_id": 1, + "poly": [ + 298, + 781, + 1401, + 781, + 1401, + 876, + 298, + 876 + ], + "score": 0.812 + }, + { + "category_id": 1, + "poly": [ + 298, + 897, + 1400, + 897, + 1400, + 963, + 298, + 963 + ], + "score": 0.807 + }, + { + "category_id": 1, + "poly": [ + 295, + 1303, + 1401, + 1303, + 1401, + 1367, + 295, + 1367 + ], + "score": 0.791 + }, + { + "category_id": 1, + "poly": [ + 304, + 229, + 1405, + 229, + 1405, + 324, + 304, + 324 + ], + "score": 0.783 + }, + { + "category_id": 1, + "poly": [ + 299, + 1738, + 1404, + 1738, + 1404, + 1832, + 299, + 1832 + ], + "score": 0.778 + }, + { + "category_id": 1, + "poly": [ + 298, + 1474, + 1405, + 1474, + 1405, + 1566, + 298, + 1566 + ], + "score": 0.776 + }, + { + "category_id": 1, + "poly": [ + 299, + 1100, + 1406, + 1100, + 1406, + 1194, + 299, + 1194 + ], + "score": 0.776 + }, + { + "category_id": 1, + "poly": [ + 297, + 1591, + 1405, + 1591, + 1405, + 1717, + 297, + 1717 + ], + "score": 0.775 + }, + { + "category_id": 1, + "poly": [ + 299, + 983, + 1405, + 983, + 1405, + 1077, + 299, + 1077 + ], + "score": 0.756 + }, + { + "category_id": 1, + "poly": [ + 299, + 346, + 1401, + 346, + 1401, + 410, + 299, + 410 + ], + "score": 0.754 + }, + { + "category_id": 1, + "poly": [ + 296, + 1972, + 1399, + 1972, + 1399, + 2034, + 296, + 2034 + ], + "score": 0.724 + }, + { + "category_id": 1, + "poly": [ + 294, + 1389, + 1403, + 1389, + 1403, + 1453, + 294, + 1453 + ], + "score": 0.711 + }, + { + "category_id": 1, + "poly": [ + 297, + 431, + 1406, + 431, + 1406, + 555, + 297, + 555 + ], + "score": 0.686 + }, + { + "category_id": 1, + "poly": [ + 300, + 1854, + 1404, + 1854, + 1404, + 1948, + 300, + 1948 + ], + "score": 0.684 + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 869.0, + 2085.0, + 869.0, + 2123.0, + 831.0, + 2123.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 665.0, + 1406.0, + 665.0, + 1406.0, + 702.0, + 295.0, + 702.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 691.0, + 1407.0, + 691.0, + 1407.0, + 736.0, + 320.0, + 736.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 725.0, + 635.0, + 725.0, + 635.0, + 759.0, + 322.0, + 759.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1215.0, + 1405.0, + 1215.0, + 1405.0, + 1252.0, + 294.0, + 1252.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1249.0, + 1007.0, + 1249.0, + 1007.0, + 1282.0, + 323.0, + 1282.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 573.0, + 1406.0, + 573.0, + 1406.0, + 620.0, + 292.0, + 620.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 610.0, + 804.0, + 610.0, + 804.0, + 644.0, + 321.0, + 644.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 781.0, + 1406.0, + 781.0, + 1406.0, + 819.0, + 294.0, + 819.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 813.0, + 1405.0, + 813.0, + 1405.0, + 847.0, + 323.0, + 847.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 842.0, + 1078.0, + 842.0, + 1078.0, + 878.0, + 323.0, + 878.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 896.0, + 1405.0, + 896.0, + 1405.0, + 934.0, + 294.0, + 934.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 930.0, + 1387.0, + 930.0, + 1387.0, + 962.0, + 322.0, + 962.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1298.0, + 1406.0, + 1298.0, + 1406.0, + 1343.0, + 294.0, + 1343.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1334.0, + 1246.0, + 1334.0, + 1246.0, + 1367.0, + 323.0, + 1367.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 227.0, + 1405.0, + 227.0, + 1405.0, + 267.0, + 297.0, + 267.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 262.0, + 1407.0, + 262.0, + 1407.0, + 296.0, + 322.0, + 296.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 290.0, + 1122.0, + 290.0, + 1122.0, + 328.0, + 322.0, + 328.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1734.0, + 1405.0, + 1734.0, + 1405.0, + 1777.0, + 292.0, + 1777.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 319.0, + 1767.0, + 1409.0, + 1767.0, + 1409.0, + 1807.0, + 319.0, + 1807.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1798.0, + 490.0, + 1798.0, + 490.0, + 1834.0, + 322.0, + 1834.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1474.0, + 1405.0, + 1474.0, + 1405.0, + 1512.0, + 295.0, + 1512.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1506.0, + 1406.0, + 1506.0, + 1406.0, + 1539.0, + 323.0, + 1539.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1538.0, + 347.0, + 1538.0, + 347.0, + 1567.0, + 322.0, + 1567.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1101.0, + 1404.0, + 1101.0, + 1404.0, + 1135.0, + 296.0, + 1135.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1132.0, + 1406.0, + 1132.0, + 1406.0, + 1166.0, + 323.0, + 1166.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1162.0, + 777.0, + 1162.0, + 777.0, + 1196.0, + 324.0, + 1196.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1589.0, + 1410.0, + 1589.0, + 1410.0, + 1628.0, + 295.0, + 1628.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 1622.0, + 1405.0, + 1622.0, + 1405.0, + 1656.0, + 321.0, + 1656.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 1654.0, + 1405.0, + 1654.0, + 1405.0, + 1689.0, + 321.0, + 1689.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 1682.0, + 1201.0, + 1682.0, + 1201.0, + 1719.0, + 321.0, + 1719.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 984.0, + 1405.0, + 984.0, + 1405.0, + 1018.0, + 297.0, + 1018.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1015.0, + 1403.0, + 1015.0, + 1403.0, + 1049.0, + 323.0, + 1049.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1045.0, + 682.0, + 1045.0, + 682.0, + 1078.0, + 323.0, + 1078.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 343.0, + 1405.0, + 343.0, + 1405.0, + 382.0, + 293.0, + 382.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 378.0, + 1382.0, + 378.0, + 1382.0, + 410.0, + 323.0, + 410.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1966.0, + 1404.0, + 1966.0, + 1404.0, + 2011.0, + 294.0, + 2011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 2000.0, + 923.0, + 2000.0, + 923.0, + 2036.0, + 322.0, + 2036.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1389.0, + 1405.0, + 1389.0, + 1405.0, + 1425.0, + 296.0, + 1425.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1418.0, + 1404.0, + 1418.0, + 1404.0, + 1454.0, + 322.0, + 1454.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 431.0, + 1406.0, + 431.0, + 1406.0, + 470.0, + 292.0, + 470.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 463.0, + 1407.0, + 463.0, + 1407.0, + 499.0, + 321.0, + 499.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 493.0, + 1411.0, + 493.0, + 1411.0, + 529.0, + 323.0, + 529.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 525.0, + 616.0, + 525.0, + 616.0, + 557.0, + 323.0, + 557.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1851.0, + 1408.0, + 1851.0, + 1408.0, + 1891.0, + 293.0, + 1891.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 1882.0, + 1405.0, + 1882.0, + 1405.0, + 1924.0, + 320.0, + 1924.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1916.0, + 1119.0, + 1916.0, + 1119.0, + 1950.0, + 324.0, + 1950.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 9, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 2, + "poly": [ + 300, + 75, + 813, + 75, + 813, + 105, + 300, + 105 + ], + "score": 0.877 + }, + { + "category_id": 1, + "poly": [ + 301, + 777, + 1401, + 777, + 1401, + 871, + 301, + 871 + ], + "score": 0.817 + }, + { + "category_id": 2, + "poly": [ + 835, + 2088, + 862, + 2088, + 862, + 2113, + 835, + 2113 + ], + "score": 0.811 + }, + { + "category_id": 1, + "poly": [ + 297, + 893, + 1403, + 893, + 1403, + 988, + 297, + 988 + ], + "score": 0.809 + }, + { + "category_id": 1, + "poly": [ + 297, + 1010, + 1399, + 1010, + 1399, + 1076, + 297, + 1076 + ], + "score": 0.809 + }, + { + "category_id": 1, + "poly": [ + 298, + 1213, + 1401, + 1213, + 1401, + 1309, + 298, + 1309 + ], + "score": 0.785 + }, + { + "category_id": 1, + "poly": [ + 297, + 690, + 1404, + 690, + 1404, + 756, + 297, + 756 + ], + "score": 0.773 + }, + { + "category_id": 1, + "poly": [ + 297, + 1330, + 1402, + 1330, + 1402, + 1426, + 297, + 1426 + ], + "score": 0.756 + }, + { + "category_id": 1, + "poly": [ + 298, + 1097, + 1405, + 1097, + 1405, + 1191, + 298, + 1191 + ], + "score": 0.753 + }, + { + "category_id": 1, + "poly": [ + 298, + 604, + 1404, + 604, + 1404, + 669, + 298, + 669 + ], + "score": 0.751 + }, + { + "category_id": 1, + "poly": [ + 299, + 518, + 1399, + 518, + 1399, + 583, + 299, + 583 + ], + "score": 0.749 + }, + { + "category_id": 1, + "poly": [ + 297, + 1447, + 1404, + 1447, + 1404, + 1542, + 297, + 1542 + ], + "score": 0.729 + }, + { + "category_id": 1, + "poly": [ + 294, + 228, + 1404, + 228, + 1404, + 293, + 294, + 293 + ], + "score": 0.717 + }, + { + "category_id": 1, + "poly": [ + 298, + 1736, + 1403, + 1736, + 1403, + 1832, + 298, + 1832 + ], + "score": 0.707 + }, + { + "category_id": 1, + "poly": [ + 296, + 1565, + 1405, + 1565, + 1405, + 1630, + 296, + 1630 + ], + "score": 0.707 + }, + { + "category_id": 1, + "poly": [ + 297, + 1650, + 1403, + 1650, + 1403, + 1715, + 297, + 1715 + ], + "score": 0.706 + }, + { + "category_id": 1, + "poly": [ + 289, + 1853, + 1404, + 1853, + 1404, + 1918, + 289, + 1918 + ], + "score": 0.697 + }, + { + "category_id": 1, + "poly": [ + 302, + 314, + 1403, + 314, + 1403, + 409, + 302, + 409 + ], + "score": 0.683 + }, + { + "category_id": 1, + "poly": [ + 300, + 431, + 1403, + 431, + 1403, + 497, + 300, + 497 + ], + "score": 0.669 + }, + { + "category_id": 1, + "poly": [ + 299, + 1942, + 1404, + 1942, + 1404, + 2034, + 299, + 2034 + ], + "score": 0.639 + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 867.0, + 2085.0, + 867.0, + 2125.0, + 831.0, + 2125.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 779.0, + 1406.0, + 779.0, + 1406.0, + 813.0, + 297.0, + 813.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 809.0, + 1409.0, + 809.0, + 1409.0, + 843.0, + 321.0, + 843.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 836.0, + 402.0, + 836.0, + 402.0, + 874.0, + 320.0, + 874.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 895.0, + 1404.0, + 895.0, + 1404.0, + 930.0, + 296.0, + 930.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 921.0, + 1409.0, + 921.0, + 1409.0, + 965.0, + 320.0, + 965.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 956.0, + 528.0, + 956.0, + 528.0, + 987.0, + 322.0, + 987.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1011.0, + 1404.0, + 1011.0, + 1404.0, + 1047.0, + 296.0, + 1047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1043.0, + 529.0, + 1043.0, + 529.0, + 1075.0, + 323.0, + 1075.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1211.0, + 1405.0, + 1211.0, + 1405.0, + 1251.0, + 296.0, + 1251.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1244.0, + 1405.0, + 1244.0, + 1405.0, + 1282.0, + 322.0, + 1282.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1276.0, + 593.0, + 1276.0, + 593.0, + 1311.0, + 324.0, + 1311.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 689.0, + 1404.0, + 689.0, + 1404.0, + 728.0, + 296.0, + 728.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 722.0, + 984.0, + 722.0, + 984.0, + 757.0, + 322.0, + 757.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1330.0, + 1406.0, + 1330.0, + 1406.0, + 1368.0, + 293.0, + 1368.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1362.0, + 1404.0, + 1362.0, + 1404.0, + 1399.0, + 322.0, + 1399.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1393.0, + 1103.0, + 1393.0, + 1103.0, + 1427.0, + 324.0, + 1427.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1097.0, + 1405.0, + 1097.0, + 1405.0, + 1134.0, + 296.0, + 1134.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1127.0, + 1405.0, + 1127.0, + 1405.0, + 1165.0, + 322.0, + 1165.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1160.0, + 619.0, + 1160.0, + 619.0, + 1189.0, + 322.0, + 1189.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 603.0, + 1405.0, + 603.0, + 1405.0, + 644.0, + 293.0, + 644.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 635.0, + 1337.0, + 635.0, + 1337.0, + 670.0, + 322.0, + 670.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 520.0, + 1404.0, + 520.0, + 1404.0, + 556.0, + 295.0, + 556.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 549.0, + 1205.0, + 549.0, + 1205.0, + 584.0, + 321.0, + 584.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1448.0, + 1405.0, + 1448.0, + 1405.0, + 1482.0, + 296.0, + 1482.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 1475.0, + 1405.0, + 1475.0, + 1405.0, + 1518.0, + 320.0, + 1518.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1510.0, + 696.0, + 1510.0, + 696.0, + 1544.0, + 322.0, + 1544.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 227.0, + 1406.0, + 227.0, + 1406.0, + 264.0, + 294.0, + 264.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 261.0, + 797.0, + 261.0, + 797.0, + 293.0, + 323.0, + 293.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1736.0, + 1407.0, + 1736.0, + 1407.0, + 1773.0, + 293.0, + 1773.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1768.0, + 1405.0, + 1768.0, + 1405.0, + 1805.0, + 323.0, + 1805.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 326.0, + 1800.0, + 948.0, + 1800.0, + 948.0, + 1834.0, + 326.0, + 1834.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1561.0, + 1405.0, + 1561.0, + 1405.0, + 1604.0, + 293.0, + 1604.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 1594.0, + 1267.0, + 1594.0, + 1267.0, + 1632.0, + 320.0, + 1632.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1646.0, + 1409.0, + 1646.0, + 1409.0, + 1692.0, + 295.0, + 1692.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1681.0, + 1057.0, + 1681.0, + 1057.0, + 1717.0, + 324.0, + 1717.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1851.0, + 1405.0, + 1851.0, + 1405.0, + 1893.0, + 293.0, + 1893.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1887.0, + 834.0, + 1887.0, + 834.0, + 1919.0, + 323.0, + 1919.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 314.0, + 1407.0, + 314.0, + 1407.0, + 352.0, + 296.0, + 352.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 346.0, + 1408.0, + 346.0, + 1408.0, + 382.0, + 321.0, + 382.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 379.0, + 346.0, + 379.0, + 346.0, + 408.0, + 322.0, + 408.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 430.0, + 1408.0, + 430.0, + 1408.0, + 468.0, + 294.0, + 468.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 463.0, + 1385.0, + 463.0, + 1385.0, + 498.0, + 321.0, + 498.