The Digital Michelangelo Project was a pioneering initiative undertaken during the 1998–1999 academic year to digitize the sculptures and architecture of Michelangelo using advanced laser scanning technology. The project was led by a team of 30 faculty, staff, and students from Stanford University and the University of Washington, with the aim of creating high-resolution 3D models of Michelangelo's works for scholarly, educational, and preservation purposes. == Objectives == The primary goals of the Digital Michelangelo Project were: To apply recent advancements in laser rangefinder technology for digitizing large cultural artifacts. To create detailed digital archives of Michelangelo's sculptures and architectural spaces for future study and analysis. To explore potential educational and curatorial applications for 3D scanned data. === Artworks digitized === The project involved scanning several iconic works by Michelangelo, including: David The Unfinished Slaves (Atlas, Awakening, Bearded, and Youthful) St. Matthew The allegorical statues from the Medici tombs (Night, Day, Dawn, and Dusk) The architectural interiors of the Tribuna del David at the Galleria dell'Accademia and the New Sacristy in the Medici Chapels. == Technology and methodology == === 3D scanning === The project's primary scanner was a laser triangulation rangefinder mounted on a motorized gantry, custom-built by Cyberware Inc. The scanner used a laser sheet to project onto an object, capturing its shape through triangulation. Multiple scans were taken from various angles and combined into a single, detailed 3D mesh. The resolution achieved was fine enough to capture even Michelangelo's chisel marks, with triangles approximately 0.25 mm on each side. In addition to shape data, color data was captured using a spotlight and a secondary camera, enabling the creation of textured 3D models. === Data processing === The project developed a software suite for processing the scanned data. This included: Aligning and merging multiple scans into a seamless 3D model. Filling holes in the geometry caused by inaccessible areas. Correcting color data for lighting inconsistencies and shadowing. Non-photorealistic rendering techniques were also applied, highlighting surface features such as Michelangelo’s chisel marks for enhanced visualization. == Logistical challenges == The scale and complexity of the project presented several challenges: Data size: The dataset for David alone comprised 2 billion polygons and 7,000 color images, occupying 60 GB of storage. Artifact safety: Ensuring the safety of the statues during scanning required extensive crew training, foam-encased equipment, and collision-prevention mechanisms. == Applications and impact == The digitized models have numerous potential applications: Art history: Allowing precise measurements and geometric analysis, such as determining chisel types or evaluating structural balance. Education: Providing new ways to study art, including interactive viewing from unconventional angles and with custom lighting. Museum curation: Enhancing visitor experiences through interactive kiosks and virtual models. The project demonstrated the potential for 3D technology to preserve and disseminate cultural heritage. == Data distribution == The project's models are available through Stanford University for scholarly purposes, under strict licensing due to Italian intellectual property laws. === ScanView === To provide public access to the 3D models while respecting usage restrictions, the project developed ScanView, a client/server rendering system. ScanView allows users to view and interact with high-resolution 3D models without downloading the data. The client component consists of a freely available viewer program and simplified 3D models. Users can navigate these models locally, adjusting position, orientation, lighting, and surface appearance. When a user finalizes a view, the client queries a remote server for a high-resolution rendering of the model, which is sent back to overwrite the simplified version on the user’s screen. A typical query-response cycle takes 1–2 seconds, depending on network conditions. To protect the models from unauthorized reconstruction, the system employs several security measures, including: Encrypting queries Perturbing viewpoint and lighting parameters Adding noise and warping rendered images Compressing images before transmission ScanView operates on Windows-based PCs and provides access to selected models, including David and St. Matthew, as well as other artifacts such as fragments of the Forma Urbis Romae and items from the Stanford 3D Scanning Repository. == Sponsors == The Digital Michelangelo Project was supported by Stanford University, Interval Research Corporation, and the Paul G. Allen Foundation for the Arts.
