// Mechanical adaptation of OA loss/rlPpoClipBwd.slang.
import attributes;
import storage;
struct PushConstants {
uint new_log_probability_index;
uint old_log_probability_index;
uint advantage_index;
uint gradient_index;
uint count;
float clip_epsilon;
};
[[vk::push_constant]] PushConstants push;
[[vk::binding(0, 0)]] RWByteAddressBuffer storage_buffers[];
[kernel_name("ppo_clipped_policy_backward")]
[domain("ml")]
[variant("generic")]
[dtype("f32")]
[status("experimental")]
[shader("compute")]
[numthreads(256, 1, 1)]
void main(uint3 dispatch_thread_id : SV_DispatchThreadID) {
uint index = dispatch_thread_id.x;
if (index >= push.count) {
return;
}
float ratio = exp(
load_f32(storage_buffers[push.new_log_probability_index], index)
- load_f32(storage_buffers[push.old_log_probability_index], index));
float advantage = load_f32(storage_buffers[push.advantage_index], index);
bool clipping_active = (advantage >= 0.0 && ratio > 1.0 + push.clip_epsilon)
|| (advantage < 0.0 && ratio < 1.0 - push.clip_epsilon);
float gradient = clipping_active ? 0.0 : -(advantage * ratio) / float(push.count);
store_f32(storage_buffers[push.gradient_index], index, gradient);
}