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1940.0, + 1408.0, + 1940.0, + 1408.0, + 1977.0, + 294.0, + 1977.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1971.0, + 1405.0, + 1971.0, + 1405.0, + 2008.0, + 322.0, + 2008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 2002.0, + 782.0, + 2002.0, + 782.0, + 2036.0, + 323.0, + 2036.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 10, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 2, + "poly": [ + 299, + 75, + 813, + 75, + 813, + 105, + 299, + 105 + ], + "score": 0.869 + }, + { + "category_id": 1, + "poly": [ + 298, + 229, + 1403, + 229, + 1403, + 323, + 298, + 323 + ], + "score": 0.826 + }, + { + "category_id": 2, + "poly": [ + 836, + 2088, + 864, + 2088, + 864, + 2113, + 836, + 2113 + ], + "score": 0.825 + }, + { + "category_id": 1, + "poly": [ + 301, + 343, + 1402, + 343, + 1402, + 435, + 301, + 435 + ], + "score": 0.763 + }, + { + "category_id": 1, + "poly": [ + 296, + 1017, + 1401, + 1017, + 1401, + 1080, + 296, + 1080 + ], + "score": 0.76 + }, + { + "category_id": 1, + "poly": [ + 297, + 1099, + 1402, + 1099, + 1402, + 1164, + 297, + 1164 + ], + "score": 0.75 + }, + { + "category_id": 1, + "poly": [ + 296, + 819, + 1400, + 819, + 1400, + 884, + 296, + 884 + ], + "score": 0.744 + }, + { + "category_id": 1, + "poly": [ + 295, + 1462, + 1403, + 1462, + 1403, + 1529, + 295, + 1529 + ], + "score": 0.739 + }, + { + "category_id": 1, + "poly": [ + 298, + 1380, + 1403, + 1380, + 1403, + 1445, + 298, + 1445 + ], + "score": 0.734 + }, + { + "category_id": 1, + "poly": [ + 299, + 1942, + 1406, + 1942, + 1406, + 2034, + 299, + 2034 + ], + "score": 0.718 + }, + { + "category_id": 1, + "poly": [ + 298, + 736, + 1400, + 736, + 1400, + 801, + 298, + 801 + ], + "score": 0.711 + }, + { + "category_id": 1, + "poly": [ + 306, + 1827, + 1408, + 1827, + 1408, + 1922, + 306, + 1922 + ], + "score": 0.71 + }, + { + "category_id": 1, + "poly": [ + 295, + 1629, + 1402, + 1629, + 1402, + 1694, + 295, + 1694 + ], + "score": 0.703 + }, + { + "category_id": 1, + "poly": [ + 297, + 1183, + 1404, + 1183, + 1404, + 1277, + 297, + 1277 + ], + "score": 0.701 + }, + { + "category_id": 1, + "poly": [ + 296, + 1546, + 1402, + 1546, + 1402, + 1611, + 296, + 1611 + ], + "score": 0.691 + }, + { + "category_id": 1, + "poly": [ + 294, + 570, + 1401, + 570, + 1401, + 635, + 294, + 635 + ], + "score": 0.684 + }, + { + "category_id": 1, + "poly": [ + 296, + 653, + 1403, + 653, + 1403, + 717, + 296, + 717 + ], + "score": 0.684 + }, + { + "category_id": 1, + "poly": [ + 305, + 1713, + 1408, + 1713, + 1408, + 1809, + 305, + 1809 + ], + "score": 0.649 + }, + { + "category_id": 1, + "poly": [ + 293, + 1297, + 1403, + 1297, + 1403, + 1361, + 293, + 1361 + ], + "score": 0.584 + }, + { + "category_id": 1, + "poly": [ + 299, + 456, + 1404, + 456, + 1404, + 550, + 299, + 550 + ], + "score": 0.552 + }, + { + "category_id": 1, + "poly": [ + 297, + 902, + 1405, + 902, + 1405, + 997, + 297, + 997 + ], + "score": 0.523 + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 868.0, + 2085.0, + 868.0, + 2124.0, + 831.0, + 2124.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 227.0, + 1407.0, + 227.0, + 1407.0, + 266.0, + 292.0, + 266.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 258.0, + 1409.0, + 258.0, + 1409.0, + 299.0, + 321.0, + 299.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 288.0, + 632.0, + 288.0, + 632.0, + 325.0, + 321.0, + 325.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 341.0, + 1407.0, + 341.0, + 1407.0, + 379.0, + 292.0, + 379.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 373.0, + 1411.0, + 373.0, + 1411.0, + 411.0, + 320.0, + 411.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 407.0, + 346.0, + 407.0, + 346.0, + 435.0, + 322.0, + 435.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1013.0, + 1406.0, + 1013.0, + 1406.0, + 1056.0, + 293.0, + 1056.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1047.0, + 830.0, + 1047.0, + 830.0, + 1082.0, + 323.0, + 1082.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1101.0, + 1407.0, + 1101.0, + 1407.0, + 1137.0, + 295.0, + 1137.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1132.0, + 1023.0, + 1132.0, + 1023.0, + 1164.0, + 324.0, + 1164.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 819.0, + 1404.0, + 819.0, + 1404.0, + 857.0, + 294.0, + 857.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 850.0, + 1077.0, + 850.0, + 1077.0, + 886.0, + 320.0, + 886.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1462.0, + 1405.0, + 1462.0, + 1405.0, + 1499.0, + 295.0, + 1499.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1493.0, + 1262.0, + 1493.0, + 1262.0, + 1529.0, + 322.0, + 1529.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1376.0, + 1404.0, + 1376.0, + 1404.0, + 1419.0, + 293.0, + 1419.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 1410.0, + 704.0, + 1410.0, + 704.0, + 1446.0, + 321.0, + 1446.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1941.0, + 1406.0, + 1941.0, + 1406.0, + 1978.0, + 294.0, + 1978.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 1972.0, + 1406.0, + 1972.0, + 1406.0, + 2006.0, + 321.0, + 2006.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 2001.0, + 1372.0, + 2001.0, + 1372.0, + 2037.0, + 322.0, + 2037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 739.0, + 1402.0, + 739.0, + 1402.0, + 771.0, + 297.0, + 771.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 769.0, + 1062.0, + 769.0, + 1062.0, + 801.0, + 322.0, + 801.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 299.0, + 1823.0, + 1408.0, + 1823.0, + 1408.0, + 1865.0, + 299.0, + 1865.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1859.0, + 1408.0, + 1859.0, + 1408.0, + 1893.0, + 322.0, + 1893.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1889.0, + 1254.0, + 1889.0, + 1254.0, + 1923.0, + 322.0, + 1923.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1629.0, + 1405.0, + 1629.0, + 1405.0, + 1665.0, + 297.0, + 1665.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1662.0, + 1317.0, + 1662.0, + 1317.0, + 1695.0, + 322.0, + 1695.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 1183.0, + 1406.0, + 1183.0, + 1406.0, + 1217.0, + 297.0, + 1217.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1215.0, + 1405.0, + 1215.0, + 1405.0, + 1249.0, + 324.0, + 1249.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1246.0, + 635.0, + 1246.0, + 635.0, + 1276.0, + 322.0, + 1276.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1544.0, + 1404.0, + 1544.0, + 1404.0, + 1585.0, + 295.0, + 1585.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1579.0, + 785.0, + 1579.0, + 785.0, + 1611.0, + 323.0, + 1611.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 566.0, + 1406.0, + 566.0, + 1406.0, + 610.0, + 293.0, + 610.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 601.0, + 1404.0, + 601.0, + 1404.0, + 636.0, + 322.0, + 636.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 652.0, + 1406.0, + 652.0, + 1406.0, + 689.0, + 294.0, + 689.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 685.0, + 1385.0, + 685.0, + 1385.0, + 720.0, + 321.0, + 720.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 300.0, + 1711.0, + 1403.0, + 1711.0, + 1403.0, + 1750.0, + 300.0, + 1750.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1745.0, + 1405.0, + 1745.0, + 1405.0, + 1779.0, + 323.0, + 1779.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 1774.0, + 1268.0, + 1774.0, + 1268.0, + 1812.0, + 324.0, + 1812.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1293.0, + 1405.0, + 1293.0, + 1405.0, + 1334.0, + 293.0, + 1334.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 1328.0, + 525.0, + 1328.0, + 525.0, + 1361.0, + 320.0, + 1361.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 454.0, + 1406.0, + 454.0, + 1406.0, + 493.0, + 294.0, + 493.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 490.0, + 1404.0, + 490.0, + 1404.0, + 524.0, + 323.0, + 524.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 516.0, + 789.0, + 516.0, + 789.0, + 554.0, + 322.0, + 554.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 901.0, + 1405.0, + 901.0, + 1405.0, + 941.0, + 294.0, + 941.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 935.0, + 1403.0, + 935.0, + 1403.0, + 969.0, + 324.0, + 969.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 967.0, + 633.0, + 967.0, + 633.0, + 997.0, + 322.0, + 997.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 11, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 2, + "poly": [ + 299, + 75, + 814, + 75, + 814, + 105, + 299, + 105 + ], + "score": 0.874 + }, + { + "category_id": 2, + "poly": [ + 835, + 2088, + 864, + 2088, + 864, + 2113, + 835, + 2113 + ], + "score": 0.831 + }, + { + "category_id": 1, + "poly": [ + 297, + 1068, + 1400, + 1068, + 1400, + 1132, + 297, + 1132 + ], + "score": 0.766 + }, + { + "category_id": 1, + "poly": [ + 295, + 1150, + 1402, + 1150, + 1402, + 1215, + 295, + 1215 + ], + "score": 0.756 + }, + { + "category_id": 1, + "poly": [ + 295, + 984, + 1403, + 984, + 1403, + 1049, + 295, + 1049 + ], + "score": 0.749 + }, + { + "category_id": 1, + "poly": [ + 300, + 229, + 1403, + 229, + 1403, + 323, + 300, + 323 + ], + "score": 0.744 + }, + { + "category_id": 1, + "poly": [ + 295, + 456, + 1402, + 456, + 1402, + 519, + 295, + 519 + ], + "score": 0.711 + }, + { + "category_id": 1, + "poly": [ + 294, + 818, + 1401, + 818, + 1401, + 883, + 294, + 883 + ], + "score": 0.711 + }, + { + "category_id": 1, + "poly": [ + 297, + 902, + 1400, + 902, + 1400, + 966, + 297, + 966 + ], + "score": 0.706 + }, + { + "category_id": 1, + "poly": [ + 300, + 343, + 1398, + 343, + 1398, + 436, + 300, + 436 + ], + "score": 0.693 + }, + { + "category_id": 1, + "poly": [ + 294, + 539, + 1404, + 539, + 1404, + 602, + 294, + 602 + ], + "score": 0.673 + }, + { + "category_id": 1, + "poly": [ + 297, + 622, + 1400, + 622, + 1400, + 686, + 297, + 686 + ], + "score": 0.665 + }, + { + "category_id": 1, + "poly": [ + 295, + 705, + 1404, + 705, + 1404, + 800, + 295, + 800 + ], + "score": 0.641 + }, + { + "category_id": 1, + "poly": [ + 299, + 1234, + 1404, + 1234, + 1404, + 1329, + 299, + 1329 + ], + "score": 0.637 + }, + { + "category_id": 15, + "poly": [ + 296.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 296.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 868.0, + 2085.0, + 868.0, + 2123.0, + 831.0, + 2123.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1063.0, + 1405.0, + 1063.0, + 1405.0, + 1108.0, + 293.0, + 1108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1099.0, + 1285.0, + 1099.0, + 1285.0, + 1133.0, + 322.0, + 1133.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1146.0, + 1406.0, + 1146.0, + 1406.0, + 1190.0, + 293.0, + 1190.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 1183.0, + 1034.0, + 1183.0, + 1034.0, + 1216.0, + 323.0, + 1216.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 984.0, + 1404.0, + 984.0, + 1404.0, + 1020.0, + 295.0, + 1020.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1017.0, + 743.0, + 1017.0, + 743.0, + 1047.0, + 322.0, + 1047.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 227.0, + 1404.0, + 227.0, + 1404.0, + 265.0, + 294.0, + 265.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 260.0, + 1408.0, + 260.0, + 1408.0, + 297.0, + 321.0, + 297.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 320.0, + 291.0, + 602.0, + 291.0, + 602.0, + 326.0, + 320.0, + 326.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 454.0, + 1408.0, + 454.0, + 1408.0, + 494.0, + 294.0, + 494.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 486.0, + 528.0, + 486.0, + 528.0, + 519.0, + 323.0, + 519.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 817.0, + 1403.0, + 817.0, + 1403.0, + 856.0, + 294.0, + 856.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 848.0, + 976.0, + 848.0, + 976.0, + 884.0, + 322.0, + 884.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 902.0, + 1405.0, + 902.0, + 1405.0, + 938.0, + 297.0, + 938.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 323.0, + 933.0, + 1405.0, + 933.0, + 1405.0, + 969.0, + 323.0, + 969.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 342.0, + 1404.0, + 342.0, + 1404.0, + 380.0, + 295.0, + 380.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 324.0, + 375.0, + 1403.0, + 375.0, + 1403.0, + 409.0, + 324.0, + 409.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 403.0, + 1051.0, + 403.0, + 1051.0, + 440.0, + 322.0, + 440.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 537.0, + 1407.0, + 537.0, + 1407.0, + 577.0, + 294.0, + 577.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 571.0, + 766.0, + 571.0, + 766.0, + 602.0, + 322.0, + 602.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 622.0, + 1402.0, + 622.0, + 1402.0, + 658.0, + 295.0, + 658.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 653.0, + 1355.0, + 653.0, + 1355.0, + 687.0, + 322.0, + 687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 703.0, + 1407.0, + 703.0, + 1407.0, + 744.0, + 293.0, + 744.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 321.0, + 737.0, + 1405.0, + 737.0, + 1405.0, + 774.0, + 321.0, + 774.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 767.0, + 885.0, + 767.0, + 885.0, + 802.0, + 322.0, + 802.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1229.0, + 1405.0, + 1229.0, + 1405.0, + 1275.0, + 292.0, + 1275.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1264.0, + 1406.0, + 1264.0, + 1406.0, + 1302.0, + 322.0, + 1302.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 322.0, + 1296.0, + 834.0, + 1296.0, + 834.0, + 1330.0, + 322.0, + 1330.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 12, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 299, + 1406, + 299, + 1406, + 971, + 297, + 971 + ], + "score": 0.983 + }, + { + "category_id": 5, + "poly": [ + 432, + 1125, + 1261, + 1125, + 1261, + 1433, + 432, + 1433 + ], + "score": 0.983, + "html": "
Training env.BlindSAC+DR+IDM+IDM (PAD)
Walker, walk235±17847±71756±71911±24895±28
Walker, stand388±10959±11928±36966±8956±20
Cartpole, swingup132±41850±28807±36849±30845±34
Cartpole,balance646±131978±22971±30982±20979±21
Ball in cup, catch150±96725±355469±339919±118910±129
Finger, spin3±2809±138686±295928±45927±45
Finger, turn_easy172±27462±146243±124462±152455±160
Cheetah, run264±75387±74195±46384±88380±91
Reacher, easy107±11264±11392±45390±126365±114
" + }, + { + "category_id": 1, + "poly": [ + 297, + 1819, + 1404, + 1819, + 1404, + 2034, + 297, + 2034 + ], + "score": 0.976 + }, + { + "category_id": 5, + "poly": [ + 297, + 1599, + 1401, + 1599, + 1401, + 1690, + 297, + 1690 + ], + "score": 0.969, + "html": "