Starlight Information Visualization System
Starlight is a software product originally developed at Pacific Northwest National Laboratory and now by Future Point Systems. It is an advanced visual analysis environment. In addition to using information visualization to show the importance of individual pieces of data by showing how they relate to one another, it also contains a small suite of tools useful for collaboration and data sharing, as well as data conversion, processing, augmentation and loading. The software, originally developed for the intelligence community, allows users to load data from XML files, databases, RSS feeds, web services, HTML files, Microsoft Word, PowerPoint, Excel, CSV, Adobe PDF, TXT files, etc. and analyze it with a variety of visualizations and tools. The system integrates structured, unstructured, geospatial, and multimedia data, offering comparisons of information at multiple levels of abstraction, simultaneously and in near real-time. In addition Starlight allows users to build their own named entity-extractors using a combination of algorithms, targeted normalization lists and regular expressions in the Starlight Data Engineer (SDE). As an example, Starlight might be used to look for correlations in a database containing records about chemical spills. An analyst could begin by grouping records according to the cause of the spill to reveal general trends. Sorting the data a second time, they could apply different colors based on related details such as the company responsible, age of equipment or geographic location. Maps and photographs could be integrated into the display, making it even easier to recognize connections among multiple variables. Starlight has been deployed to both the Iraq and Afghanistan wars and used on a number of large-scale projects. PNNL began developing Starlight in the mid-1990s, with funding from the Land Information Warfare Agency, a part of the Army Intelligence and Security Command and continued developed at the laboratory with funding from the NSA and the CIA. Starlight integrates visual representations of reports, radio transcripts, radar signals, maps and other information. The software system was recently honored with an R&D 100 Award for technical innovation. In 2006 Future Point Systems, a Silicon Valley startup, acquired rights to jointly develop and distribute the Starlight product in cooperation with the Pacific Northwest National Laboratory. The software is now also used outside of the military/intelligence communities in a number of commercial environments.
Sentence extraction
Sentence extraction is a technique used for automatic summarization of a text. In this shallow approach, statistical heuristics are used to identify the most salient sentences of a text. Sentence extraction is a low-cost approach compared to more knowledge-intensive deeper approaches which require additional knowledge bases such as ontologies or linguistic knowledge. In short, sentence extraction works as a filter that allows only meaningful sentences to pass. The major downside of applying sentence-extraction techniques to the task of summarization is the loss of coherence in the resulting summary. Nevertheless, sentence extraction summaries can give valuable clues to the main points of a document and are frequently sufficiently intelligible to human readers. == Procedure == Usually, a combination of heuristics is used to determine the most important sentences within the document. Each heuristic assigns a (positive or negative) score to the sentence. After all heuristics have been applied, the highest-scoring sentences are included in the summary. The individual heuristics are weighted according to their importance. === Early approaches and some sample heuristics === Seminal papers which laid the foundations for many techniques used today have been published by Hans Peter Luhn in 1958 and H. P Edmundson in 1969. Luhn proposed to assign more weight to sentences at the beginning of the document or a paragraph. Edmundson stressed the importance of title-words for summarization and was the first to employ stop-lists in order to filter uninformative words of low semantic content (e.g. most grammatical words such as of, the, a). He also distinguished between bonus words and stigma words, i.e. words that probably occur together with important (e.g. the word form significant) or unimportant information. His idea of using key-words, i.e. words which occur significantly frequently in the document, is still one of the core heuristics of today's summarizers. With large linguistic corpora available today, the tf–idf value which originated in information retrieval, can be successfully applied to identify the key words of a text: If for example the word cat occurs significantly more often in the text to be summarized (TF = "term frequency") than in the corpus (IDF means "inverse document frequency"; here the corpus is meant by document), then cat is likely to be an important word of the text; the text may in fact be a text about cats.
Tango (platform)
Tango (named Project Tango while in testing) was an augmented reality computing platform, developed and authored by the Advanced Technology and Projects (ATAP), a skunkworks division of Google. It used computer vision to enable mobile devices, such as smartphones and tablets, to detect their position relative to the world around them without using GPS or other external signals. This allowed application developers to create user experiences that include indoor navigation, 3D mapping, physical space measurement, environmental recognition, augmented reality, and windows into a virtual world. The first product to emerge from ATAP, Tango was developed by a team led by computer scientist Johnny Lee, a core contributor to Microsoft's Kinect. In an interview in June 2015, Lee said, "We're developing the hardware and software technologies to help everything and everyone understand precisely where they are, anywhere." Google produced two devices to demonstrate the Tango technology: the Peanut phone and the Yellowstone 7-inch tablet. More than 3,000 of these devices had been sold as of June 2015, chiefly to researchers