no result

[Ying-Zi66]

i use this model to convert pdf into mmd，but the mmd is [MISSING_PAGE_EMPTY:1]。no result

[lukas-blecher]

Duplicate of #7, #11

[Ying-Zi66]

You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference. 0%| | 0/1 [00:00<?, ?it/s]WARNING:root:Found repetitions in sample 0 [nltk_data] Error loading words: <urlopen error [Errno 11004] [nltk_data] getaddrinfo failed>

[lukas-blecher]

You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.

Looks like you are loading the model in a weird way. What are you running?

[Ying-Zi66]

i download from https://huggingface.co/nielsr/nougat/tree/main， i dowload this model too slowly

[lukas-blecher]

How did you run it?

[Ying-Zi66]

anaconda prompt (base) C:\Users\yingzi>nougat C:\Users\yingzi\Desktop\358.pdf --out C:\Users\yingzi\Desktop

[lukas-blecher]

Can you give more information about the pdf? What language is the file in? If it's not english, nougat won't be able to process it

[Ying-Zi66]

english,

[lukas-blecher]

Seems to work for me, see below. Please try to download the official release checkpoint: https://github.com/facebookresearch/nougat/releases/tag/0.1.0-small

Raw output

# Continuum Simulations of Hypersonic Flows in Chemical and Thermal Nonequilibrium R. M. Wagnild # Abstract The ability of the Navier-Stokes equations to capture the effects of strong chemical and thermal nonequilibrium on gas composition in a manner similar to the direct simulation Monte Carlo (DSMC) method is tested with the introduction of a nonequilibrium chemistry model. Although this chemistry model has the ability to reasonably reproduce measured Arrhenius rates in conditions of thermal equilibrium, it results in reaction rates that vary significantly for the most commonly used Arrhenius rates in conditions of thermal nonequilibrium. The nonequilibrium chemistry model is implemented in the three-temperature Navier-Stokes computational fluid dynamics (CFD) solver, Data-Parallel Line Relaxation (DPLR), and tested on high-altitude, hypersonic flow conditions. The results of the simulations show that the model predicts a greater amount of NO compared with previous Navier-Stokes using nonequilibrium quasi-classical trajectory derived rates and remains consistent with DSMC computations over altitudes ranging from 53.5 to 87.5 km. The most commonly used chemistry model for Navier-Stokes solvers was found unable to match this performance, indicating the importance of including nonequilibrium effects when modeling chemically reacting hypersonic flowfields. ## 1 Introduction ALTHOUGH the problem of calculating chemically reacting flowfields in the rarefied hypersonic regime has been studied since the Apollo era [1], the influence of thermal nonequilibrium on atmospheric reaction rates is still not fully understood. Multiple theoretical and empirical methods have been proposed that aim to capture the impact of a nonequilibrium energy distribution between the available energy modes on chemical reaction rates [2]. A common issue with these models is that they rely on experimental rate data taken at conditions that do not match the flight environment and therefore employ empirical fits extrapolated to flight conditions. The most widely used models in reacting, nonequilibrium, fluid simulations are Park's model [3] for computational fluid mechanics and Bird's total collision energy model [4] for direct simulation Monte Carlo (DSMC) simulations. Recently, Bird proposed new types of phenomenological chemical reaction models that aim to reproduce chemical reaction rates without using measured rate data to calibrate model parameters. The recently proposed quantum-kinetic (QK) model [5, 6], in its simplest form, can be based solely on properties of the colliding molecules. The QK theory sets the reaction rate based on the probability of a vibrational transition from the current state to the state corresponding to the activation energy of the reaction for all possible collisions. The sum of these probabilities is considered to correspond to the full reaction rate because there is zero probability of transition to a state with energy greater than the collision energy. In this form, QK theory assumes a zero-energy barrier and sets the activation energy equal to the heat of reaction. When there is a serious discrepancy, the QK result is often an overprediction that would be reduced by a finite energy barrier. Reliable information on the energy barrier is not generally available, and the rates would no longer be a function of the well-established molecular properties alone. The microscopic properties used in these reaction models include the available collision energy, dissociation energies, and quantized vibrational energy levels. These chemical reaction models link the chemical reaction process to the energy content of the vibrational modes of the colliding molecules. Application of chemical reaction procedures for collisions between molecules that could lead to endothermic reactions is conceptually straightforward. The models for these "forward" endothermic reactions and the principle of microscopic reversibility are then used to develop models for the corresponding "reverse" exothermic reactions. These models satisfy microscopic reversibility by balancing the fluxes into and out of each state and do not require any macroscopic rate information. For dissociation, ionization, and exchange reaction types, the models produce equilibrium reaction rates that are in good qualitative and quantitative agreement with the best available theoretical and measured or extrapolated reaction rates [7, 8, 6]. The differences between the most reliable reaction rates and the corresponding QK values are usually less than an order of magnitude, which typically is comparable to the uncertainties in measurements. Because the QK models predict reaction rates that can be based on the properties of colliding molecules, these models have the ability to predict reaction rates even under conditions of thermal nonequilibrium. The technical difficulties in extracting information from nonequilibrium flows suggest that a sufficient number of reliable measured values of the nonequilibrium rates are unlikely to become available. As long as this is the case, validation of chemistry models is dependent on measurements of observable quantities that depend on the combination of rates in particular models. Where measured equilibrium or nonequilibrium rates are available, the agreement the QK models have shown is so satisfactory that using QK rates would be preferable to estimated rates in the absence of measured data. Besides their conceptual simplicity and their apparent ability to reproduce known rates, a significant advantage of the QK models is that, although originally proposed for use in DSMC codes, these models allow for the derivation of closed-form solutions for the reaction rates. Thus, these rates, and in particular the nonequilibrium ones, can be used within the context of the Navier-Stokes equations. A significant portion of the nonequilibrium flow phase of a reentry flight takes place in the continuum regime, where the Navier-Stokes equations are valid to a very good approximation. This portion of the flight is particularly important due to higher densities, resulting in