search

arXiv / html_feedback

Supports a student project developing a UI for feedback on arXiv articles rendered as html.
MIT License
15 stars 3 forks source link

Figures rendered overlapping with captions #906

Open anish-bhattacharya opened 4 months ago

anish-bhattacharya commented 4 months ago

Description

The figures themselves (pictures) are rendered behind the captions of the figures.

(Optional:) Please add any files, screenshots, or other information here.

No response

(Required) What is this issue most closely related to? Select one.

Choose One

Internal issue ID

91d09bb6-685a-4794-a9d2-bae46f2f961c

Paper URL

https://arxiv.org/html/2403.12203v1

Browser

Chrome/125.0.0.0

Device Type

Desktop

html-feedback-bot[bot] commented 4 months ago

Location in document: undefined

Selected HTML: tℒI⁢L=‖ateacher−astudent‖2subscriptℒ𝐼𝐿superscriptnormsubscript𝑎teachersubscript𝑎student2\mathcal{L}{IL}=\left|a{\text{teacher}}-a_{\text{student}}\right|^{2}caligraphic_L start_POSTSUBSCRIPT italic_I italic_L end_POSTSUBSCRIPT = ∥ italic_a start_POSTSUBSCRIPT teacher end_POSTSUBSCRIPT - italic_a start_POSTSUBSCRIPT student end_POSTSUBSCRIPT ∥ start_POSTSUPERSCRIPT 2 endPOSTSUPERSCRIPT<math alttext="\mathcal{I}{t}" class="ltx_Math" display="inline" id="S1.F1.pic1.10.10.10.10.10.10.10.10.10.10.10.10.10.1.1.1.1.1.1.1.1.1.m1.1">ℐtsubscriptℐ𝑡\mathcal{I}_{t}caligraphic_I start_POSTSUBSCRIPT italic_t endPOSTSUBSCRIPT<math alttext="\mathcal{I}{t-N+1}" class="ltx_Math" display="inline" id="S1.F1.pic1.11.11.11.11.11.11.11.11.11.11.11.11.11.1.1.1.1.1.1.1.1.1.m1.1">ℐt−N+1subscriptℐ𝑡𝑁1\mathcal{I}_{t-N+1}caligraphic_I start_POSTSUBSCRIPT italic_t - italic_N + 1 end_POSTSUBSCRIPTState Inputℐtsubscriptℐ𝑡\mathcal{I}_{t}caligraphic_I start_POSTSUBSCRIPT italic_t endPOSTSUBSCRIPT<math alttext="\mathcal{I}{t-N+1}" class="ltx_Math" display="inline" id="S1.F1.pic1.13.13.13.13.13.13.13.13.13.13.13.13.13.1.1.1.1.1.1.1.1.1.m1.1">ℐt−N+1subscriptℐ𝑡𝑁1\mathcal{I}_{t-N+1}caligraphic_I start_POSTSUBSCRIPT italic_t - italic_N + 1 end_POSTSUBSCRIPT⋯⋯\mathbf{\cdots}⋯⋯⋯\mathbf{\cdots}⋯CNNCNNTCNCommandptℒP⁢G⁢(θ)superscriptℒ𝑃𝐺𝜃\mathcal{L}^{PG}(\theta)caligraphic_L start_POSTSUPERSCRIPT italic_P italic_G end_POSTSUPERSCRIPT ( italic_θ )

Figure 1: Overview of our approach. We demonstrate visuomotor policy learning in three different stages. In stage I, we train a state-based teacher policy using RL. Then in stage II we used the teacher policy and perceptual inputs to perform an imitation learning. In stage III, we bootstrap the actor using the student policy from stage II and fine-tune the policy through vision-based RL.
github-actions[bot] commented 4 months ago

Hello @anish-bhattacharya, thanks for the issue report! We are reviewing your report and will address it as soon as possible.