and software developers interested in building applications for the platform. In the summer of 2015, Qualcomm and Intel both announced that they were developing Tango reference devices as models for device manufacturers who use their mobile chipsets. At CES, in January 2016, Google announced a partnership with Lenovo to release a consumer smartphone during the summer of 2016 to feature Tango technology marketed at consumers, noting a less than $500 price-point and a small form factor below 6.5 inches. At the same time, both companies also announced an application incubator to get applications developed to be on the device on launch. On 15 December 2017, Google announced that they would be ending support for Tango on March 1, 2018, in favor of ARCore. == Overview == Tango was different from other contemporary 3D-sensing computer vision products, in that it was designed to run on a standalone mobile phone or tablet and was chiefly concerned with determining the device's position and orientation within the environment. The software worked by integrating three types of functionality: Motion-tracking: using visual features of the environment, in combination with accelerometer and gyroscope data, to closely track the device's movements in space Area learning: storing environment data in a map that can be re-used later, shared with other Tango devices, and enhanced with metadata such as notes, instructions, or points of interest Depth perception: detecting distances, sizes, and surfaces in the environment Together, these generate data about the device in "six degrees of freedom" (3 axes of orientation plus 3 axes of position) and detailed three-dimensional information about the environment. Project Tango was also the first project to graduate from Google X in 2012 Applications on mobile devices use Tango's C and Java APIs to access this data in real time. In addition, an API was also provided for integrating Tango with the Unity game engine; this enabled the conversion or creation of games that allow the user to interact and navigate in the game space by moving and rotating a Tango device in real space. These APIs were documented on the Google developer website. == Applications == Tango enabled apps to track a device's position and orientation within a detailed 3D environment, and to recognize known environments. This allowed the creations of applications such as in-store navigation, visual measurement and mapping utilities, presentation and design tools, and a variety of immersive games. At Augmented World Expo 2015, Johnny Lee demonstrated a construction game that builds a virtual structure in real space, an AR showroom app that allows users to view a full-size virtual automobile and customize its features, a hybrid Nerf gun with mounted Tango screen for dodging and shooting AR monsters superimposed on reality, and a multiplayer VR app that lets multiple players converse in a virtual space where their avatar movements match their real-life movements. Tango apps are distributed through Play. Google has encouraged the development of more apps with hackathons, an app contest, and promotional discounts on the development tablet. == Devices == As a platform for software developers and a model for device manufacturers, Google created two Tango devices. === The Peanut phone === "Peanut" was the first production Tango device, released in the first quarter of 2014. It was a small Android phone with a Qualcomm MSM8974 quad-core processor and additional special hardware including a fisheye motion camera, "RGB-IR" camera for color image and infrared depth detection, and Movidius Vision processing units. A high-performance accelerometer and gyroscope were added after testing several competing models in the MARS lab at the University of Minnesota. Several hundred Peanut devices were distributed to early-access partners including university researchers in computer vision and robotics, as well as application developers and technology startups. Google stopped supporting the Peanut device in September 2015, as by then the Tango software stack had evolved beyond the versions of Android that run on the device. === The Yellowstone tablet === "Yellowstone" was a 7-inch tablet with full Tango functionality, released in June 2014, and sold as the Project Tango Tablet Development Kit. It featured a 2.3 GHz quad-core Nvidia Tegra K1 processor, 128GB flash memory, 1920x1200-pixel touchscreen, 4MP color camera, fisheye-lens (motion-tracking) camera, an IR projector with RGB-IR camera for integrated depth sensing, and 4G LTE connectivity. As of May 27, 2017, the Tango tablet is considered officially unsupported by Google. ==== Testing by NASA ==== In May 2014, two Peanut phones were delivered to the International Space Station to be part of a NASA project to develop autonomous robots that navigate in a variety of environments, including outer space. The soccer-ball-sized, 18-sided polyhedral SPHERES robots were developed at the NASA Ames Research Center, adjacent to the Google campus in Mountain View, California. Andres Martinez, SPHERES manager at NASA, said "We are researching how effective [Tango's] vision-based navigation abilities are for performing localization and navigation of a mobile free flyer on ISS. === Intel RealSense smartphone === Announced at Intel's Developer Forum in August 2015, and offered to public through a Developer Kit since January 2016. It incorporated a RealSense ZR300 camera which had optical features required for Tango, such as the fisheye camera. === Lenovo Phab 2 Pro === Lenovo Phab 2 Pro was the first commercial smartphone with the Tango Technology, the device was announced at the beginning of 2016, launched in August, and available for purchase in the US in November. The Phab 2 Pro had a 6.4 inch screen, a Snapdragon 652 processor, and 64 GB of internal storage, with a rear facing 16 Megapixels camera and 8 MP front camera. === Asus Zenfone AR === Asus Zenfone AR, announced at CES 2017, was the second commercial smartphone with the Tango Technology. It ran Tango AR & Daydream VR on Snapdragon 821, with 6GB or 8GB of RAM and 128 or 256GB of internal memory depending on the configuration.