CRLMazeRandomA2C+DR+IDM+IDM (PAD)+Rot+Rot (PAD)
Training env.-868±34371±198-355±93585±246-416±135729±148681±99
" + }, + { + "category_id": 6, + "poly": [ + 300, + 1006, + 1404, + 1006, + 1404, + 1100, + 300, + 1100 + ], + "score": 0.952 + }, + { + "category_id": 0, + "poly": [ + 298, + 1746, + 998, + 1746, + 998, + 1782, + 298, + 1782 + ], + "score": 0.913 + }, + { + "category_id": 6, + "poly": [ + 298, + 1490, + 1399, + 1490, + 1399, + 1582, + 298, + 1582 + ], + "score": 0.911 + }, + { + "category_id": 2, + "poly": [ + 300, + 76, + 813, + 76, + 813, + 104, + 300, + 104 + ], + "score": 0.887 + }, + { + "category_id": 2, + "poly": [ + 835, + 2088, + 863, + 2088, + 863, + 2112, + 835, + 2112 + ], + "score": 0.843 + }, + { + "category_id": 0, + "poly": [ + 300, + 229, + 1063, + 229, + 1063, + 261, + 300, + 261 + ], + "score": 0.802 + }, + { + "category_id": 13, + "poly": [ + 1105, + 666, + 1163, + 666, + 1163, + 695, + 1105, + 695 + ], + "score": 0.87, + "latex": "15 \\%" + }, + { + "category_id": 13, + "poly": [ + 297, + 878, + 352, + 878, + 352, + 908, + 297, + 908 + ], + "score": 0.86, + "latex": "51 \\%" + }, + { + "category_id": 13, + "poly": [ + 1271, + 1037, + 1346, + 1037, + 1346, + 1066, + 1271, + 1066 + ], + "score": 0.84, + "latex": "+ \\mathrm { I D M }" + }, + { + "category_id": 13, + "poly": [ + 409, + 1881, + 537, + 1881, + 537, + 1911, + 409, + 1911 + ], + "score": 0.67, + "latex": "\\mathrm { S A C + I D M }" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1004.0, + 1408.0, + 1004.0, + 1408.0, + 1042.0, + 293.0, + 1042.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1036.0, + 1270.0, + 1036.0, + 1270.0, + 1070.0, + 294.0, + 1070.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1347.0, + 1036.0, + 1404.0, + 1036.0, + 1404.0, + 1070.0, + 1347.0, + 1070.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1066.0, + 1332.0, + 1066.0, + 1332.0, + 1104.0, + 294.0, + 1104.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1743.0, + 1004.0, + 1743.0, + 1004.0, + 1788.0, + 291.0, + 1788.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1488.0, + 1404.0, + 1488.0, + 1404.0, + 1524.0, + 293.0, + 1524.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1519.0, + 1405.0, + 1519.0, + 1405.0, + 1556.0, + 294.0, + 1556.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1552.0, + 770.0, + 1552.0, + 770.0, + 1584.0, + 293.0, + 1584.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 72.0, + 816.0, + 72.0, + 816.0, + 110.0, + 294.0, + 110.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2084.0, + 869.0, + 2084.0, + 869.0, + 2123.0, + 831.0, + 2123.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 225.0, + 1066.0, + 225.0, + 1066.0, + 267.0, + 294.0, + 267.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 301.0, + 1403.0, + 301.0, + 1403.0, + 333.0, + 296.0, + 333.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 330.0, + 1403.0, + 330.0, + 1403.0, + 362.0, + 296.0, + 362.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 362.0, + 1404.0, + 362.0, + 1404.0, + 398.0, + 294.0, + 398.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 391.0, + 1404.0, + 391.0, + 1404.0, + 427.0, + 295.0, + 427.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 420.0, + 1406.0, + 420.0, + 1406.0, + 457.0, + 292.0, + 457.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 452.0, + 1406.0, + 452.0, + 1406.0, + 488.0, + 294.0, + 488.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 481.0, + 1407.0, + 481.0, + 1407.0, + 518.0, + 292.0, + 518.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 513.0, + 1408.0, + 513.0, + 1408.0, + 549.0, + 295.0, + 549.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 544.0, + 1406.0, + 544.0, + 1406.0, + 579.0, + 294.0, + 579.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 577.0, + 1403.0, + 577.0, + 1403.0, + 608.0, + 296.0, + 608.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 605.0, + 1406.0, + 605.0, + 1406.0, + 640.0, + 295.0, + 640.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 635.0, + 1407.0, + 635.0, + 1407.0, + 669.0, + 295.0, + 669.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 665.0, + 1104.0, + 665.0, + 1104.0, + 697.0, + 296.0, + 697.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1164.0, + 665.0, + 1404.0, + 665.0, + 1404.0, + 697.0, + 1164.0, + 697.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 696.0, + 1408.0, + 696.0, + 1408.0, + 730.0, + 294.0, + 730.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 725.0, + 1411.0, + 725.0, + 1411.0, + 763.0, + 291.0, + 763.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 756.0, + 1406.0, + 756.0, + 1406.0, + 794.0, + 292.0, + 794.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 787.0, + 1404.0, + 787.0, + 1404.0, + 823.0, + 294.0, + 823.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 818.0, + 1406.0, + 818.0, + 1406.0, + 852.0, + 292.0, + 852.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 846.0, + 1407.0, + 846.0, + 1407.0, + 884.0, + 294.0, + 884.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 876.0, + 296.0, + 876.0, + 296.0, + 913.0, + 292.0, + 913.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 353.0, + 876.0, + 1406.0, + 876.0, + 1406.0, + 913.0, + 353.0, + 913.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 909.0, + 1406.0, + 909.0, + 1406.0, + 945.0, + 295.0, + 945.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 941.0, + 757.0, + 941.0, + 757.0, + 973.0, + 296.0, + 973.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1821.0, + 1406.0, + 1821.0, + 1406.0, + 1855.0, + 295.0, + 1855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1850.0, + 1406.0, + 1850.0, + 1406.0, + 1887.0, + 294.0, + 1887.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1880.0, + 408.0, + 1880.0, + 408.0, + 1915.0, + 294.0, + 1915.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 538.0, + 1880.0, + 1406.0, + 1880.0, + 1406.0, + 1915.0, + 538.0, + 1915.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1910.0, + 1406.0, + 1910.0, + 1406.0, + 1947.0, + 294.0, + 1947.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1939.0, + 1409.0, + 1939.0, + 1409.0, + 1981.0, + 292.0, + 1981.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1973.0, + 1406.0, + 1973.0, + 1406.0, + 2008.0, + 295.0, + 2008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 2000.0, + 446.0, + 2000.0, + 446.0, + 2036.0, + 294.0, + 2036.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 13, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1472, + 1405, + 1472, + 1405, + 1840, + 297, + 1840 + ], + "score": 0.983 + }, + { + "category_id": 1, + "poly": [ + 297, + 873, + 1404, + 873, + 1404, + 1058, + 297, + 1058 + ], + "score": 0.979 + }, + { + "category_id": 5, + "poly": [ + 497, + 1223, + 1189, + 1223, + 1189, + 1369, + 497, + 1369 + ], + "score": 0.973, + "html": "
Random colorsIDMIDM (PAD, fixed π s)IDM (PAD)
Walker, walk406±29452±38468±47
Walker, stand743±37802±41797±46
Cartpole, swingup585±73623±57630±63
" + }, + { + "category_id": 3, + "poly": [ + 332, + 227, + 1358, + 227, + 1358, + 567, + 332, + 567 + ], + "score": 0.969 + }, + { + "category_id": 1, + "poly": [ + 298, + 1942, + 1402, + 1942, + 1402, + 2034, + 298, + 2034 + ], + "score": 0.966 + }, + { + "category_id": 4, + "poly": [ + 296, + 580, + 1407, + 580, + 1407, + 764, + 296, + 764 + ], + "score": 0.959 + }, + { + "category_id": 6, + "poly": [ + 300, + 1083, + 1406, + 1083, + 1406, + 1208, + 300, + 1208 + ], + "score": 0.933 + }, + { + "category_id": 0, + "poly": [ + 297, + 1417, + 1020, + 1417, + 1020, + 1450, + 297, + 1450 + ], + "score": 0.919 + }, + { + "category_id": 0, + "poly": [ + 298, + 1887, + 1044, + 1887, + 1044, + 1919, + 298, + 1919 + ], + "score": 0.908 + }, + { + "category_id": 0, + "poly": [ + 299, + 817, + 1044, + 817, + 1044, + 851, + 299, + 851 + ], + "score": 0.906 + }, + { + "category_id": 2, + "poly": [ + 299, + 76, + 812, + 76, + 812, + 104, + 299, + 104 + ], + "score": 0.886 + }, + { + "category_id": 2, + "poly": [ + 836, + 2088, + 863, + 2088, + 863, + 2112, + 836, + 2112 + ], + "score": 0.842 + }, + { + "category_id": 13, + "poly": [ + 658, + 642, + 713, + 642, + 713, + 671, + 658, + 671 + ], + "score": 0.88, + "latex": "9 5 \\%" + }, + { + "category_id": 13, + "poly": [ + 1177, + 1151, + 1208, + 1151, + 1208, + 1176, + 1177, + 1176 + ], + "score": 0.87, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 1151, + 1120, + 1182, + 1120, + 1182, + 1146, + 1151, + 1146 + ], + "score": 0.87, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 1092, + 1150, + 1123, + 1150, + 1123, + 1176, + 1092, + 1176 + ], + "score": 0.86, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 1092, + 879, + 1123, + 879, + 1123, + 905, + 1092, + 905 + ], + "score": 0.86, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 551, + 1629, + 584, + 1629, + 584, + 1656, + 551, + 1656 + ], + "score": 0.86, + "latex": "\\pi _ { a }" + }, + { + "category_id": 13, + "poly": [ + 866, + 1001, + 897, + 1001, + 897, + 1027, + 866, + 1027 + ], + "score": 0.86, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 921, + 909, + 951, + 909, + 951, + 935, + 921, + 935 + ], + "score": 0.86, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 1323, + 970, + 1354, + 970, + 1354, + 997, + 1323, + 997 + ], + "score": 0.86, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 465, + 1630, + 497, + 1630, + 497, + 1656, + 465, + 1656 + ], + "score": 0.85, + "latex": "\\pi _ { e }" + }, + { + "category_id": 13, + "poly": [ + 492, + 822, + 527, + 822, + 527, + 851, + 492, + 851 + ], + "score": 0.85, + "latex": "\\pi _ { s }" + }, + { + "category_id": 13, + "poly": [ + 659, + 906, + 681, + 906, + 681, + 932, + 659, + 932 + ], + "score": 0.78, + "latex": "L" + }, + { + "category_id": 13, + "poly": [ + 346, + 611, + 475, + 611, + 475, + 640, + 346, + 640 + ], + "score": 0.77, + "latex": "\\mathrm { S A C + I D M }" + }, + { + "category_id": 13, + "poly": [ + 735, + 1119, + 767, + 1119, + 767, + 1147, + 735, + 1147 + ], + "score": 0.74, + "latex": "\\pi _ { s . }" + }, + { + "category_id": 13, + "poly": [ + 297, + 672, + 423, + 672, + 423, + 702, + 297, + 702 + ], + "score": 0.73, + "latex": "\\mathrm { S A C + I D M }" + }, + { + "category_id": 13, + "poly": [ + 1225, + 581, + 1336, + 581, + 1336, + 612, + 1225, + 612 + ], + "score": 0.5, + "latex": "S A C ( D R )" + }, + { + "category_id": 13, + "poly": [ + 396, + 1146, + 533, + 1146, + 533, + 1177, + 396, + 1177 + ], + "score": 0.4, + "latex": "I D M \\left( P A D \\right)" + }, + { + "category_id": 15, + "poly": [ + 457.0, + 231.0, + 621.0, + 231.0, + 621.0, + 263.0, + 457.0, + 263.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 806.0, + 233.0, + 933.0, + 233.0, + 933.0, + 259.0, + 806.0, + 259.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1141.0, + 233.0, + 1259.0, + 233.0, + 1259.0, + 259.0, + 1141.0, + 259.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 350.0, + 247.0, + 393.0, + 247.0, + 393.0, + 271.0, + 350.0, + 271.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 292.0, + 391.0, + 292.0, + 391.0, + 313.0, + 358.0, + 313.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 331.0, + 305.0, + 357.0, + 305.0, + 357.0, + 433.0, + 331.0, + 433.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 336.0, + 391.0, + 336.0, + 391.0, + 358.0, + 358.0, + 358.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 381.0, + 391.0, + 381.0, + 391.0, + 402.0, + 358.0, + 402.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 358.0, + 425.0, + 392.0, + 425.0, + 392.0, + 446.0, + 358.0, + 446.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 368.0, + 465.0, + 412.0, + 465.0, + 412.0, + 506.0, + 368.0, + 506.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 436.0, + 479.0, + 467.0, + 479.0, + 467.0, + 501.0, + 436.0, + 501.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 494.0, + 479.0, + 526.0, + 479.0, + 526.0, + 501.0, + 494.0, + 501.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 553.0, + 479.0, + 585.0, + 479.0, + 585.0, + 501.0, + 553.0, + 501.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 609.0, + 478.0, + 644.0, + 478.0, + 644.0, + 504.0, + 609.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 668.0, + 478.0, + 741.0, + 478.0, + 741.0, + 505.0, + 668.0, + 505.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 765.0, + 481.0, + 797.0, + 481.0, + 797.0, + 502.0, + 765.0, + 502.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 824.0, + 478.0, + 857.0, + 478.0, + 857.0, + 504.0, + 824.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 883.0, + 479.0, + 915.0, + 479.0, + 915.0, + 501.0, + 883.0, + 501.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 940.0, + 478.0, + 974.0, + 478.0, + 974.0, + 504.0, + 940.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 997.0, + 478.0, + 1069.0, + 478.0, + 1069.0, + 501.0, + 997.0, + 501.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1096.0, + 481.0, + 1128.0, + 481.0, + 1128.0, + 502.0, + 1096.0, + 502.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1152.0, + 478.0, + 1187.0, + 478.0, + 1187.0, + 504.0, + 1152.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1212.0, + 479.0, + 1245.0, + 479.0, + 1245.0, + 501.0, + 1212.0, + 501.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1271.0, + 478.0, + 1305.0, + 478.0, + 1305.0, + 504.0, + 1271.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1329.0, + 478.0, + 1363.0, + 478.0, + 1363.0, + 504.0, + 1329.0, + 504.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 450.0, + 494.0, + 626.0, + 494.0, + 626.0, + 520.0, + 450.0, + 520.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 782.0, + 496.0, + 958.0, + 496.0, + 958.0, + 521.0, + 782.0, + 521.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1111.0, + 496.0, + 1287.0, + 496.0, + 1287.0, + 521.0, + 1111.0, + 521.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 684.0, + 535.0, + 736.0, + 535.0, + 736.0, + 567.0, + 684.0, + 567.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 800.0, + 536.0, + 892.0, + 536.0, + 892.0, + 565.0, + 800.0, + 565.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 958.0, + 536.0, + 1054.0, + 536.0, + 1054.0, + 564.0, + 958.0, + 564.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 579.0, + 1224.0, + 579.0, + 1224.0, + 615.0, + 294.0, + 615.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1337.0, + 579.0, + 1409.0, + 579.0, + 1409.0, + 615.0, + 1337.0, + 615.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 610.0, + 345.0, + 610.0, + 345.0, + 644.0, + 294.0, + 644.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 476.0, + 610.0, + 1405.0, + 610.0, + 1405.0, + 644.0, + 476.0, + 644.