在花小茶馆🍵
Downloaded a safetensors file, this is almost certainly not trained. Some claude plots of the distribution of the weights vs a deepseek model. There was effort put in, but it's fake, and for no real purpose. ▎Source
Downloaded a safetensors file, this is almost certainly not trained. Some claude plots of the distribution of the weights vs a deepseek model. There was effort put in, but it's fake, and for no real purpose. ▎Source