Visual temporal attention
Visual temporal attention is a special case of visual attention that involves directing attention to specific instant of time. Similar to its spatial counterpart visual spatial attention, these attention modules have been widely implemented in video analytics in computer vision to provide enhanced performance and human interpretable explanation of deep learning models. As visual spatial attention mechanism allows human and/or computer vision systems to focus more on semantically more substantial regions in space, visual temporal attention modules enable machine learning algorithms to emphasize more on critical video frames in video analytics tasks, such as human action recognition. In convolutional neural network-based systems, the prioritization introduced by the attention mechanism is regularly implemented as a linear weighting layer with parameters determined by labeled training data. == Application in Action Recognition == Recent video segmentation algorithms often exploits both spatial and temporal attention mechanisms. Research in human action recognition has accelerated significantly since the introduction of powerful tools such as Convolutional Neural Networks (CNNs). However, effective methods for incorporation of temporal information into CNNs are still being actively explored. Motivated by the popular recurrent attention models in natural language processing, the Attention-aware Temporal Weighted CNN (ATW CNN) is proposed in videos, which embeds a visual attention model into a temporal weighted multi-stream CNN. This attention model is implemented as temporal weighting and it effectively boosts the recognition performance of video representations. Besides, each stream in the proposed ATW CNN framework is capable of end-to-end training, with both network parameters and temporal weights optimized by stochastic gradient descent (SGD) with back-propagation. Experimental results show that the ATW CNN attention mechanism contributes substantially to the performance gains with the more discriminative snippets by focusing on more relevant video segments. == Literature == Seibold VC, Balke J and Rolke B (2023): Temporal attention. Front. Cognit. 2:1168320. doi: 10.3389/fcogn.2023.1168320.
Learning curve (machine learning)
In machine learning (ML), a learning curve (or training curve) is a graphical representation that shows how a model's performance on a training set (and usually a validation set) changes with the number of training iterations (epochs) or the amount of training data. Typically, the number of training epochs or training set size is plotted on the x-axis, and the value of the loss function (and possibly some other metric such as the cross-validation score) on the y-axis. Synonyms include error curve, experience curve, improvement curve and generalization curve. More abstractly, learning curves plot the difference between learning effort and predictive performance, where "learning effort" usually means the number of training samples, and "predictive performance" means accuracy on testing samples. Learning curves have many useful purposes in ML, including: choosing model parameters during design, adjusting optimization to improve convergence, and diagnosing problems such as overfitting (or underfitting). Learning curves can also be tools for determining how much a model benefits from adding more training data, and whether the model suffers more from a variance error or a bias error. If both the validation score and the training score converge to a certain value, then the model will no longer significantly benefit from more training data. == Formal definition == When creating a function to approximate the distribution of some data, it is necessary to define a loss function L ( f θ ( X ) , Y ) {\displaystyle L(f_{\theta }(X),Y)} to measure how good the model output is (e.g., accuracy for classification tasks or mean squared error for regression). We then define an optimization process which finds model parameters θ {\displaystyle \theta } such that L ( f θ ( X ) , Y ) {\displaystyle L(f_{\theta }(X),Y)} is minimized, referred to as θ ∗ {\displaystyle \theta ^{}} . === Training curve for amount of data === If the training data is { x 1 , x 2 , … , x n } , { y 1 , y 2 , … y n } {\displaystyle \{x_{1},x_{2},\dots ,x_{n}\},\{y_{1},y_{2},\dots y_{n}\}} and the validation data is { x 1 ′ , x 2 ′ , … x m ′ } , { y 1 ′ , y 2 ′ , … y m ′ } {\displaystyle \{x_{1}',x_{2}',\dots x_{m}'\},\{y_{1}',y_{2}',\dots y_{m}'\}} , a learning curve is the plot of the two curves i ↦ L ( f θ ∗ ( X i , Y i ) ( X i ) , Y i ) {\displaystyle i\mapsto L(f_{\theta ^{}(X_{i},Y_{i})}(X_{i}),Y_{i})} i ↦ L ( f θ ∗ ( X i , Y i ) ( X i ′ ) , Y i ′ ) {\displaystyle i\mapsto L(f_{\theta ^{}(X_{i},Y_{i})}(X_{i}'),Y_{i}')} where X i = { x 1 , x 2 , … x i } {\displaystyle X_{i}=\{x_{1},x_{2},\dots x_{i}\}} === Training curve for number of iterations === Many optimization algorithms are iterative, repeating the same step (such as backpropagation) until the process converges to an optimal value. Gradient descent is one such algorithm. If θ i ∗ {\displaystyle \theta _{i}^{}} is the approximation of the optimal θ {\displaystyle \theta } after i {\displaystyle i} steps, a learning curve is the plot of i ↦ L ( f θ i ∗ ( X , Y ) ( X ) , Y ) {\displaystyle i\mapsto L(f_{\theta _{i}^{}(X,Y)}(X),Y)} i ↦ L ( f θ i ∗ ( X , Y ) ( X ′ ) , Y ′ ) {\displaystyle i\mapsto L(f_{\theta _{i}^{}(X,Y)}(X'),Y')}
