[feat]: Wire creation studio config to backend session init and generation - #39
Draft
aryan5v wants to merge 6 commits into
Draft
[feat]: Wire creation studio config to backend session init and generation#39aryan5v wants to merge 6 commits into
aryan5v wants to merge 6 commits into
Conversation
Connect lobby model/mode/aspect/resolution/duration and reference images to session_init_v2 and project_init_v1. Backend parses creation config, validates mode assets, resolves frame size and segment cap, and passes per-step sampling overrides to the GPU worker. Frontend builds init payloads from creation state and applies echoed creation_config. Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>
|
Important Draft PR not reviewedDraft PRs are not automatically reviewed by default.
To automatically review draft PRs, update your CodeRabbit configuration: reviews:
auto_review:
drafts: trueThanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>
Use canonical t2va/fl2va/ref2va from hao-ai-lab#1834/hao-ai-lab#1835 instead of a parallel creation_mode field. Wire model, resolution, duration, and image assets through session_creation_config while leaving H3 mode routing to upstream. Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>
Combined stack verification (UI #38 + backend #39 + mock server)Tested on branch E2E (9/9 core checks passed):
Manual checks:
Fix included: Note: Full GPU |
Playwright was matching placeholder copy as visible text; the lobby composer only exposes it via the textarea placeholder attribute. Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>
Expose GET /creation-capabilities from backend and mirror it in the frontend so only supported models, modes (t2va/ref2va), resolutions (480p-1080p), and image references are selectable. Reject FL2VA, 4K, and invalid assets client-side and in session_creation_config. Lock in-session config pills read-only until live updates exist. Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>
Derive lobby model_ids from MODEL_REGISTRY and expose per-model creation capabilities so FastH3 appears in the composer alongside FastLTX models. H3 keeps fixed 16:9 / 720p-class geometry while LTX models retain the full aspect and resolution surface. Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Update: FastH3 in model list
The lobby model list now mirrors
MODEL_REGISTRYand includesfast-h3(MiniMax H3 / FastH3), which is already wired in the backend via the merged H3 integration.Per-model capabilities
fast-ltx2,fast-ltx23fast-h3GET /creation-capabilitiesnow returns amodelsmap with per-model caps. The UI clamps aspect/resolution when switching to H3.FL2VA and 4K remain rejected across models.