asm rejects packets with policy in transparent mode

If the policy is in transparent mode, and if blocking is disabled for all violations, then there must be some other existing condition that is causing the issue. Is there anything other than the ASM policy applied to the virtual server? Do you have any other profiles applied to the virtual server, or are you using a mitigation that injects JavaScript into responses--think web scraping and/or proactive bot defense. Can you verify that packets are traveling from the client to the BIG-IP? Do you have SNAT/Auto Map configured on the virtual server?