VTP Base F16d64
Minimax's VTP Base F 16d 64 is a visual tokenizer model that integrates contrastive, self-supervised, and reconstruction learning, demonstrating improved generation capabilities at the same FLOPs as other models like DiT. It is particularly effective in understanding and generation tasks, with a zero-shot accuracy of 73.28 and a generation score of 3.88. Notably, the model features a scalable architecture that allows for better representation learning and improved performance with increased parameters and data.
Input
Output
Context
-
Max Output
-
Parameters
295.6M
Technical Specifications
Input Modalities
Output Modalities
Capabilities
Estimated GPU Requirements for VTP Base F16d64
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
Related Tools
Data sourced from official provider APIs and documentation
Last updated: Jun 23, 2026
Start building with the right model.
Automatically route workloads to the right model for every task, every time.