Tesla Dojo
Tesla Dojo is a series of supercomputers designed and built by Tesla for computer vision video processing and recognition. It was used for training Tesla's machine learning models to improve its Full Self-Driving (FSD) advanced driver-assistance system. It went into production in July 2023. Dojo's goal was to efficiently process millions of terabytes of video data captured from real-life driving situations from Tesla's 4+ million cars. This goal led to a considerably different architecture than conventional supercomputer designs. In August 2025, Bloomberg News reported that the Dojo project had been disbanded, though it was restarted in January 2026. == History == Tesla operates several massively parallel computing clusters for developing its Autopilot advanced driver assistance system. Its primary unnamed cluster using 5,760 Nvidia A100 graphics processing units (GPUs) was touted by Andrej Karpathy in 2021 at the fourth International Joint Conference on Computer Vision and Pattern Recognition (CCVPR 2021) to be "roughly the number five supercomputer in the world" at approximately 81.6 petaflops, based on scaling the performance of the Nvidia Selene supercomputer, which uses similar components. However, the performance of the primary Tesla GPU cluster has been disputed, as it was not clear if this was measured using single-precision or double-precision floating point numbers (FP32 or FP64). Tesla also operates a second 4,032 GPU cluster for training and a third 1,752 GPU cluster for automatic labeling of objects. The primary unnamed Tesla GPU cluster has been used for processing one million video clips, each ten seconds long, taken from Tesla Autopilot cameras operating in Tesla cars in the real world, running at 36 frames per second. Collectively, these video clips contained six billion object labels, with depth and velocity data; the total size of the data set was 1.5 petabytes. This data set was used for training a neural network intended to help Autopilot computers in Tesla cars understand roads. By August 2022, Tesla had upgraded the primary GPU cluster to 7,360 GPUs. Dojo was first mentioned by Elon Musk in April 2019 during Tesla's "Autonomy Investor Day". In August 2020, Musk stated it was "about a year away" due to power and thermal issues. Dojo was officially announced at Tesla's Artificial Intelligence (AI) Day on August 19, 2021. Tesla revealed details of the D1 chip and its plans for "Project Dojo", a datacenter that would house 3,000 D1 chips; the first "Training Tile" had been completed and delivered the week before. In October 2021, Tesla released a "Dojo Technology" whitepaper describing the Configurable Float8 (CFloat8) and Configurable Float16 (CFloat16) floating point formats and arithmetic operations as an extension of Institute of Electrical and Electronics Engineers (IEEE) standard 754. At the follow-up AI Day in September 2022, Tesla announced it had built several System Trays and one Cabinet. During a test, the company stated that Project Dojo drew 2.3 megawatts (MW) of power before tripping a local San Jose, California power substation. At the time, Tesla was assembling one Training Tile per day. In August 2023, Tesla powered on Dojo for production use as well as a new training cluster configured with 10,000 Nvidia H100 GPUs. In January 2024, Musk described Dojo as "a long shot worth taking because the payoff is potentially very high. But it's not something that is a high probability." In June 2024, Musk explained that ongoing construction work at Gigafactory Texas is for a computing cluster claiming that it is planned to comprise an even mix of "Tesla AI" and Nvidia/other hardware with a total thermal design power of at first 130 MW and eventually exceeding 500 MW. In August 2025, Bloomberg News reported that the Dojo project was disbanded, though Musk announced it would be restarted in January 2026 with a new chip iteration. == Technical architecture == The fundamental unit of the Dojo supercomputer is the D1 chip, designed by a team at Tesla led by ex-AMD CPU designer Ganesh Venkataramanan, including Emil Talpes, Debjit Das Sarma, Douglas Williams, Bill Chang, and