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 640.0, + 657.0, + 640.0, + 657.0, + 676.0, + 294.0, + 676.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 714.0, + 640.0, + 1407.0, + 640.0, + 1407.0, + 676.0, + 714.0, + 676.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 669.0, + 296.0, + 669.0, + 296.0, + 706.0, + 293.0, + 706.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 424.0, + 669.0, + 1405.0, + 669.0, + 1405.0, + 706.0, + 424.0, + 706.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 702.0, + 1404.0, + 702.0, + 1404.0, + 734.0, + 295.0, + 734.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 733.0, + 983.0, + 733.0, + 983.0, + 768.0, + 295.0, + 768.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1086.0, + 1408.0, + 1086.0, + 1408.0, + 1115.0, + 295.0, + 1115.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1116.0, + 734.0, + 1116.0, + 734.0, + 1149.0, + 295.0, + 1149.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 768.0, + 1116.0, + 1150.0, + 1116.0, + 1150.0, + 1149.0, + 768.0, + 1149.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1183.0, + 1116.0, + 1408.0, + 1116.0, + 1408.0, + 1149.0, + 1183.0, + 1149.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1147.0, + 395.0, + 1147.0, + 395.0, + 1179.0, + 295.0, + 1179.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 534.0, + 1147.0, + 1091.0, + 1147.0, + 1091.0, + 1179.0, + 534.0, + 1179.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1124.0, + 1147.0, + 1176.0, + 1147.0, + 1176.0, + 1179.0, + 1124.0, + 1179.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1209.0, + 1147.0, + 1406.0, + 1147.0, + 1406.0, + 1179.0, + 1209.0, + 1179.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1177.0, + 795.0, + 1177.0, + 795.0, + 1210.0, + 296.0, + 1210.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1412.0, + 1022.0, + 1412.0, + 1022.0, + 1456.0, + 294.0, + 1456.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1881.0, + 1047.0, + 1881.0, + 1047.0, + 1926.0, + 293.0, + 1926.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 810.0, + 491.0, + 810.0, + 491.0, + 859.0, + 292.0, + 859.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 528.0, + 810.0, + 1050.0, + 810.0, + 1050.0, + 859.0, + 528.0, + 859.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2085.0, + 867.0, + 2085.0, + 867.0, + 2124.0, + 832.0, + 2124.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1472.0, + 1406.0, + 1472.0, + 1406.0, + 1509.0, + 294.0, + 1509.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1501.0, + 1406.0, + 1501.0, + 1406.0, + 1541.0, + 291.0, + 1541.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1535.0, + 1405.0, + 1535.0, + 1405.0, + 1569.0, + 295.0, + 1569.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1565.0, + 1406.0, + 1565.0, + 1406.0, + 1599.0, + 292.0, + 1599.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1596.0, + 1405.0, + 1596.0, + 1405.0, + 1630.0, + 295.0, + 1630.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1622.0, + 464.0, + 1622.0, + 464.0, + 1663.0, + 292.0, + 1663.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 498.0, + 1622.0, + 550.0, + 1622.0, + 550.0, + 1663.0, + 498.0, + 1663.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 585.0, + 1622.0, + 1406.0, + 1622.0, + 1406.0, + 1663.0, + 585.0, + 1663.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1655.0, + 1406.0, + 1655.0, + 1406.0, + 1693.0, + 294.0, + 1693.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1684.0, + 1407.0, + 1684.0, + 1407.0, + 1722.0, + 294.0, + 1722.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1718.0, + 1405.0, + 1718.0, + 1405.0, + 1752.0, + 294.0, + 1752.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1744.0, + 1406.0, + 1744.0, + 1406.0, + 1785.0, + 292.0, + 1785.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1775.0, + 1406.0, + 1775.0, + 1406.0, + 1814.0, + 292.0, + 1814.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1807.0, + 1227.0, + 1807.0, + 1227.0, + 1844.0, + 294.0, + 1844.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 297.0, + 876.0, + 1091.0, + 876.0, + 1091.0, + 904.0, + 297.0, + 904.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1124.0, + 876.0, + 1402.0, + 876.0, + 1402.0, + 904.0, + 1124.0, + 904.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 903.0, + 658.0, + 903.0, + 658.0, + 939.0, + 295.0, + 939.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 682.0, + 903.0, + 920.0, + 903.0, + 920.0, + 939.0, + 682.0, + 939.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 952.0, + 903.0, + 1406.0, + 903.0, + 1406.0, + 939.0, + 952.0, + 939.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 935.0, + 1406.0, + 935.0, + 1406.0, + 971.0, + 292.0, + 971.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 964.0, + 1322.0, + 964.0, + 1322.0, + 1000.0, + 295.0, + 1000.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1355.0, + 964.0, + 1404.0, + 964.0, + 1404.0, + 1000.0, + 1355.0, + 1000.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 995.0, + 865.0, + 995.0, + 865.0, + 1033.0, + 292.0, + 1033.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 898.0, + 995.0, + 1405.0, + 995.0, + 1405.0, + 1033.0, + 898.0, + 1033.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1027.0, + 623.0, + 1027.0, + 623.0, + 1062.0, + 295.0, + 1062.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1942.0, + 1404.0, + 1942.0, + 1404.0, + 1976.0, + 296.0, + 1976.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1971.0, + 1406.0, + 1971.0, + 1406.0, + 2007.0, + 293.0, + 2007.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 2004.0, + 511.0, + 2004.0, + 511.0, + 2035.0, + 296.0, + 2035.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 14, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 297, + 1787, + 1404, + 1787, + 1404, + 2035, + 297, + 2035 + ], + "score": 0.98 + }, + { + "category_id": 1, + "poly": [ + 297, + 1161, + 1406, + 1161, + 1406, + 1286, + 297, + 1286 + ], + "score": 0.977 + }, + { + "category_id": 3, + "poly": [ + 303, + 608, + 1396, + 608, + 1396, + 919, + 303, + 919 + ], + "score": 0.974 + }, + { + "category_id": 5, + "poly": [ + 400, + 423, + 1286, + 423, + 1286, + 565, + 400, + 565 + ], + "score": 0.968, + "html": "
Random colorsIDMIDM (PAD)1 episode10 episodes100 episodes
Walker, walk406±29468±47395±78489±104561±62
Walker, stand743±37797±46661±65728±44784±31
Cartpole, swingup585±73630±63538±53605±51650±58
" + }, + { + "category_id": 3, + "poly": [ + 460, + 1316, + 1240, + 1316, + 1240, + 1608, + 460, + 1608 + ], + "score": 0.964 + }, + { + "category_id": 4, + "poly": [ + 299, + 1628, + 1405, + 1628, + 1405, + 1724, + 299, + 1724 + ], + "score": 0.93 + }, + { + "category_id": 0, + "poly": [ + 300, + 1088, + 743, + 1088, + 743, + 1124, + 300, + 1124 + ], + "score": 0.904 + }, + { + "category_id": 2, + "poly": [ + 299, + 75, + 813, + 75, + 813, + 105, + 299, + 105 + ], + "score": 0.901 + }, + { + "category_id": 6, + "poly": [ + 950, + 395, + 1190, + 395, + 1190, + 422, + 950, + 422 + ], + "score": 0.861 + }, + { + "category_id": 2, + "poly": [ + 836, + 2088, + 864, + 2088, + 864, + 2113, + 836, + 2113 + ], + "score": 0.858 + }, + { + "category_id": 4, + "poly": [ + 297, + 943, + 1400, + 943, + 1400, + 1037, + 297, + 1037 + ], + "score": 0.831 + }, + { + "category_id": 1, + "poly": [ + 296, + 223, + 1406, + 223, + 1406, + 377, + 296, + 377 + ], + "score": 0.819 + }, + { + "category_id": 6, + "poly": [ + 296, + 223, + 1406, + 223, + 1406, + 377, + 296, + 377 + ], + "score": 0.268 + }, + { + "category_id": 1, + "poly": [ + 297, + 943, + 1400, + 943, + 1400, + 1037, + 297, + 1037 + ], + "score": 0.138 + }, + { + "category_id": 13, + "poly": [ + 483, + 1943, + 647, + 1943, + 647, + 1972, + 483, + 1972 + ], + "score": 0.91, + "latex": "3 2 \\times 2 1 \\times 2 1" + }, + { + "category_id": 13, + "poly": [ + 298, + 1852, + 425, + 1852, + 425, + 1880, + 298, + 1880 + ], + "score": 0.91, + "latex": "9 \\times 8 4 \\times 8 4" + }, + { + "category_id": 13, + "poly": [ + 519, + 1821, + 640, + 1821, + 640, + 1850, + 519, + 1850 + ], + "score": 0.91, + "latex": "1 0 0 \\times 1 0 0" + }, + { + "category_id": 13, + "poly": [ + 1239, + 1943, + 1402, + 1943, + 1402, + 1972, + 1239, + 1972 + ], + "score": 0.9, + "latex": "3 2 \\times 2 5 \\times 2 5" + }, + { + "category_id": 13, + "poly": [ + 1106, + 1660, + 1226, + 1660, + 1226, + 1689, + 1106, + 1689 + ], + "score": 0.9, + "latex": "1 0 0 \\times 1 0 0" + }, + { + "category_id": 13, + "poly": [ + 813, + 1821, + 906, + 1821, + 906, + 1850, + 813, + 1850 + ], + "score": 0.9, + "latex": "8 4 \\times 8 4" + }, + { + "category_id": 13, + "poly": [ + 357, + 2004, + 390, + 2004, + 390, + 2031, + 357, + 2031 + ], + "score": 0.86, + "latex": "\\pi ^ { a }" + }, + { + "category_id": 13, + "poly": [ + 761, + 1662, + 792, + 1662, + 792, + 1688, + 761, + 1688 + ], + "score": 0.86, + "latex": "\\pi ^ { e }" + }, + { + "category_id": 13, + "poly": [ + 380, + 1661, + 413, + 1661, + 413, + 1688, + 380, + 1688 + ], + "score": 0.86, + "latex": "\\pi ^ { a }" + }, + { + "category_id": 13, + "poly": [ + 669, + 1974, + 701, + 1974, + 701, + 2000, + 669, + 2000 + ], + "score": 0.86, + "latex": "\\pi ^ { e }" + }, + { + "category_id": 13, + "poly": [ + 1274, + 1975, + 1306, + 1975, + 1306, + 2001, + 1274, + 2001 + ], + "score": 0.86, + "latex": "\\pi ^ { s }" + }, + { + "category_id": 13, + "poly": [ + 1369, + 1883, + 1401, + 1883, + 1401, + 1909, + 1369, + 1909 + ], + "score": 0.85, + "latex": "\\pi ^ { e }" + }, + { + "category_id": 13, + "poly": [ + 297, + 1662, + 329, + 1662, + 329, + 1688, + 297, + 1688 + ], + "score": 0.85, + "latex": "\\pi ^ { s }" + }, + { + "category_id": 13, + "poly": [ + 305, + 1821, + 378, + 1821, + 378, + 1850, + 305, + 1850 + ], + "score": 0.84, + "latex": "k = 3 ," + }, + { + "category_id": 15, + "poly": [ + 772.0, + 740.0, + 792.0, + 740.0, + 792.0, + 764.0, + 772.0, + 764.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 354.0, + 884.0, + 519.0, + 884.0, + 519.0, + 921.0, + 354.0, + 921.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 606.0, + 884.0, + 814.0, + 884.0, + 814.0, + 919.0, + 606.0, + 919.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 903.0, + 884.0, + 1068.0, + 884.0, + 1068.0, + 920.0, + 903.0, + 920.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1173.0, + 884.0, + 1349.0, + 884.0, + 1349.0, + 921.0, + 1173.0, + 921.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 933.0, + 1319.0, + 944.0, + 1319.0, + 944.0, + 1333.0, + 933.0, + 1333.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 524.0, + 1337.0, + 755.0, + 1337.0, + 755.0, + 1369.0, + 524.0, + 1369.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1078.0, + 1382.0, + 1240.0, + 1382.0, + 1240.0, + 1412.0, + 1078.0, + 1412.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 781.0, + 1406.0, + 789.0, + 1406.0, + 789.0, + 1415.0, + 781.0, + 1415.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 838.0, + 1401.0, + 882.0, + 1401.0, + 882.0, + 1422.0, + 838.0, + 1422.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 972.0, + 1397.0, + 1003.0, + 1397.0, + 1003.0, + 1424.0, + 972.0, + 1424.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1075.0, + 1406.0, + 1186.0, + 1406.0, + 1186.0, + 1441.0, + 1075.0, + 1441.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 653.0, + 1449.0, + 697.0, + 1449.0, + 697.0, + 1470.0, + 653.0, + 1470.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 975.0, + 1466.0, + 1001.0, + 1466.0, + 1001.0, + 1490.0, + 975.0, + 1490.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1017.0, + 1457.0, + 1154.0, + 1457.0, + 1154.0, + 1499.0, + 1017.0, + 1499.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 837.0, + 1497.0, + 881.0, + 1497.0, + 881.0, + 1518.0, + 837.0, + 1518.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 459.0, + 1517.0, + 588.0, + 1517.0, + 588.0, + 1552.0, + 459.0, + 1552.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 658.0, + 1515.0, + 694.0, + 1515.0, + 694.0, + 1549.0, + 658.0, + 1549.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 926.0, + 1524.0, + 939.0, + 1524.0, + 939.0, + 1538.0, + 926.0, + 1538.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 971.0, + 1516.0, + 1012.0, + 1516.0, + 1012.0, + 1550.0, + 971.0, + 1550.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1017.0, + 1516.0, + 1148.0, + 1516.0, + 1148.0, + 1549.0, + 1017.0, + 1549.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 927.0, + 1578.0, + 945.0, + 1578.0, + 945.0, + 1596.0, + 927.0, + 1596.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 912.0, + 1588.0, + 932.0, + 1588.0, + 932.0, + 1608.0, + 912.0, + 1608.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 699.0, + 1378.0, + 719.0, + 1378.0, + 719.0, + 1390.5, + 699.0, + 1390.5 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1625.0, + 1409.0, + 1625.0, + 1409.0, + 1666.0, + 292.0, + 1666.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 330.0, + 1658.0, + 379.0, + 1658.0, + 379.0, + 1692.0, + 330.0, + 1692.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 414.0, + 1658.0, + 760.0, + 1658.0, + 760.0, + 1692.0, + 414.0, + 1692.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 793.0, + 1658.0, + 1105.0, + 1658.0, + 1105.0, + 1692.0, + 793.0, + 1692.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1227.0, + 1658.0, + 1407.0, + 1658.0, + 1407.0, + 1692.0, + 1227.0, + 1692.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1691.0, + 1187.0, + 1691.0, + 1187.0, + 1725.0, + 295.0, + 1725.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 1084.0, + 748.0, + 1084.0, + 748.0, + 1129.0, + 291.0, + 1129.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 948.0, + 392.0, + 1192.0, + 392.0, + 1192.0, + 426.0, + 948.0, + 426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 831.0, + 2085.0, + 868.0, + 2085.0, + 868.0, + 2124.0, + 831.0, + 2124.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 940.0, + 1405.0, + 940.0, + 1405.0, + 979.0, + 291.0, + 979.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 973.0, + 1404.0, + 973.0, + 1404.0, + 1008.0, + 293.0, + 1008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1005.0, + 924.0, + 1005.0, + 924.0, + 1038.0, + 295.0, + 1038.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 225.0, + 1406.0, + 225.0, + 1406.0, + 258.0, + 295.0, + 258.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 255.0, + 1407.0, + 255.0, + 1407.0, + 288.0, + 295.0, + 288.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 284.0, + 1404.0, + 284.0, + 1404.0, + 321.0, + 295.0, + 321.