Spaces:
Running on Zero
Running on Zero
Commit History
-
Port the studio from gradio.Server + custom index.html to gr.Workflow
- app.py: replace Server/@app.api/FastAPI routes with a gr.Workflow canvas (workflow.json) bound to a generate_video fn operator; x-ip-token now forwarded via LocalContext. All pipeline features unchanged: NCII guard, larry/lightx/off LoRAs, keyframe cover-crop/canvas-fit, GPU duration estimator, AoTI blocks. - workflow.json: the canvas graph (references -> generate_video -> subjects). - requirements.txt: Gradio from main via the prebuilt pypi-previews wheel at a9ce60a (git installs ship no frontend assets); README sdk_version bumped to 6.25.0 to match, hf_oauth: true for owner canvas editing. - index.html removed: the Workflow canvas replaces the custom studio.89d839f
-
[Admin] Modify name for distinction with previous demos (#1)
- [Admin] Modify name for distinction with previous demos (2bfc8cd189c701d29cfc1ad5501166fc29298432)6597d77
-
Add the lightx2v Minimax-h3-Turbo LoRA as a per-request alternative
h3_lora.py now loads both turbo LoRAs and switches the folded set in place per request (unfold old / fold new, one bf16 rounding, AoTI-safe). The lightx checkpoint is PEFT-style against the diffusers tree (rank 128, alpha 8 -> fold scale 0.0625); the larry one keeps the reference key conversion. The UI checkbox becomes a LoRA dropdown (larry / lightx / off) that suggests each design point in steps (6 / 4 / 28), and the API gains a 'lora' string parameter — the legacy use_lora bool still works.456aa27
-
Default steps 4 -> 6
942d78c -
Switch to the recommended ckpt850 EMA LoRA; default steps back to 4
ckpt850 is the final checkpoint of the round and sharp at 4 steps per the updated card, so the UI/API default drops 8 -> 4. Also adds H3_LORA_STRENGTH (default 1.0), the card's sharpness/artifact dial, applied at fold time and in the per-request toggle.4512613 -
Default steps 6 -> 8 (better quality at the ckpt500 weights)
e35aff5 -
Default to the ckpt500 turbo weights, add a per-request LoRA toggle
- H3_LORA default is now minimax_h3_turbo_4step_ckpt500.safetensors, the model card's recommended default (newest, sharpest, non-EMA) - the low-rank factors stay resident after the startup fold, so h3_lora.set_enabled() folds/unfolds B@A in place per request; the AoTI blocks read the same live storage. UI gets a 'Turbo LoRA' checkbox (default on), the API a use_lora parameter (default true) - default steps 4 -> 6 per the card's updated guidance (comfort zone is 6-8 at this checkpoint; 4 is softer)868ff3a -
Rebuild the UI as a custom studio frontend on gradio.Server
Server mode: @app.api(name="generate") keeps the queue / ZeroGPU / gradio_client contract (same name and signature), @app.get("/") serves a self-contained index.html — dark studio layout with a control deck (prompt, keyframe dropzones, canvas / duration / steps / seed), a preview monitor, and a transport-style report bar, driven by the @gradio/client JS package. /status and /config feed the readiness pill and the canvas list. Keyframe cover-crop moved server-side into generate so API callers get it too.299d59b -
Fold the 4-step Turbo LoRA into the transformer, default to 4 steps
Downloads larryvrh/MiniMax-H3-Turbo-Lora at startup and merges W + B@A into the bf16 weights, mirroring the diffusers key conversion (fused-QKV thirds, SwiGLU gate/value swap, shared AdaLN row layout). Folded before the AoTI package is patched in so compiled blocks carry the update. H3_LORA / H3_LORA_REPO configure it, H3_LORA=off disables.d8b48ff
-
Cache the conditioner client with functools.cache, tidy comments
Verified signature5029048 -
Let gradio_client forward the caller's ZeroGPU token itself instead of pinning one by hand
Verified signaturebf1199a -
Bill the conditioner call to the requesting user's own ZeroGPU token, never an org token
Verified signature9a812e9 -
Pay for the conditioner call with the first ZeroGPU identity that can: the caller's token, this Space's HF_TOKEN, then none
Verified signature8e346cd -
Fall back to an anonymous conditioner call when the forwarded ZeroGPU token is refused
Verified signature1ccbe3f -
Forward the caller's ZeroGPU token to the conditioner so one request bills as one request
Verified signature9c16554 -
Load the AoTI packages from the public multimodalart/minimax-h3-aoti, move the upsample toggle under the prompt
Verified signaturea0bf481 -
Migrate to canonical diffusers PR 14371 + public MiniMaxAI/MiniMax-H3 weights, add prompt upsampling toggle
Verified signature9da256f
-
Update README.md
Verified signatureef0bfcc -
Demo short description
Verified signaturecba1972 -
Rename to MiniMax H3
Verified signaturee63a181 -
Wire the AoTI load path (env-gated) and the sub-768p canvases
Verified signature578be89 -
Sync the split MiniMax-H3 Spaces
Verified signaturec938c2e -
Sync the split MiniMax-H3 Spaces
Verified signature5ca1dce -
Sync the split MiniMax-H3 Spaces
Verified signature0453841 -
Sync the split MiniMax-H3 Spaces
Verified signature186aa49 -
initial commit
Verified signatured11a565