Rajiv Kurian. The D1 chip is manufactured by the Taiwan Semiconductor Manufacturing Company (TSMC) using 7 nanometer (nm) semiconductor nodes, has 50 billion transistors and a large die size of 645 mm2 (1.0 square inch). Updating at Artificial Intelligence (AI) Day in 2022, Tesla announced that Dojo would scale by deploying multiple ExaPODs, in which there would be: 10 Cabinets per ExaPOD (1,062,000 cores, 3,000 D1 chips) 2 System Trays per Cabinet (106,200 cores, 300 D1 chips) 6 Training Tiles per System Tray (53,100 cores, along with host interface hardware) 25 D1 chips per Training Tile (8,850 cores) 354 computing cores per D1 chip According to Venkataramanan, Tesla's senior director of Autopilot hardware, Dojo will have more than an exaflop (a million teraflops) of computing power. For comparison, according to Nvidia, in August 2021, the (pre-Dojo) Tesla AI-training center used 720 nodes, each with eight Nvidia A100 Tensor Core GPUs for 5,760 GPUs in total, providing up to 1.8 exaflops of performance. === D1 chip === Each node (computing core) of the D1 processing chip is a general purpose 64-bit CPU with a superscalar core. It supports internal instruction-level parallelism, and includes simultaneous multithreading (SMT). It doesn't support virtual memory and uses limited memory protection mechanisms. Dojo software/applications manage chip resources. The D1 instruction set supports both 64-bit scalar and 64-byte single instruction, multiple data (SIMD) vector instructions. The integer unit mixes reduced instruction set computer (RISC-V) and custom instructions, supporting 8, 16, 32, or 64 bit integers. The custom vector math unit is optimized for machine learning kernels and supports multiple data formats, with a mix of precisions and numerical ranges, many of which are compiler composable. Up to 16 vector formats can be used simultaneously. ==== Node ==== Each D1 node uses a 32-byte fetch window holding up to eight instructions. These instructions are fed to an eight-wide decoder which supports two threads per cycle, followed by a four-wide, four-way SMT scalar scheduler that has two integer units, two address units, and one register file per thread. Vector instructions are passed further down the pipeline to a dedicated vector scheduler with two-way SMT, which feeds either a 64-byte SIMD unit or four 8×8×4 matrix multiplication units. The network on-chip (NOC) router links cores into a two-dimensional mesh network. It can send one packet in and one packet out in all four directions to/from each neighbor node, along with one 64-byte read and one 64-byte write to local SRAM per clock cycle. Hardware native operations transfer data, semaphores and barrier constraints across memories and CPUs. System-wide double data rate 4 (DDR4) synchronous dynamic random-access memory (SDRAM) memory works like bulk storage. ==== Memory ==== Each core has a 1.25 megabytes (MB) of SRAM main memory. Load and store speeds reach 400 gigabytes (GB) per second and 270 GB/sec, respectively. The chip has explicit core-to-core data transfer instructions. Each SRAM has a unique list parser that feeds a pair of decoders and a gather engine that feeds the vector register file, which together can directly transfer information across nodes. ==== Die ==== Twelve nodes (cores) are grouped into a local block. Nodes are arranged in an 18×20 array on a single die, of which 354 cores are available for applications. The die runs at 2 gigahertz (GHz) and totals 440 MB of SRAM (360 cores × 1.25 MB/core). It reaches 376 teraflops using 16-bit brain floating point (BF16) numbers or using configurable 8-bit floating point (CFloat8) numbers, which is a Tesla proposal, and 22 teraflops at FP32. Each die comprises 576 bi-directional serializer/deserializer (SerDes) channels along the perimeter to link to other dies, and moves 8 TB/sec across all four die edges. Each D1 chip has a thermal design power of approximately 400 watts. === Training Tile === The water-cooled Training Tile packages 25 D1 chips into a 5×5 array. Each tile supports 36 TB/sec of aggregate bandwidth via 40 input/output (I/O) chips - half the bandwidth of the chip mesh network. Each tile supports 10 TB/sec of on-tile bandwidth. Each tile has 11 GB of SRAM memory (25 D1 chips × 360 cores/D1 × 1.25 MB/core). Each tile achieves 9 petaflops at BF16/CFloat8 precision (25 D1 chips × 376 TFLOP/D1). Each tile consumes 15 kilowatts; 288 amperes at 52 volts. === System Tray === Six tiles are aggregated into a System Tray, which is integrated with a host interface. Each host interface includes 512 x86 cores, providing a Linux-based user environment. Previously, the Dojo System Tray was known as the Training Matrix, which includes six Training Tiles, 20 Dojo Interface Processor cards across four host servers, and Ethernet-l