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 313.0, + 1407.0, + 313.0, + 1407.0, + 354.0, + 293.0, + 354.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 348.0, + 690.0, + 348.0, + 690.0, + 381.0, + 296.0, + 381.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1788.0, + 1405.0, + 1788.0, + 1405.0, + 1823.0, + 294.0, + 1823.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1819.0, + 304.0, + 1819.0, + 304.0, + 1853.0, + 294.0, + 1853.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 379.0, + 1819.0, + 518.0, + 1819.0, + 518.0, + 1853.0, + 379.0, + 1853.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 641.0, + 1819.0, + 812.0, + 1819.0, + 812.0, + 1853.0, + 641.0, + 1853.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 907.0, + 1819.0, + 1404.0, + 1819.0, + 1404.0, + 1853.0, + 907.0, + 1853.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1848.0, + 297.0, + 1848.0, + 297.0, + 1886.0, + 292.0, + 1886.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 426.0, + 1848.0, + 1404.0, + 1848.0, + 1404.0, + 1886.0, + 426.0, + 1886.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1881.0, + 1368.0, + 1881.0, + 1368.0, + 1916.0, + 292.0, + 1916.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1402.0, + 1881.0, + 1405.0, + 1881.0, + 1405.0, + 1916.0, + 1402.0, + 1916.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1912.0, + 1406.0, + 1912.0, + 1406.0, + 1946.0, + 295.0, + 1946.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1941.0, + 482.0, + 1941.0, + 482.0, + 1975.0, + 292.0, + 1975.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 648.0, + 1941.0, + 1238.0, + 1941.0, + 1238.0, + 1975.0, + 648.0, + 1975.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1403.0, + 1941.0, + 1406.0, + 1941.0, + 1406.0, + 1975.0, + 1403.0, + 1975.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1971.0, + 668.0, + 1971.0, + 668.0, + 2005.0, + 292.0, + 2005.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 702.0, + 1971.0, + 1273.0, + 1971.0, + 1273.0, + 2005.0, + 702.0, + 2005.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 1307.0, + 1971.0, + 1406.0, + 1971.0, + 1406.0, + 2005.0, + 1307.0, + 2005.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 2003.0, + 356.0, + 2003.0, + 356.0, + 2037.0, + 294.0, + 2037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 391.0, + 2003.0, + 1405.0, + 2003.0, + 1405.0, + 2037.0, + 391.0, + 2037.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1161.0, + 1406.0, + 1161.0, + 1406.0, + 1194.0, + 295.0, + 1194.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1192.0, + 1408.0, + 1192.0, + 1408.0, + 1225.0, + 295.0, + 1225.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1223.0, + 1406.0, + 1223.0, + 1406.0, + 1256.0, + 295.0, + 1256.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1250.0, + 1304.0, + 1250.0, + 1304.0, + 1288.0, + 294.0, + 1288.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 225.0, + 1406.0, + 225.0, + 1406.0, + 258.0, + 295.0, + 258.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 255.0, + 1407.0, + 255.0, + 1407.0, + 288.0, + 295.0, + 288.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 284.0, + 1404.0, + 284.0, + 1404.0, + 321.0, + 295.0, + 321.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 313.0, + 1407.0, + 313.0, + 1407.0, + 354.0, + 293.0, + 354.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 348.0, + 690.0, + 348.0, + 690.0, + 381.0, + 296.0, + 381.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 291.0, + 940.0, + 1405.0, + 940.0, + 1405.0, + 979.0, + 291.0, + 979.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 973.0, + 1404.0, + 973.0, + 1404.0, + 1008.0, + 293.0, + 1008.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1005.0, + 924.0, + 1005.0, + 924.0, + 1038.0, + 295.0, + 1038.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 15, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 1499, + 1404, + 1499, + 1404, + 1775, + 298, + 1775 + ], + "score": 0.98 + }, + { + "category_id": 5, + "poly": [ + 840, + 312, + 1387, + 312, + 1387, + 813, + 840, + 813 + ], + "score": 0.979, + "html": "
HyperparameterValue
Frame rendering3×100×100
Frame after crop3×84×84
Stacked frames3
Action repeat4
Discount factor y0.99
Episode length1,000
Learning algorithmAdvantage Actor-Critic
Self-supervised taskRotation Prediction
Number of training episodes1,000 (dom. rand.) 500 (otherwise)
Number of processes20
OptimizerAdam (β=0.9,β=0.999)
Learning rate1e-4
Learning rate (test-time)1e-5
Batch size20
32
Batch size (test-time) π,πloss coefficient0.5
1
πe,πloss coefficient (test-time)1
πe,π update freq. πe,π update freq.(test-time)1
" + }, + { + "category_id": 1, + "poly": [ + 298, + 1009, + 1404, + 1009, + 1404, + 1223, + 298, + 1223 + ], + "score": 0.979 + }, + { + "category_id": 5, + "poly": [ + 296, + 311, + 819, + 311, + 819, + 834, + 296, + 834 + ], + "score": 0.978, + "html": "
HyperparameterValue
Frame rendering3 ×100×100
Frame after crop3×84×84
Stacked frames3
Action repeat2 (finger)
8 (cartpole) 4(otherwise)
Discount factor y0.99
Episode length1,000
Learning algorithmSoft Actor-Critic
Self-supervised taskInverse Dynamics Model
Number of training steps500,000
Replay buffer size500,000
Optimizer(πe,πä,π)Adam (β=0.9,β=0.999)
Optimizer (α)Adam(β=0.5,β=0.999)
Learning rate (πe,π,π$)3e-4 (cheetah)
Learning rate (α)le-3 (otherwise) 1e-4
Batch size128
Batch size (test-time)32
πe,π update freq.2
πe,π update freq. (test-time)1
" + }, + { + "category_id": 1, + "poly": [ + 298, + 1819, + 1403, + 1819, + 1403, + 2033, + 298, + 2033 + ], + "score": 0.978 + }, + { + "category_id": 1, + "poly": [ + 298, + 1271, + 1404, + 1271, + 1404, + 1453, + 298, + 1453 + ], + "score": 0.977 + }, + { + "category_id": 6, + "poly": [ + 296, + 225, + 820, + 225, + 820, + 285, + 296, + 285 + ], + "score": 0.941 + }, + { + "category_id": 1, + "poly": [ + 298, + 902, + 1401, + 902, + 1401, + 964, + 298, + 964 + ], + "score": 0.923 + }, + { + "category_id": 6, + "poly": [ + 840, + 225, + 1389, + 225, + 1389, + 286, + 840, + 286 + ], + "score": 0.914 + }, + { + "category_id": 2, + "poly": [ + 300, + 76, + 812, + 76, + 812, + 104, + 300, + 104 + ], + "score": 0.892 + }, + { + "category_id": 2, + "poly": [ + 836, + 2088, + 863, + 2088, + 863, + 2112, + 836, + 2112 + ], + "score": 0.842 + }, + { + "category_id": 13, + "poly": [ + 1296, + 1713, + 1403, + 1713, + 1403, + 1746, + 1296, + 1746 + ], + "score": 0.91, + "latex": "\\left( \\mathbf { s } _ { t } , \\mathbf { s } _ { t + 1 } \\right)" + }, + { + "category_id": 13, + "poly": [ + 620, + 1622, + 739, + 1622, + 739, + 1651, + 620, + 1651 + ], + "score": 0.91, + "latex": "1 0 0 \\times 1 0 0" + }, + { + "category_id": 13, + "poly": [ + 769, + 1622, + 860, + 1622, + 860, + 1651, + 769, + 1651 + ], + "score": 0.9, + "latex": "8 4 \\times 8 4" + }, + { + "category_id": 13, + "poly": [ + 854, + 1362, + 923, + 1362, + 923, + 1390, + 854, + 1390 + ], + "score": 0.9, + "latex": "k = 3" + }, + { + "category_id": 13, + "poly": [ + 353, + 1392, + 422, + 1392, + 422, + 1420, + 353, + 1420 + ], + "score": 0.89, + "latex": "k = 1" + }, + { + "category_id": 13, + "poly": [ + 386, + 624, + 484, + 624, + 484, + 647, + 386, + 647 + ], + "score": 0.85, + "latex": "( \\pi ^ { e } , \\pi ^ { a } , \\pi ^ { s } )" + }, + { + "category_id": 13, + "poly": [ + 541, + 1749, + 569, + 1749, + 569, + 1774, + 541, + 1774 + ], + "score": 0.85, + "latex": "\\mathbf { a } _ { t }" + }, + { + "category_id": 13, + "poly": [ + 413, + 669, + 512, + 669, + 512, + 693, + 413, + 693 + ], + "score": 0.85, + "latex": "( \\pi ^ { e } , \\pi ^ { a } , \\pi ^ { s } )" + }, + { + "category_id": 13, + "poly": [ + 627, + 647, + 815, + 647, + 815, + 669, + 627, + 669 + ], + "score": 0.81, + "latex": "( \\beta _ { 1 } = 0 . 5 , \\beta _ { 2 } = 0 . 9 9 9 )" + }, + { + "category_id": 13, + "poly": [ + 298, + 805, + 351, + 805, + 351, + 827, + 298, + 827 + ], + "score": 0.73, + "latex": "\\pi ^ { e } , \\pi ^ { s }" + }, + { + "category_id": 13, + "poly": [ + 298, + 782, + 320, + 782, + 320, + 802, + 298, + 802 + ], + "score": 0.72, + "latex": "\\pi ^ { e }" + }, + { + "category_id": 13, + "poly": [ + 842, + 714, + 865, + 714, + 865, + 734, + 842, + 734 + ], + "score": 0.69, + "latex": "\\pi ^ { e }" + }, + { + "category_id": 13, + "poly": [ + 873, + 714, + 896, + 714, + 896, + 734, + 873, + 734 + ], + "score": 0.66, + "latex": "\\pi ^ { s }" + }, + { + "category_id": 13, + "poly": [ + 627, + 623, + 815, + 623, + 815, + 646, + 627, + 646 + ], + "score": 0.64, + "latex": "( \\beta _ { 1 } = 0 . 9 , \\beta _ { 2 } = 0 . 9 9 9 )" + }, + { + "category_id": 13, + "poly": [ + 842, + 781, + 896, + 781, + 896, + 804, + 842, + 804 + ], + "score": 0.61, + "latex": "\\pi ^ { e } , \\pi ^ { s }" + }, + { + "category_id": 13, + "poly": [ + 842, + 759, + 896, + 759, + 896, + 780, + 842, + 780 + ], + "score": 0.57, + "latex": "\\pi ^ { e } , \\pi ^ { s }" + }, + { + "category_id": 13, + "poly": [ + 842, + 736, + 897, + 736, + 897, + 758, + 842, + 758 + ], + "score": 0.56, + "latex": "\\pi ^ { e } , \\pi ^ { s }" + }, + { + "category_id": 13, + "poly": [ + 328, + 782, + 352, + 782, + 352, + 802, + 328, + 802 + ], + "score": 0.53, + "latex": "\\pi ^ { s }" + }, + { + "category_id": 13, + "poly": [ + 1195, + 600, + 1380, + 600, + 1380, + 625, + 1195, + 625 + ], + "score": 0.39, + "latex": "( \\beta _ { 1 } = 0 . 9 , \\beta _ { 2 } = 0 . 9 9 9 )" + }, + { + "category_id": 13, + "poly": [ + 429, + 493, + 445, + 493, + 445, + 513, + 429, + 513 + ], + "score": 0.38, + "latex": "\\gamma" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 223.0, + 824.0, + 223.0, + 824.0, + 257.0, + 295.0, + 257.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 298.0, + 254.0, + 674.0, + 254.0, + 674.0, + 286.0, + 298.0, + 286.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 223.0, + 1392.0, + 223.0, + 1392.0, + 258.0, + 839.0, + 258.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 839.0, + 254.0, + 1364.0, + 254.0, + 1364.0, + 288.0, + 839.0, + 288.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 73.0, + 815.0, + 73.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2085.0, + 867.0, + 2085.0, + 867.0, + 2124.0, + 832.0, + 2124.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1500.0, + 1406.0, + 1500.0, + 1406.0, + 1533.0, + 296.0, + 1533.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1530.0, + 1405.0, + 1530.0, + 1405.0, + 1563.0, + 296.0, + 1563.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1557.0, + 1406.0, + 1557.0, + 1406.0, + 1594.0, + 293.0, + 1594.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 1585.0, + 1406.0, + 1585.0, + 1406.0, + 1628.0, + 292.0, + 1628.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1620.0, + 619.0, + 1620.0, + 619.0, + 1655.0, + 293.0, + 1655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 740.0, + 1620.0, + 768.0, + 1620.0, + 768.0, + 1655.0, + 740.0, + 1655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 861.0, + 1620.0, + 1404.0, + 1620.0, + 1404.0, + 1655.0, + 861.0, + 1655.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1647.0, + 1405.0, + 1647.0, + 1405.0, + 1687.0, + 295.0, + 1687.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1681.0, + 1405.0, + 1681.0, + 1405.0, + 1718.0, + 294.0, + 1718.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1710.0, + 1295.0, + 1710.0, + 1295.0, + 1749.0, + 293.0, + 1749.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 1741.0, + 540.0, + 1741.0, + 540.0, + 1779.0, + 295.0, + 1779.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 570.0, + 1741.0, + 581.0, + 1741.0, + 581.0, + 1779.0, + 570.0, + 1779.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1009.0, + 1405.0, + 1009.0, + 1405.0, + 1044.0, + 294.0, + 1044.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1040.0, + 1405.0, + 1040.0, + 1405.0, + 1074.0, + 294.0, + 1074.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1069.0, + 1405.0, + 1069.0, + 1405.0, + 1106.0, + 293.0, + 1106.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1104.0, + 1405.0, + 1104.0, + 1405.0, + 1134.0, + 296.0, + 1134.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1131.0, + 1406.0, + 1131.0, + 1406.0, + 1165.0, + 294.0, + 1165.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1164.0, + 1405.0, + 1164.0, + 1405.0, + 1194.0, + 296.0, + 1194.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1191.0, + 1076.0, + 1191.0, + 1076.0, + 1228.0, + 293.0, + 1228.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1821.0, + 1405.0, + 1821.0, + 1405.0, + 1855.0, + 294.0, + 1855.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1850.0, + 1405.0, + 1850.0, + 1405.0, + 1884.0, + 293.0, + 1884.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1882.0, + 1405.0, + 1882.0, + 1405.0, + 1916.0, + 296.0, + 1916.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1913.0, + 1405.0, + 1913.0, + 1405.0, + 1946.0, + 293.0, + 1946.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1942.0, + 1405.0, + 1942.0, + 1405.0, + 1980.0, + 293.0, + 1980.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1973.0, + 1405.0, + 1973.0, + 1405.0, + 2011.0, + 293.0, + 2011.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 2005.0, + 577.0, + 2005.0, + 577.0, + 2035.0, + 296.0, + 2035.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1269.0, + 1409.0, + 1269.0, + 1409.0, + 1304.0, + 294.0, + 1304.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1298.0, + 1406.0, + 1298.0, + 1406.0, + 1334.0, + 293.0, + 1334.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 293.0, + 1329.0, + 1406.0, + 1329.0, + 1406.0, + 1365.0, + 293.0, + 1365.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1362.0, + 853.0, + 1362.0, + 853.0, + 1393.0, + 296.0, + 1393.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 924.0, + 1362.0, + 1405.0, + 1362.0, + 1405.0, + 1393.0, + 924.0, + 1393.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 1391.0, + 352.0, + 1391.0, + 352.0, + 1426.0, + 294.0, + 1426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 423.0, + 1391.0, + 1406.0, + 1391.0, + 1406.0, + 1426.0, + 423.0, + 1426.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 1422.0, + 1086.0, + 1422.0, + 1086.0, + 1454.0, + 296.0, + 1454.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 900.0, + 1406.0, + 900.0, + 1406.0, + 936.0, + 296.0, + 936.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 295.0, + 929.0, + 523.0, + 929.0, + 523.0, + 969.0, + 295.0, + 969.