Search for: All records

Vision-language models are integral to computer vision research, yet many high-performing models remain closed-source, obscuring their data, design and training recipe. The research community has responded by using distillation from black-box models to label training data, achieving strong benchmark results, at the cost of measurable scientific progress. However, without knowing the details of the teacher model and its data sources, scientific progress remains difficult to measure. In this paper, we study building a Perception Language Model (PLM) in a fully open and reproducible framework for transparent research in image and video understanding. We analyze standard training pipelines without distillation from proprietary models and explore large-scale synthetic data to identify critical data gaps, particularly in detailed video understanding. To bridge these gaps, we release 2.8M human-labeled instances of fine-grained video question-answer pairs and spatio-temporally grounded video captions. Additionally, we introduce PLM-VideoBench, a suite for evaluating challenging video understanding tasks focusing on the ability to reason about "what", "where", "when", and "how" of a video. We make our work fully reproducible by providing data, training recipes, code & models. 

60 and 120 nm thick epitaxial films of isotopically enriched bcc iron (α-57Fe) grown on (100) MgO substrates are studied using x-ray diffraction, reflection high-energy electron diffraction (RHEED), and conversion electron Mössbauer spectroscopy (CEMS). X-ray diffraction and RHEED data indicate that each film behaves as a single crystal material consistent with the relative intensity ratios of the spectral lines observed in the CEMS spectrum. Data further confirm that the easy axis of magnetization lies along the ⟨100⟩ family of directions of the cubic α-iron film. The relevant theory to understand the relative intensities in a magnetic Mössbauer spectrum is outlined and is applied to interpret the intensity ratio of the Mössbauer spectral lines of a more complex hexaferrite magnetic system, BaFe12O19, grown on a single crystal substrate of Sr1.03Ga10.81Mg0.58Zr0.58O19. The conclusion that the magnetic moment in (0001)-oriented epitaxial BaFe12O19 film lies perpendicular to the plane of the substrate is deduced from the absence of the second and fifth lines by comparing the CEMS spectrum of the epitaxial (0001) BaFe12O19 film with the spectrum of a polycrystalline BaFe12O19 powder. Our measurements using CEMS corroborate what is known about the direction of the magnetic easy axis in α-iron and BaFe12O19 and motivate the use of CEMS to probe more complex atomically engineered epitaxial heterostructures, including superlattices.