n_samples_per_prompt outputs per prompt to normalize advantages within a group. The naive way is N separate
requests, which encodes the same prompt N times and denoises N batches of size 1. Miles-diffusion instead sends one
request per microgroup: sglang-diffusion encodes the conditioning once, expands it engine-side, and denoises all N
outputs as a batch. Engine-side expansion is opt-in per model family in sglang-diffusion.
1. Microgroups
--rollout-microgroup-size M splits each prompt group of n_samples_per_prompt samples into requests of at most M
(generate_and_rm_group in miles/rollout/sglang_diffusion_rollout.py). Each request carries
num_outputs_per_prompt = M:
--sglang-server-concurrency.
Canonical values on main:
LTX-2.3 supports engine-side expansion, but the recipe generates one output per request: a multi-output video response
is large enough that deserializing and scoring it stops overlapping with other in-flight requests, so single-output
requests pipeline better end to end.
2. Seed layout
Rollout stays deterministic and collision-free per sample:- sgl-d expands one request’s
seedintoseed, seed+1, …, seed+M−1— one RNG stream per output. - The trainer spaces request seeds so streams never overlap:
group_index is monotonic across the run, so every (rollout, prompt-group, sample) triple gets a distinct seed. This
is what makes rollout replay and deterministic mode possible at microgroup granularity.
3. Pairs well with
- Streaming Reward and Deserialization — what happens to a microgroup response after generation.
- Deterministic Training — seed layout is half of run reproducibility.
- Core Concepts — where microgroup size sits among the batch knobs.