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 16, + "width": 1700, + "height": 2200 + } + }, + { + "layout_dets": [ + { + "category_id": 1, + "poly": [ + 298, + 229, + 1403, + 229, + 1403, + 415, + 298, + 415 + ], + "score": 0.976 + }, + { + "category_id": 2, + "poly": [ + 300, + 75, + 813, + 75, + 813, + 104, + 300, + 104 + ], + "score": 0.868 + }, + { + "category_id": 2, + "poly": [ + 836, + 2089, + 863, + 2089, + 863, + 2112, + 836, + 2112 + ], + "score": 0.835 + }, + { + "category_id": 15, + "poly": [ + 295.0, + 72.0, + 815.0, + 72.0, + 815.0, + 108.0, + 295.0, + 108.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 832.0, + 2085.0, + 868.0, + 2085.0, + 868.0, + 2122.0, + 832.0, + 2122.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 231.0, + 1403.0, + 231.0, + 1403.0, + 263.0, + 296.0, + 263.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 294.0, + 262.0, + 1405.0, + 262.0, + 1405.0, + 298.0, + 294.0, + 298.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 293.0, + 1404.0, + 293.0, + 1404.0, + 325.0, + 296.0, + 325.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 319.0, + 1404.0, + 319.0, + 1404.0, + 360.0, + 292.0, + 360.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 292.0, + 350.0, + 1404.0, + 350.0, + 1404.0, + 388.0, + 292.0, + 388.0 + ], + "score": 1.0, + "text": "" + }, + { + "category_id": 15, + "poly": [ + 296.0, + 385.0, + 1016.0, + 385.0, + 1016.0, + 416.0, + 296.0, + 416.0 + ], + "score": 1.0, + "text": "" + } + ], + "page_info": { + "page_no": 17, + "width": 1700, + "height": 2200 + } + } +] \ No newline at end of file diff --git a/vlm/train/B1spAqUp-/0.png b/vlm/train/B1spAqUp-/0.png new file mode 100644 index 0000000000000000000000000000000000000000..3b5b4f42c4d30187e6379199d7dda8b86b2e565d --- /dev/null +++ b/vlm/train/B1spAqUp-/0.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e004405d482e7ac29f1b67e6202279aaddba8080e01754525cfeb2c0d61a2eb5 +size 514868 diff --git a/vlm/train/B1spAqUp-/1.png b/vlm/train/B1spAqUp-/1.png new file mode 100644 index 0000000000000000000000000000000000000000..0d94c8c3edbcad791c4a00a05a029fa080c70f46 --- /dev/null +++ b/vlm/train/B1spAqUp-/1.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c74d54b8d66413f159a99fe0441ae8b03405f25e79b1eddb88fc83d24d82a659 +size 667099 diff --git a/vlm/train/B1spAqUp-/10.png b/vlm/train/B1spAqUp-/10.png new file mode 100644 index 0000000000000000000000000000000000000000..70835a8be306c04cfaf3ca53be11c3f2e2291b31 --- /dev/null +++ b/vlm/train/B1spAqUp-/10.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d97284a87af970df3234d8ea0bfe397bc6b01b313300ef46e3003e956efe77f4 +size 438760 diff --git a/vlm/train/B1spAqUp-/2.png b/vlm/train/B1spAqUp-/2.png new file mode 100644 index 0000000000000000000000000000000000000000..c451552c085045558e6b554f4046495b07d14301 --- /dev/null +++ b/vlm/train/B1spAqUp-/2.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cf561f6f933564c182d7fc1b168b94a1383fa0d31d8a8c0e766f2f8b34c78bb3 +size 456563 diff --git a/vlm/train/B1spAqUp-/3.png b/vlm/train/B1spAqUp-/3.png new file mode 100644 index 0000000000000000000000000000000000000000..2fe39d653868a638173c473fc767f74ac97523fc --- /dev/null +++ b/vlm/train/B1spAqUp-/3.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9846f88dd7887d074c8c16dd0bfa5b5ce9ac935af75bee41adc43d3902bbe248 +size 757545 diff --git a/vlm/train/B1spAqUp-/4.png b/vlm/train/B1spAqUp-/4.png new file mode 100644 index 0000000000000000000000000000000000000000..ee8ee0a4bcc72b5a5fc2b456a03d138b930200ce --- /dev/null +++ b/vlm/train/B1spAqUp-/4.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e5bad19bad1c178979e40ec6b8fc9f7c988a43028b9398fd8185f91c3f05939d +size 523182 diff --git a/vlm/train/B1spAqUp-/5.png b/vlm/train/B1spAqUp-/5.png new file mode 100644 index 0000000000000000000000000000000000000000..e19a38b8c9a59e50f9711dc1afc77c81387879c8 --- /dev/null +++ b/vlm/train/B1spAqUp-/5.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c99854095ec633a636b807723bd74fb0dde5157a25d0dd54cff7c78e4d24144f +size 496838 diff --git a/vlm/train/B1spAqUp-/6.png b/vlm/train/B1spAqUp-/6.png new file mode 100644 index 0000000000000000000000000000000000000000..0f89322d6a8b1d94213e818d05999238d2946852 --- /dev/null +++ b/vlm/train/B1spAqUp-/6.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cf3707714a7c3b7669857a5bda0fd7daa2b42b2728313d57b5d5c2eb69c4d4a2 +size 687050 diff --git a/vlm/train/B1spAqUp-/7.png b/vlm/train/B1spAqUp-/7.png new file mode 100644 index 0000000000000000000000000000000000000000..a2bb61248a34651c86bc059405fd49cf00262f07 --- /dev/null +++ b/vlm/train/B1spAqUp-/7.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ad533b7c47eecd329a3a2f1377d07411b824def38bddb8d23a5b728f73a564b +size 664016 diff --git a/vlm/train/B1spAqUp-/8.png b/vlm/train/B1spAqUp-/8.png new file mode 100644 index 0000000000000000000000000000000000000000..b132e9fc4e6bbfbaf37d9d5bd86b211fda958a74 --- /dev/null +++ b/vlm/train/B1spAqUp-/8.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:26798f7a58d20e7a723903375dd8eb66adbbc89c51f6ffd60d7e0dd4ddda6a81 +size 1020084 diff --git a/vlm/train/B1spAqUp-/9.png b/vlm/train/B1spAqUp-/9.png new file mode 100644 index 0000000000000000000000000000000000000000..6ae3dca92d2f8b977ee12c09792afd360943f0cf --- /dev/null +++ b/vlm/train/B1spAqUp-/9.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:669b180a87920b30b7ec6a39f662ab3c6d303d7028299b5340d97554f0227d91 +size 553164 diff --git a/vlm/train/BJgVaG-Ab/0.png b/vlm/train/BJgVaG-Ab/0.png new file mode 100644 index 0000000000000000000000000000000000000000..1177a3c48c6070f4457a67e7941d5a8a86d765fd --- /dev/null +++ b/vlm/train/BJgVaG-Ab/0.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1847c7d1732b8d9712afdb571e81f4ae697bc95658992c08450530cbc7f95252 +size 473861 diff --git a/vlm/train/BJgVaG-Ab/1.png b/vlm/train/BJgVaG-Ab/1.png new file mode 100644 index 0000000000000000000000000000000000000000..6f8f28dd88e77e09393565bfcda613661319a2a5 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/1.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:66ac28cb44ff0acff27d388229c5b376cc792daf26adeb55836b0182579c692b +size 490432 diff --git a/vlm/train/BJgVaG-Ab/10.png b/vlm/train/BJgVaG-Ab/10.png new file mode 100644 index 0000000000000000000000000000000000000000..d4c5ced207aec101f58acc90ba95d1d774a99188 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/10.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:196eb98158b6282f3ebabb349a824b60116f9172d5d20afdc52018c83456f51a +size 430316 diff --git a/vlm/train/BJgVaG-Ab/11.png b/vlm/train/BJgVaG-Ab/11.png new file mode 100644 index 0000000000000000000000000000000000000000..dd3784d5b7fbd848c2512abcabcb5a0b2485cca3 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/11.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:837d66e11ba11b832423c2ca3fdb5cfd20b30134aaabb311ca8fba307f8ba50c +size 288210 diff --git a/vlm/train/BJgVaG-Ab/12.png b/vlm/train/BJgVaG-Ab/12.png new file mode 100644 index 0000000000000000000000000000000000000000..f74dc02dc9915a9da5f5808531c1a7254daf8fa1 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/12.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b397cb84975af7376dd5a49e3ef4561d315f01db8442239e66fd0a309301236 +size 280116 diff --git a/vlm/train/BJgVaG-Ab/13.png b/vlm/train/BJgVaG-Ab/13.png new file mode 100644 index 0000000000000000000000000000000000000000..dcb9428861899dfe962029d326d0fa08912fa929 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/13.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:967ba96bc01be746bc5ba0e9e80f95d939807c8fb520213086ca39bfa59836ce +size 213639 diff --git a/vlm/train/BJgVaG-Ab/14.png b/vlm/train/BJgVaG-Ab/14.png new file mode 100644 index 0000000000000000000000000000000000000000..7a46a757891397042b27d14ef9d91c54e6acbea9 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/14.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1f0e8482bcb74f4795c5a7d26acc774ee0c3dc0b5bd8cb0b0af909e931389ec +size 318498 diff --git a/vlm/train/BJgVaG-Ab/2.png b/vlm/train/BJgVaG-Ab/2.png new file mode 100644 index 0000000000000000000000000000000000000000..e97c8f426fc4b2e06239a3d945a7f14867234dc6 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/2.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e67f497c408f5b68d4d9834438e1d58409cb2cea45b89f2cebc83bef78f0cd1e +size 488314 diff --git a/vlm/train/BJgVaG-Ab/3.png b/vlm/train/BJgVaG-Ab/3.png new file mode 100644 index 0000000000000000000000000000000000000000..8a5c4bfbdc512475689ee687838ed4a7879259c5 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/3.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f37ff939ea8b45f142e4ee7d1e3c9303b8456cb82c08bee23b79867bbf3b12a3 +size 479994 diff --git a/vlm/train/BJgVaG-Ab/4.png b/vlm/train/BJgVaG-Ab/4.png new file mode 100644 index 0000000000000000000000000000000000000000..25801c0838d2d2c840ee81621bf18fd643b16d9d --- /dev/null +++ b/vlm/train/BJgVaG-Ab/4.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ec48be241a7f4affc1d635a029b831e9d5c82a75eeda90311a2a13a4b845249e +size 486406 diff --git a/vlm/train/BJgVaG-Ab/5.png b/vlm/train/BJgVaG-Ab/5.png new file mode 100644 index 0000000000000000000000000000000000000000..f27df9ca8ed507afe72e65adc9bc860f401885b5 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/5.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:88b4f9859f6e0c1badaf6ab974ad64a5b2488973bbb0b2998e909c06556073ee +size 340442 diff --git a/vlm/train/BJgVaG-Ab/6.png b/vlm/train/BJgVaG-Ab/6.png new file mode 100644 index 0000000000000000000000000000000000000000..95ec619ee42eecfcbafed7185e4d0275be580546 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/6.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c7b7f784abfc96936fbb590a8e1591bd99c452f32712b0dbf3518aa48eeffea0 +size 520819 diff --git a/vlm/train/BJgVaG-Ab/7.png b/vlm/train/BJgVaG-Ab/7.png new file mode 100644 index 0000000000000000000000000000000000000000..7044cc4a0fe93aa9e2ea98c49ad933ea300c2af2 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/7.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a4a22e651f0d4b4d74d43856a1b8319a56a5e7c49f6e3175a75a03b2bd83bf2c +size 572030 diff --git a/vlm/train/BJgVaG-Ab/8.png b/vlm/train/BJgVaG-Ab/8.png new file mode 100644 index 0000000000000000000000000000000000000000..577fadad214754829e5cc0ace0f11755a798d463 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/8.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:284433c5b73ea458d66a41f9d88acf9914b3fea3adcf9e60885352d1aca922bc +size 535055 diff --git a/vlm/train/BJgVaG-Ab/9.png b/vlm/train/BJgVaG-Ab/9.png new file mode 100644 index 0000000000000000000000000000000000000000..3b1f25585dd9e0d0b3372677795e026a72abab06 --- /dev/null +++ b/vlm/train/BJgVaG-Ab/9.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:750e57ca77979892a42a47a97a8fa21597d0d0ff9bd1792c739595aaa561c73e +size 473860 diff --git a/vlm/train/HCSgyPUfeDj/0.png b/vlm/train/HCSgyPUfeDj/0.png new file mode 100644 index 0000000000000000000000000000000000000000..98a3b1e2b00acd3dbdd84614ca2dd0ef84882d08 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/0.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:95797fe40104755aa3d054fa13c4176c5852f281a4c6ff58cdff0e452bbefbec +size 485525 diff --git a/vlm/train/HCSgyPUfeDj/1.png b/vlm/train/HCSgyPUfeDj/1.png new file mode 100644 index 0000000000000000000000000000000000000000..de289479facf9adc1d0d3dd0cb8282e8168ff7fb --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/1.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4e9b71ed1e902c9f5e333be5a37523c7cb82e04306b63ee2df453c7175c3a747 +size 594407 diff --git a/vlm/train/HCSgyPUfeDj/10.png b/vlm/train/HCSgyPUfeDj/10.png new file mode 100644 index 0000000000000000000000000000000000000000..8c61cabec4e539c9715ba532a43ab42ccd39c3bd --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/10.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:44988c412137c8a8878a81eb840399856b354bff0eb5bb6cc0661e7300cfb9b1 +size 618805 diff --git a/vlm/train/HCSgyPUfeDj/11.png b/vlm/train/HCSgyPUfeDj/11.png new file mode 100644 index 0000000000000000000000000000000000000000..6545fc33a4e19c972819484fe1585dd1243db828 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/11.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:298afd0887d73299ddb96efebfe1c2d3cc336f2c65dbcd6e71f5a058bfca19a0 +size 628372 diff --git a/vlm/train/HCSgyPUfeDj/12.png b/vlm/train/HCSgyPUfeDj/12.png new file mode 100644 index 0000000000000000000000000000000000000000..3707c38e72f4f4f3eb9beb1acfe581b925e1e3dd --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/12.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b22c43dd9f250465b879c9ad0162be2bd735f49605098ff88eb5c0229b64c159 +size 645656 diff --git a/vlm/train/HCSgyPUfeDj/13.png b/vlm/train/HCSgyPUfeDj/13.png new file mode 100644 index 0000000000000000000000000000000000000000..90cfb776f0aa13df238816a85e7d2b008e10bf79 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/13.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab9b0544521771b0beed1e30a25f18b44746731cfe99a8bdab570305e1e087f9 +size 67517 diff --git a/vlm/train/HCSgyPUfeDj/14.png b/vlm/train/HCSgyPUfeDj/14.png new file mode 100644 index 0000000000000000000000000000000000000000..68e91b744b0884ac34a8c02ebd8b8c583cc92142 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/14.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4275f3e17d8a491d6870dc4744f1754b720972adedb4463fad704fd2670381e2 +size 452925 diff --git a/vlm/train/HCSgyPUfeDj/15.png b/vlm/train/HCSgyPUfeDj/15.png new file mode 100644 index 0000000000000000000000000000000000000000..912ddda0d24e7373118e970ce68219238cc431e5 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/15.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f49be4cf2c13d6fd7ea1a70af6dc6dcd6ded5f73316db8578f063609823e71d5 +size 589031 diff --git a/vlm/train/HCSgyPUfeDj/16.png b/vlm/train/HCSgyPUfeDj/16.png new file mode 100644 index 0000000000000000000000000000000000000000..d47799e2027c5765b7071e7ee69edf04c877fdfa --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/16.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4701e2c73f8f5be23780b4ade81f4972c823dc4a311c36c7c34b68ec017e54b1 +size 474457 diff --git a/vlm/train/HCSgyPUfeDj/17.png b/vlm/train/HCSgyPUfeDj/17.png new file mode 100644 index 0000000000000000000000000000000000000000..e3dc41a184512a7cb09d4384e526ab22471845d8 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/17.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:259ac16d2d939c15fc2f6d1f4ea73987fadab3d42e0ed97ea96adf34e01f12b6 +size 502108 diff --git a/vlm/train/HCSgyPUfeDj/18.png b/vlm/train/HCSgyPUfeDj/18.png new file mode 100644 index 0000000000000000000000000000000000000000..33c31a26266c1bc8277b13a07220bbd92bd7bd1b --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/18.