Technical Updates 🤖 Machine-Learning(ML), CortexVirtual Machine(CVM) & Model RepresentationTool(MRT) R&D Add the accuracy and performance records of the pre-quantized model, and use the cvm-runtime framework to test the accuracy and time-consuming after quantization; ZkRollup 1. ZkContract adds poseidon hash calculation and merkle path verification, realizes the balance correctness verification of the user state…

Technical Updates 🤖 Machine-Learning(ML), CortexVirtual Machine(CVM) & Model RepresentationTool(MRT) R&D Update the MRT documentation, and improve the introduction and quantification of mnist-tutorial Zk Rollup 1. ZkContract refactors the code, adjusts the processing logic and reduces the size of the compiled bytecode of the main contract; 2. ZkContract adds contract-related functions (event, etc.) required for…

Technical Updates 🤖 Machine-Learning(ML), CortexVirtual Machine(CVM) & Model RepresentationTool(MRT) R&D Zk Rollup 1. ZkNode fixes the State Keeper processing logic, processes the received transactions as soon as possible, and optimizes the maximum interval block time; 2. ZkNode optimizes the redundancy problem of State Keeper empty transaction processing, greatly improves the execution speed of transactions…

Cortex NFTs showing up in your wallet without spending a fraction of your precious crypto asset enforces the !vibe of the collection. Some airdrops can even turn into blue-chip pieces that can make holders substantially wealthier than they were before (Mutant ape serum, anyone?) If you’re a newcomer to NFTs…

Technical Updates 🤖 Machine-Learning(ML), CortexVirtual Machine(CVM) & Model RepresentationTool(MRT) R&D 1. ZkNode deletes the redundant hash cache in Merkle Tree, which greatly reduces the amount of data stored in the database; 2. ZkNode adds breakpoints, exits quickly and safely and saves the state; 3. ZkNode adds a more realistic transaction simulation and concurrently generates a…

Technical Updates 🤖 Machine-Learning(ML), CortexVirtual Machine(CVM) & Model RepresentationTool(MRT) R&D 1. The yamrt preprocessing module implements the symbol map zero preset and convolution operator zero quantization for quantization_aware training 2. YaMRT adds ImageNet dataset training and dynamic quantization preparation ZkRollup 1. ZKNode updates the startup script, adds workload quick start test, proof server…

Technical Updates 🤖 Machine-Learning(ML), CortexVirtual Machine(CVM) & Model RepresentationTool(MRT) R&D 1. Add an operator-level variable-dimensional inference function to implement the variable-dimensional inference feature of art.tfm_ops 2. The git repository of the yamrt framework is initialized and related documents are maintained ZkRollup 1. Analyze the libfqfft code and implement the GPU calculation of the…

It’s been months since we released our Cortex NFT Collector Program The Cortex team would like to express our heartfelt gratitude to every user who holds Cortex and thank you all for being there along with Cortex’s growth. The Cortex team has now come up with a new Incentive Program…

Technical Updates 🤖 Machine-Learning(ML), CortexVirtual Machine(CVM) & Model RepresentationTool(MRT) R&D ML&MRT&CVM 研发 1. Finish training the stylegan2 model with NFT pictures. 2. Cifar series quantization accuracy tweaking in the MRT model; 3. MRT adds variable batch dimension inference at the Op, Graph level, and boosts prediction SCTF model adaption. ZkRollup 1. ZKNode resolves…