CLI Reference
Command Shape To Run A Demo/Model
This section will only cover <System Arguments>. Refer to the Running a model section for more information on <Demo Preset> and <Demo Arguments>:
uv run flashdreams-run-v2 <Demo Preset> <System Arguments> -- <Demo Arguments>
System Arguments
General
| Argument | Description |
|---|---|
--timeout {SECONDS,unbound} |
Bound the whole application run, including initialization and replacement sessions. |
--total-model-steps {N,unbound} |
Bound model steps across the whole application run, including replacement sessions. |
--mode {mp4,webrtc,native-window} |
Select file, browser, or local-window presentation. Default: mp4. |
--stats-path PATH |
Write JSON model-step measurements and enable synchronized per-stage pipeline profiling. |
MP4 Specific
| Argument | Description |
|---|---|
--output-path PATH |
MP4 destination. Required in MP4 mode. |
WebRTC Specific
| Argument | Description |
|---|---|
--host HOST |
Interface to serve on. Default: 127.0.0.1; use 0.0.0.0 for remote clients. |
--port PORT |
Port to serve on. Default: 0, which selects an available port. |
Native Window Specific
| Argument | Description |
|---|---|
--window-title TITLE |
Native window title. Default: FlashDreams. |
Default Demo Session Overrides
These options override the default demo session values requested by a particular demo implementation. Omit them to use the application's default values.
| Argument | Description |
|---|---|
--pixel-width N |
Override generated-frame width. |
--pixel-height N |
Override generated-frame height. |
--fps N |
Override the generated-frame playback rate. |
--layout {tchw,btchw,bcthw,bvtchw} |
Override generated tensor layout. |
--backpressure-mode {block,drop_oldest} |
Choose full-queue behavior: block waits for capacity to retain generated chunks; drop_oldest discards the oldest queued chunk to favor recent output. |
--presentation-mode {on_demand,continuous} |
Choose rendering behavior: on_demand presents each selected model frame when it arrives; continuous keeps the UI responsive between model frames by reusing the newest frame. |
Examples
Stream to a browser:
uv run flashdreams-run-v2 DEMO_MODEL_SLUG --mode webrtc --host 0.0.0.0 --port 8089
Write every generated frame in order to an MP4:
uv run flashdreams-run-v2 DEMO_MODEL_SLUG --mode mp4 --output-path output.mp4 \
--timeout unbound --backpressure-mode block --presentation-mode on_demand
MP4 mode requires --timeout and/or --total-model-steps; pass unbound to
name an explicit unlimited policy. WebRTC and native-window modes may omit both
limits. When a session is replaced, it receives the time and model steps left
for the whole run.
Preload and validate a v2 application without opening a window:
uv run --no-sync flashdreams-run-v2 cam2v-lingbot \
--preload-application -- --example-data
Preload runs IApplication.init and one model block. It defaults to warning;
set FLASHDREAMS_PREPARATION_POLICY to none, warn, or error. Findings
are written to preparation_issues.txt in the FlashDreams cache, or to
FLASHDREAMS_PREPARATION_ISSUES_PATH when set. Use
--skip-preload-validation to skip the first-block validation.
Open a local native window:
uv run flashdreams-run-v2 DEMO_MODEL_SLUG --mode native-window --window-title FlashDreams