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86d50e0f73e21b62223dda072132cb112fd39373382e05711a4de6f493e1aa4d +size 540226 diff --git a/vlm/train/HCSgyPUfeDj/19.png b/vlm/train/HCSgyPUfeDj/19.png new file mode 100644 index 0000000000000000000000000000000000000000..562a4a79213433da74b5d13924ec28e3c654810c --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/19.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f959c040e60d2d66e84ac3f7d3853c0a2884399fb789236e8179121f1137b10a +size 576244 diff --git a/vlm/train/HCSgyPUfeDj/2.png b/vlm/train/HCSgyPUfeDj/2.png new file mode 100644 index 0000000000000000000000000000000000000000..b2afed4dc5b14e712dbacfa82049d34640fe37ad --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/2.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1a0031af2ef591db6e91952fa913b9faf737541dc2429e7fef0c9c7955309585 +size 527786 diff --git a/vlm/train/HCSgyPUfeDj/20.png b/vlm/train/HCSgyPUfeDj/20.png new file mode 100644 index 0000000000000000000000000000000000000000..4510ce4e271e7f6bc20aedfcf08f90d2469ce220 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/20.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a3336f6532f2c42d5c64af1157237014b3306e7ed29edcc9bead0b97088b6619 +size 917205 diff --git a/vlm/train/HCSgyPUfeDj/21.png b/vlm/train/HCSgyPUfeDj/21.png new file mode 100644 index 0000000000000000000000000000000000000000..dae98e773d4901be95687e1c984d5170f73fa8f4 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/21.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:029f43e30c7fff3a553216b359d0adf9720ffd5239ae20e308dd04b18aa548b8 +size 1157609 diff --git a/vlm/train/HCSgyPUfeDj/22.png b/vlm/train/HCSgyPUfeDj/22.png new file mode 100644 index 0000000000000000000000000000000000000000..d427c015be597cc947e2c5d64252eb41dec5dd58 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/22.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e10adaf6ead5f272545fc150ae9ef416bf8b5a3176eaf3c46eddb1e87fb29976 +size 1327730 diff --git a/vlm/train/HCSgyPUfeDj/23.png b/vlm/train/HCSgyPUfeDj/23.png new file mode 100644 index 0000000000000000000000000000000000000000..e4aeac429d688d5a46c7584f7fec8810a1979c8e --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/23.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4376eee14e8d067c519151bb5abcc9cee32820555cfa629709cd5d966be41a5b +size 981362 diff --git a/vlm/train/HCSgyPUfeDj/24.png b/vlm/train/HCSgyPUfeDj/24.png new file mode 100644 index 0000000000000000000000000000000000000000..59fc7199de7d496c5c4390de078c47d2a2df0ab9 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/24.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fdf4dcc5c572c50c6e921a992f61e9a93af2e75d68488d1b2a2cce1eee4ad38c +size 528886 diff --git a/vlm/train/HCSgyPUfeDj/25.png b/vlm/train/HCSgyPUfeDj/25.png new file mode 100644 index 0000000000000000000000000000000000000000..9c34c131a4c11c4b34041466582fc9d34cfe8890 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/25.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:351472e8adf9b02446b08bbd035c1f7f86301b0d974e1c23c04671c1f52ebeef +size 1158461 diff --git a/vlm/train/HCSgyPUfeDj/26.png b/vlm/train/HCSgyPUfeDj/26.png new file mode 100644 index 0000000000000000000000000000000000000000..8bd252a9b15ceab08f88fbebfe084fe47c12aa8b --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/26.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:803487fb180f41f23ac821859d64f35b399ccddb5e895908e2a362907ce0639b +size 1633168 diff --git a/vlm/train/HCSgyPUfeDj/27.png b/vlm/train/HCSgyPUfeDj/27.png new file mode 100644 index 0000000000000000000000000000000000000000..dfd95ef821c6342ea026ecc8091d6500743b4211 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/27.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9342e70226a43be4481f6fccfed9ac9379185cf901073f0fdef8d63ef4046586 +size 838749 diff --git a/vlm/train/HCSgyPUfeDj/28.png b/vlm/train/HCSgyPUfeDj/28.png new file mode 100644 index 0000000000000000000000000000000000000000..cb58ac24fb7a288857aa480d9eb81fd852f8fa5d --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/28.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cbdd76fe12ab6f1dd34773266dc5a3ed700f19301bb99585a8baa5404ca98df2 +size 849173 diff --git a/vlm/train/HCSgyPUfeDj/29.png b/vlm/train/HCSgyPUfeDj/29.png new file mode 100644 index 0000000000000000000000000000000000000000..53c6547b65dbbd2da3c2e34d0b177dfd9a138a3c --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/29.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1fc03177c959db8be01614f0cfa2b976d7502f965e0e9ce34a77ea3fc485a132 +size 1226004 diff --git a/vlm/train/HCSgyPUfeDj/3.png b/vlm/train/HCSgyPUfeDj/3.png new file mode 100644 index 0000000000000000000000000000000000000000..4e8d855bcb273a8900f6dec2384dacac0904ece6 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/3.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:da1a64cb1fd48c4e6a6ad01e37e9c8db2f7625009a6b10d7712f7c7f9f4495c2 +size 616122 diff --git a/vlm/train/HCSgyPUfeDj/30.png b/vlm/train/HCSgyPUfeDj/30.png new file mode 100644 index 0000000000000000000000000000000000000000..fcdaf6b40edfaca345a84d841542c01aa8dcb38c --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/30.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6f56aa3148c138d0cb3bce2820220efffb91341dc5ac3b48522c2d9080ab68f3 +size 1472457 diff --git a/vlm/train/HCSgyPUfeDj/31.png b/vlm/train/HCSgyPUfeDj/31.png new file mode 100644 index 0000000000000000000000000000000000000000..e42aac560fc2e4be70729aa143a4ba53acfc9a16 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/31.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:003e66e8803ca3f8b516df215d720ffe8f6e716673f1ae03a3d136d663a055a1 +size 1320186 diff --git a/vlm/train/HCSgyPUfeDj/4.png b/vlm/train/HCSgyPUfeDj/4.png new file mode 100644 index 0000000000000000000000000000000000000000..9f66ef4e023312fd9c871cb39d60c9645ff241e7 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/4.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a691b1af7588121a678da84d2f522a4174ae3e9b9eacf309ce1cd166b72a1cee +size 635822 diff --git a/vlm/train/HCSgyPUfeDj/5.png b/vlm/train/HCSgyPUfeDj/5.png new file mode 100644 index 0000000000000000000000000000000000000000..eb991e4a2e7e41e5760439d6cbd816717836237a --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/5.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e2c92a5f123742f0e8bcc588725bdd135d56cee071e7b433d533eb95ec3982ee +size 601582 diff --git a/vlm/train/HCSgyPUfeDj/6.png b/vlm/train/HCSgyPUfeDj/6.png new file mode 100644 index 0000000000000000000000000000000000000000..fb57a92dd9b77962754c758ca8947c8c14a123d5 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/6.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fcb85e3499122a891bb0f64c354f49550ca35a71962736522bc943036c106a3 +size 565309 diff --git a/vlm/train/HCSgyPUfeDj/7.png b/vlm/train/HCSgyPUfeDj/7.png new file mode 100644 index 0000000000000000000000000000000000000000..3687de41f3a51669c458af1eceacaed522c90647 --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/7.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4d9a808697515247b8be61d0d87db3cf495654ed1f0e4f933478855a97ad5e76 +size 598341 diff --git a/vlm/train/HCSgyPUfeDj/8.png b/vlm/train/HCSgyPUfeDj/8.png new file mode 100644 index 0000000000000000000000000000000000000000..1eb4185993bc59e5d37d57091ac49ce878fc0e2e --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/8.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c8f52816843fc75935918b20fa8f8cc651064b6424d2033b55fa9f88ac0c142e +size 851574 diff --git a/vlm/train/HCSgyPUfeDj/9.png b/vlm/train/HCSgyPUfeDj/9.png new file mode 100644 index 0000000000000000000000000000000000000000..e999890800b5ec3384f3f240322ccdb7974fe45d --- /dev/null +++ b/vlm/train/HCSgyPUfeDj/9.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca85ef5aa096bad59a3fa863dedd80a23f1dfe2185cad96daaabf488e1849664 +size 594857 diff --git a/vlm/train/HJgBA2VYwH/0.png b/vlm/train/HJgBA2VYwH/0.png new file mode 100644 index 0000000000000000000000000000000000000000..0a5d032e2a452d677b5519d775e3f1c90b939402 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/0.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:847455dcbb295588bdec18b9ccdf3918c6c20f4ac97501b353c01e5ab8300dc4 +size 429281 diff --git a/vlm/train/HJgBA2VYwH/1.png b/vlm/train/HJgBA2VYwH/1.png new file mode 100644 index 0000000000000000000000000000000000000000..346dc66259dd19753f280dd8641b1fcca329abdf --- /dev/null +++ b/vlm/train/HJgBA2VYwH/1.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:95536b9d2e052a13092a675c3707016243c30a5c85bdbe3320ab2956b3176eef +size 459426 diff --git a/vlm/train/HJgBA2VYwH/10.png b/vlm/train/HJgBA2VYwH/10.png new file mode 100644 index 0000000000000000000000000000000000000000..e54e1fe9395095924b640b92e1cd2de5831b88da --- /dev/null +++ b/vlm/train/HJgBA2VYwH/10.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c316db3a2c14b8f1fa02415fff500fa2d155f991a84be2949b875e29bed0cb76 +size 533744 diff --git a/vlm/train/HJgBA2VYwH/11.png b/vlm/train/HJgBA2VYwH/11.png new file mode 100644 index 0000000000000000000000000000000000000000..c19f4d39e8840c1fb8a1a2b79b2e817d7736992b --- /dev/null +++ b/vlm/train/HJgBA2VYwH/11.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:76a3c3c5999a7ad9b46df56f25c310cfbfe5776d42c59c54577ea3c911e4bc92 +size 273770 diff --git a/vlm/train/HJgBA2VYwH/12.png b/vlm/train/HJgBA2VYwH/12.png new file mode 100644 index 0000000000000000000000000000000000000000..b0f24be4311c5af7063c4bf86ae8bef7b78f7015 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/12.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5d33f13f17177ab089ee8758710cfde0e3ab789e680330ec662ad23c92058e00 +size 393779 diff --git a/vlm/train/HJgBA2VYwH/13.png b/vlm/train/HJgBA2VYwH/13.png new file mode 100644 index 0000000000000000000000000000000000000000..027639eb22173929c40ef03e3ae37c31026aa3f6 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/13.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1246813777a37102e1f6e4e710a78b9349be933fcad4bd223f8b00efdb1440de +size 411575 diff --git a/vlm/train/HJgBA2VYwH/14.png b/vlm/train/HJgBA2VYwH/14.png new file mode 100644 index 0000000000000000000000000000000000000000..fa0ee2d9bbb98affcd6e07c94dbe7f55c2b74172 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/14.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ff802060780ab41c3648f5cdc583ce186f2b469be796324d8481a65ed1c32e9 +size 411333 diff --git a/vlm/train/HJgBA2VYwH/15.png b/vlm/train/HJgBA2VYwH/15.png new file mode 100644 index 0000000000000000000000000000000000000000..92fc7c61a39ece32ed7d467c22737deee2c12b96 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/15.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:23c31f198e46a73700967affab52cdb32087061cadd5d3f26114099d231abf3b +size 317131 diff --git a/vlm/train/HJgBA2VYwH/16.png b/vlm/train/HJgBA2VYwH/16.png new file mode 100644 index 0000000000000000000000000000000000000000..b0d6b8eab365111bc4c4c5884074915e137974fe --- /dev/null +++ b/vlm/train/HJgBA2VYwH/16.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:64bb1b4801a9d3605051689799f5133870967202a9c4c4174b277693144b0a8c +size 471953 diff --git a/vlm/train/HJgBA2VYwH/17.png b/vlm/train/HJgBA2VYwH/17.png new file mode 100644 index 0000000000000000000000000000000000000000..35081e4cdd9d8b3a6a6ad04a6b4cac2809b0e73f --- /dev/null +++ b/vlm/train/HJgBA2VYwH/17.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f1a395ac3833e08b067c6ce9b3a3165e1d8d0c5396e23f963237f2afc4183564 +size 504816 diff --git a/vlm/train/HJgBA2VYwH/18.png b/vlm/train/HJgBA2VYwH/18.png new file mode 100644 index 0000000000000000000000000000000000000000..393528acb9cd000dd183d5b01e9b7b3555dbb012 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/18.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:08bf095f7f85a15caa5c690a6ca26f50cbd4e4a42763170391dceefbc47b2039 +size 467073 diff --git a/vlm/train/HJgBA2VYwH/19.png b/vlm/train/HJgBA2VYwH/19.png new file mode 100644 index 0000000000000000000000000000000000000000..ebae61f41413d701eb549237b1a509a6b9cfbaff --- /dev/null +++ b/vlm/train/HJgBA2VYwH/19.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:978a989fab1c482eadf0d96d40ff371778ccaf112ecf9e62660eb9b7037745b7 +size 444441 diff --git a/vlm/train/HJgBA2VYwH/2.png b/vlm/train/HJgBA2VYwH/2.png new file mode 100644 index 0000000000000000000000000000000000000000..5d9aed92d468c071078a6adc7420f7adf2d34eb9 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/2.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7d9230034513d7c9628793f4d5c9606ea2c82e100eb55e51e0e74b89bb866402 +size 514138 diff --git a/vlm/train/HJgBA2VYwH/3.png b/vlm/train/HJgBA2VYwH/3.png new file mode 100644 index 0000000000000000000000000000000000000000..60ba1c3e3dd16189b6f8f97648427c4835c5d1d6 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/3.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7ddfaad2bc7f9bef993a5a2d32f859dc601ee4d3218b68804d145af3ef4b6a03 +size 427032 diff --git a/vlm/train/HJgBA2VYwH/4.png b/vlm/train/HJgBA2VYwH/4.png new file mode 100644 index 0000000000000000000000000000000000000000..a88a582b9eebd917139c8855b4a0a75be41d0f28 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/4.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:883c8ed0d082b98a3f2e56ae21c3134ba54ad8777133027e6f5c5c43749d1195 +size 440167 diff --git a/vlm/train/HJgBA2VYwH/5.png b/vlm/train/HJgBA2VYwH/5.png new file mode 100644 index 0000000000000000000000000000000000000000..20986e6246e2c1e76766f3f461e425b8c7667398 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/5.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f7ba6014ba2459ee383bc81c1f6b8ed856212f80996f607a2e093f5cc1c6651c +size 564527 diff --git a/vlm/train/HJgBA2VYwH/6.png b/vlm/train/HJgBA2VYwH/6.png new file mode 100644 index 0000000000000000000000000000000000000000..90b158568ac49d5d84d8e1c657a83da01683e6f5 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/6.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6f834f87ca66de4dca7ca5dae7ea4554cf39012d57d6526e3d2958f9a3af7d2 +size 672363 diff --git a/vlm/train/HJgBA2VYwH/7.png b/vlm/train/HJgBA2VYwH/7.png new file mode 100644 index 0000000000000000000000000000000000000000..6d273057bff30e4b11fb93ac043d23dffe10be19 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/7.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ca08b926eb3fc3bd6668b8e920c13389ff70a9d2a9460acbd928381031c18c9 +size 506450 diff --git a/vlm/train/HJgBA2VYwH/8.png b/vlm/train/HJgBA2VYwH/8.png new file mode 100644 index 0000000000000000000000000000000000000000..9972faf9344ce58f34e910dcecfd2d570a0ae5c0 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/8.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9692291da88a81fdacffc5264438cc389d12a129913bdef326ab4852fbb3bda1 +size 499009 diff --git a/vlm/train/HJgBA2VYwH/9.png b/vlm/train/HJgBA2VYwH/9.png new file mode 100644 index 0000000000000000000000000000000000000000..a945e2b468c7b3b893fcc219a2f9ed038b222068 --- /dev/null +++ b/vlm/train/HJgBA2VYwH/9.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f9cdadeaa3136ec0c0d09b002b6ca9224972100dc746b874eec259fce19cc0d8 +size 547474 diff --git a/vlm/train/a1wQOh27zcy/12.png b/vlm/train/a1wQOh27zcy/12.png new file mode 100644 index 0000000000000000000000000000000000000000..1e1d7316257e5f87a64e48637d8a02ae55bc5b1a --- /dev/null +++ b/vlm/train/a1wQOh27zcy/12.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5587f2835bbd35f43b9f9f708a77d56c56c9eaa22f003e1e0b7af0566f31c40c +size 577927 diff --git a/vlm/train/a1wQOh27zcy/13.png b/vlm/train/a1wQOh27zcy/13.png new file mode 100644 index 0000000000000000000000000000000000000000..c60a472c6d8f244b812619f794ccdff0d0d23742 --- /dev/null +++ b/vlm/train/a1wQOh27zcy/13.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:85dec40a285ebd3cbd28dafc81a94b79439344be798f0506c51795ca307893e6 +size 590079 diff --git a/vlm/train/a1wQOh27zcy/4.png b/vlm/train/a1wQOh27zcy/4.png new file mode 100644 index 0000000000000000000000000000000000000000..63790a7d0bcedeaf3064b70e954a677d7d6bf425 --- /dev/null +++ b/vlm/train/a1wQOh27zcy/4.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5bbf194523ab8691a2b438e7d096f4178dd2097be8d6b23068f110fd372ead8f +size 751288 diff --git a/vlm/train/a1wQOh27zcy/5.png b/vlm/train/a1wQOh27zcy/5.png new file mode 100644 index 0000000000000000000000000000000000000000..0c0bff30fcd6e4df1e09efa7d7df77db966d1d02 --- /dev/null +++ b/vlm/train/a1wQOh27zcy/5.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5f60d6e7e04e57a588d74314f487110724d21bc19cfbfb1bfbcf8629f64b6795 +size 520069 diff --git a/vlm/train/a1wQOh27zcy/7.png b/vlm/train/a1wQOh27zcy/7.png new file mode 100644 index 0000000000000000000000000000000000000000..811ef57bb42ee61f6a7d0c0a4ff93b4b21f9b515 --- /dev/null +++ b/vlm/train/a1wQOh27zcy/7.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2d072ef6c5767c5f3e359d55959629a07d7d845bdbb67a18b1a9a0340ae25088 +size 707777 diff --git a/vlm/train/a1wQOh27zcy/8.png b/vlm/train/a1wQOh27zcy/8.png new file mode 100644 index 0000000000000000000000000000000000000000..8da5d5ef52a4b2583b0739dde66ea9a33f97817b --- /dev/null +++ b/vlm/train/a1wQOh27zcy/8.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:596d85f4ebec538b6268042bf99ad221aefbf59addb5eba7922a882d14091d5e +size 695516 diff --git a/vlm/train/jznizqvr15J/0.png b/vlm/train/jznizqvr15J/0.png new file mode 100644 index 0000000000000000000000000000000000000000..aca3439e184b5cfb11fc57db2790ff4315816933 --- /dev/null +++ b/vlm/train/jznizqvr15J/0.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7d680e1446ef067539e86585b5f1a020bfcbbfe992195c3509c20c6dfcc393a7 +size 537552 diff --git a/vlm/train/jznizqvr15J/1.png b/vlm/train/jznizqvr15J/1.png new file mode 100644 index 0000000000000000000000000000000000000000..3035a6e29742d1693ee2caf92a2ed3c62a401113 --- /dev/null +++ b/vlm/train/jznizqvr15J/1.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4cc881ce0f86a3ba6cb35d2a64f55f9efe4faf2ee30e5c8f121770c56f025cca +size 593588 diff --git a/vlm/train/jznizqvr15J/10.png b/vlm/train/jznizqvr15J/10.png new file mode 100644 index 0000000000000000000000000000000000000000..4b498576af3d5a7eb5c11c63188d746b1b14fed6 --- /dev/null +++ b/vlm/train/jznizqvr15J/10.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b8df6ec0879f46e17e35e6f7e8d5ce228defd27b646c5ff937e3c2ae236e0307 +size 597832 diff --git a/vlm/train/jznizqvr15J/11.png b/vlm/train/jznizqvr15J/11.png new file mode 100644 index 0000000000000000000000000000000000000000..617e9cbbd0ab4e13ee590548386bb8fbdb3eecf1 --- /dev/null +++ b/vlm/train/jznizqvr15J/11.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d7315165482e0cf8cef08ab03745640f8f01242ed0775e191260559e0342c596 +size 565018 diff --git a/vlm/train/jznizqvr15J/12.png b/vlm/train/jznizqvr15J/12.png new file mode 100644 index 0000000000000000000000000000000000000000..8cf4e73b87c9763d159587469084d7a112978c70 --- /dev/null +++ b/vlm/train/jznizqvr15J/12.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d78d5099df11dcae20b6137be902b20f4157a76d34dc33554eb13c90b3472a9a +size 277650 diff --git a/vlm/train/jznizqvr15J/2.png b/vlm/train/jznizqvr15J/2.png new file mode 100644 index 0000000000000000000000000000000000000000..8878be20f9bc1c53c062f4ed01e289d6f7744959 --- /dev/null +++ b/vlm/train/jznizqvr15J/2.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ec5c914221b463d38a1e8090fdb6904ab5b8e4fd377aab595410f36663618ebd +size 432970 diff --git a/vlm/train/jznizqvr15J/3.png b/vlm/train/jznizqvr15J/3.png new file mode 100644 index 0000000000000000000000000000000000000000..3f8ea3b0582a22aa4d295a3ce9ecabf0f0ead4bb --- /dev/null +++ b/vlm/train/jznizqvr15J/3.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:92d1ad7d31ec99945a13029301ee24f248987c5f4ecdd24b5bbc3edb29b1be1b +size 579638 diff --git a/vlm/train/jznizqvr15J/4.png b/vlm/train/jznizqvr15J/4.png new file mode 100644 index 0000000000000000000000000000000000000000..9a05c3a7719751258ef82de681518d70e008102b --- /dev/null +++ b/vlm/train/jznizqvr15J/4.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e39efcb1cc3126cdda50310e2462cdcf3055765cbf84575d0cb423e00345cf18 +size 563885 diff --git a/vlm/train/jznizqvr15J/5.png b/vlm/train/jznizqvr15J/5.png new file mode 100644 index 0000000000000000000000000000000000000000..83d4bfff3614e271923e6432bc200565567ba440 --- /dev/null +++ b/vlm/train/jznizqvr15J/5.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:567b44c5bb9dc30f4b6be2982302ebabe351883679bd3793084e82e51f96b7a4 +size 615334 diff --git a/vlm/train/jznizqvr15J/6.png b/vlm/train/jznizqvr15J/6.png new file mode 100644 index 0000000000000000000000000000000000000000..4162774f7421d116722be94992e6dd1f063d495a --- /dev/null +++ b/vlm/train/jznizqvr15J/6.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:82b541e68c131227c119fd7c7adb08bd67067e8d51e939bd4e1825cc40a75f7c +size 659274 diff --git a/vlm/train/jznizqvr15J/7.png b/vlm/train/jznizqvr15J/7.png new file mode 100644 index 0000000000000000000000000000000000000000..0e0fdba4711e38e9af06ef83065e10f22fb51d31 --- /dev/null +++ b/vlm/train/jznizqvr15J/7.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2a48e8ab03fad0ab2a3bf7332f1656f699772eae6002c42bffa97edfdcc28a0 +size 613849 diff --git a/vlm/train/jznizqvr15J/8.png b/vlm/train/jznizqvr15J/8.png new file mode 100644 index 0000000000000000000000000000000000000000..34fbe913630ae0738b3127b01b9a9e19de65bc5c --- /dev/null +++ b/vlm/train/jznizqvr15J/8.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9c8bc626452a5a9ce900577cccb7483e0a14ba1d192bd9c75155a1f604667081 +size 627916 diff --git a/vlm/train/jznizqvr15J/9.png b/vlm/train/jznizqvr15J/9.png new file mode 100644 index 0000000000000000000000000000000000000000..b7625b8da63fc985c9cb2db067c803fac7e89b73 --- /dev/null +++ b/vlm/train/jznizqvr15J/9.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4052bbc3347d50dbed134647287f9314488d42de6959298575b03f76ca78de8a +size 557356 diff --git a/vlm/train/rJl0r3R9KX/10.png b/vlm/train/rJl0r3R9KX/10.png new file mode 100644 index 0000000000000000000000000000000000000000..dee7aad34b9f2900bf02357e5006366f417de446 --- /dev/null +++ b/vlm/train/rJl0r3R9KX/10.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:133c8f0b5e66bdba9018c8e8626cc56bf60d502982d54b6009f248ee9c98660f +size 531457 diff --git a/vlm/train/rJl0r3R9KX/17.png b/vlm/train/rJl0r3R9KX/17.png new file mode 100644 index 0000000000000000000000000000000000000000..8860aba09a0f5d1e9bc0bbe6ccc7bfbba2232c5a --- /dev/null +++ b/vlm/train/rJl0r3R9KX/17.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5b433aeb1647e74173e7c348f31f8d3be86bc59f90a733d07efa66cc5375eb2e +size 362853 diff --git a/vlm/train/rJl0r3R9KX/21.png b/vlm/train/rJl0r3R9KX/21.png new file mode 100644 index 0000000000000000000000000000000000000000..f0bfdb853b1dab9e99a859f49b3c82644bea6f0b --- /dev/null +++ b/vlm/train/rJl0r3R9KX/21.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:16c51ac2f95d60736868fb95579bffad9382ef174388fa9de8893225e2d4c8f0 +size 337861 diff --git a/vlm/train/rJl0r3R9KX/23.png b/vlm/train/rJl0r3R9KX/23.png new file mode 100644 index 0000000000000000000000000000000000000000..e14516185aa4b1ef5e134a43331286e532b5249c --- /dev/null +++ b/vlm/train/rJl0r3R9KX/23.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1c2c7e483783624673c316579002c26ea27ed6d17353e36f21219dca37368fe +size 364647 diff --git a/vlm/train/rJl0r3R9KX/24.png b/vlm/train/rJl0r3R9KX/24.png new file mode 100644 index 0000000000000000000000000000000000000000..07bfd91ff05520be08edcd4cffa0291deb9f6213 --- /dev/null +++ b/vlm/train/rJl0r3R9KX/24.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:77dcfcec2cd66aa62f56e554a461068022a75fbe29fc611ccca638b07128147a +size 356372 diff --git a/vlm/train/rJl0r3R9KX/3.png b/vlm/train/rJl0r3R9KX/3.png new file mode 100644 index 0000000000000000000000000000000000000000..e1f29dfbb46fc4fcde7b2b1db1924ef3926963d0 --- /dev/null +++ b/vlm/train/rJl0r3R9KX/3.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f4719b78dbad2421f98d40c37f4d2b804173757878b706f56a322a4cf69a45dc +size 500573 diff --git a/vlm/train/rJl0r3R9KX/6.png b/vlm/train/rJl0r3R9KX/6.png new file mode 100644 index 0000000000000000000000000000000000000000..d93b96cf59721171fd219936bedccd12f730e9be --- /dev/null +++ b/vlm/train/rJl0r3R9KX/6.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b88c2808fb386af0c0b900934b5d2ff8555bd4adeb38345ad1b6bd1b87fb916a +size 561872 diff --git a/vlm/train/uJDmJp7kn2F/0.png b/vlm/train/uJDmJp7kn2F/0.png new file mode 100644 index 0000000000000000000000000000000000000000..00eb2c454f97f3e3df93d00bc4525189996f1c6a --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/0.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2cbe5f69f9f0700102693846141e2aea3bece8d2b7c72be64afeef6019eb9b78 +size 397882 diff --git a/vlm/train/uJDmJp7kn2F/1.png b/vlm/train/uJDmJp7kn2F/1.png new file mode 100644 index 0000000000000000000000000000000000000000..3ad06f6f2d530f12a7e48e64fa168f2b4995b5ef --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/1.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fe17f67079da81b5f03008ae18ba2f12113e6a1107d18edeb87cd47c2975b8f8 +size 550494 diff --git a/vlm/train/uJDmJp7kn2F/10.png b/vlm/train/uJDmJp7kn2F/10.png new file mode 100644 index 0000000000000000000000000000000000000000..33e654ed54c56edba56e2a1bfe5d28a9ffaf27ab --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/10.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:93da732a152c28fdb4d579139a84af562544401ad36a665fad79fe9c4b0e08d9 +size 508729 diff --git a/vlm/train/uJDmJp7kn2F/11.png b/vlm/train/uJDmJp7kn2F/11.png new file mode 100644 index 0000000000000000000000000000000000000000..64440d31633b7b40235df627981a4ee03f12a902 --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/11.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:67aa1329bf422ad535932013414b8cdd40f0b30446613fd72805638476c05927 +size 167953 diff --git a/vlm/train/uJDmJp7kn2F/12.png b/vlm/train/uJDmJp7kn2F/12.png new file mode 100644 index 0000000000000000000000000000000000000000..8cf75073f95efe010c10a21d3f7dd293e2a14ff9 --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/12.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad8e26b033880bb7843409f0ed97c154066184f6bfa9589cff671ba54661f4eb +size 395339 diff --git a/vlm/train/uJDmJp7kn2F/2.png b/vlm/train/uJDmJp7kn2F/2.png new file mode 100644 index 0000000000000000000000000000000000000000..65691588c36b0a0f32228010624181e17f36fcbe --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/2.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2afe1373a41f26ae024e3ed99639108d04a3d523402193b9dbf4f0c5b2d57a0d +size 495161 diff --git a/vlm/train/uJDmJp7kn2F/3.png b/vlm/train/uJDmJp7kn2F/3.png new file mode 100644 index 0000000000000000000000000000000000000000..2ec73bd7da1d01ea2dcba4bcc8a23b72b5022830 --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/3.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6564bdd44d919ed95ac1a9206b8bcf8d8af5c03d4770d4cfb7b812db7da2041 +size 467815 diff --git a/vlm/train/uJDmJp7kn2F/4.png b/vlm/train/uJDmJp7kn2F/4.png new file mode 100644 index 0000000000000000000000000000000000000000..ebc3fc7dc267ab6efe754c8636b3ba9d656b8518 --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/4.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9a0816c301457bd5beb1bdaa80b97e44c2a8b277ecae60b22fcb31df19a5afe2 +size 468612 diff --git a/vlm/train/uJDmJp7kn2F/5.png b/vlm/train/uJDmJp7kn2F/5.png new file mode 100644 index 0000000000000000000000000000000000000000..f0887fcbc77d60e22cc663b0a5c6ab16b3694555 --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/5.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2581bbcb0f92c79882968d0cd4e55538fc1858db7a90f8ae17ed608516a4bbc +size 519200 diff --git a/vlm/train/uJDmJp7kn2F/6.png b/vlm/train/uJDmJp7kn2F/6.png new file mode 100644 index 0000000000000000000000000000000000000000..97c4839188420333c82421e4b260bbcc47c867ab --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/6.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:025ea3a9f596a64c192476537cbceee4872af73278ced62324a419a918f4f29e +size 454321 diff --git a/vlm/train/uJDmJp7kn2F/7.png b/vlm/train/uJDmJp7kn2F/7.png new file mode 100644 index 0000000000000000000000000000000000000000..ab151a10a63824435c6b691e6039ff80045bd3f0 --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/7.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc5616bf66df0e96794957bb95b15c30369f3d6037708d6e1757ca7b628d5431 +size 475919 diff --git a/vlm/train/uJDmJp7kn2F/8.png b/vlm/train/uJDmJp7kn2F/8.png new file mode 100644 index 0000000000000000000000000000000000000000..d97740b56974a6611cb98fa328547a07d895de2e --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/8.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:90fb521707fec58fc4700f58d3c5ff2d322b5d2da3114fe1725b36228a424b6e +size 376853 diff --git a/vlm/train/uJDmJp7kn2F/9.png b/vlm/train/uJDmJp7kn2F/9.png new file mode 100644 index 0000000000000000000000000000000000000000..3b6672d2f29c766febd7f6bec07adc7b3ca3510a --- /dev/null +++ b/vlm/train/uJDmJp7kn2F/9.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b230b9748497dafedb9ad00e296bf602adac7f8267a3f109c5f0984955f39498 +size 529572