Spaces:
Running on Zero
Running on Zero
Commit History
-
Update app.py Running
Verified signature8580bc6 -
Update app.py
Verified signaturef6f9a22 -
Update app.py
Verified signature9d53740
-
Update app.py
Verified signature2e2333e -
Upload film_net.py
Verified signature2070802 -
Update app.py
Verified signature8183acc -
Upload pk_workflow.py
Verified signature0925136 -
Update app.py
Verified signature4649024 -
Update app.py
Verified signature36aa8df -
Update app.py
Verified signatured0cfcec -
Update app.py
Verified signature7cf7fe0 -
Update app.py
Verified signature503efa7 -
Update app.py
Verified signature027bb7f -
Update app.py
Verified signature1a8274b -
Update app.py
Verified signaturea1be217
-
feat(app): add auto-canvas selection and update conditioner space
Verified signature- Update `CONDITIONER_SPACE` default to `mpasila/qwen3vl-conditioner`. - Add 3:2 and 2:3 aspect ratio options to `CANVASES`. - Implement `auto_canvas` logic to automatically select the closest canvas aspect ratio based on uploaded image or video dimensions. - Bind `auto_canvas` to image and video component change events in the Gradio UI.35d08d7 -
feat: add Turbo LoRA presets and ComfyUI support
Verified signature- Update Turbo LoRA presets to use EMA versions for better quality. - Add support for ComfyUI-trained LoRAs via on-the-fly module remapping. - Update README and app introduction to highlight 4-8 step Turbo LoRA speedups.49c5b9b -
feat(lora): add ComfyUI MiniMax-H3 LoRA conversion
Verified signatureImplement automatic remapping of ComfyUI-format MiniMax-H3 Turbo LoRAs to diffusers module names. - Added `_is_comfyui_lora` to detect ComfyUI naming conventions. - Added `_convert_comfyui_lora` to handle: - Block prefix remapping (`blocks` -> `transformer_blocks`). - Token refiner path updates. - Splitting fused QKV projections into separate `to_q`, `to_k`, and `to_v` weights. - Remapping MLP and output projection keys. - Integrated conversion into `apply_loras` to support Turbo LoRAs without manual conversion.b3babf9 -
Upload app.py
Verified signaturee7c3a50
-
Update app.py
Verified signature9c9e9f5 -
Update app.py
Verified signaturebfeb6c8 -
Update README.md
Verified signature46deca9
-
Upload app.py
Verified signature0f38077 -
Upload app.py
Verified signature4762142 -
Update requirements.txt
Verified signature2c8da25 -
Upload app.py
Verified signatured0795e0
-
Open badge links in a new tab
Verified signature3320577 -
Price bookings against the AoTI blocks, generate from 2 s again, cross-link the demos
get_duration is refit on the measured AoTI per-step table, AoTI is opt-in behind H3_AOTI with a card and torch check before it loads, and the duration slider starts at 2 s again by lowering the pipeline's own floor.92693c1 -
Start the duration slider at the 5 s the model actually generates
Verified signature74a631a -
Restore probe
Verified signature97142dd -
Cache the conditioner client with functools.cache, tidy comments
Verified signature1f48523 -
Let gradio_client forward the caller's ZeroGPU token itself instead of pinning one by hand
Verified signatureb2bc645 -
Forward the caller's ZeroGPU token to the conditioner, plainly
Verified signatureafea884 -
Bill the conditioner call to the requesting user's own ZeroGPU token, never an org token
Verified signaturea4b3082 -
Pay for the conditioner call with the first ZeroGPU identity that can: the caller's token, this Space's HF_TOKEN, then none
Verified signature9539b35 -
Fall back to an anonymous conditioner call when the forwarded ZeroGPU token is refused
Verified signaturee0f6bde -
Forward the caller's ZeroGPU token to the conditioner so one request bills as one request
Verified signaturee76d22a -
Load the AoTI packages from the public multimodalart/minimax-h3-aoti, move the upsample toggle under the prompt
Verified signature4218806 -
Migrate to canonical diffusers PR 14371 + public MiniMaxAI/MiniMax-H3 weights, add prompt upsampling toggle
Verified signature705b884
-
Update README.md
Verified signature2195330 -
Duration slider from 2s
Verified signatured9f8e44 -
Lower minimum duration to 2s
Verified signature38389da -
The keyframe half moved to multimodalart/minimax-h3
Verified signaturedeb4167 -
Mirror the conditioner's canonical table, grouped by ratio
Verified signature6fe31a4 -
Update README.md
Verified signature85b133d -
Mirror the conditioner's 16-canvas table (fast 4:3, 3:4 and 21:9 added)
Verified signature2225cd2 -
Point at the public multimodalart/qwen3vl-conditioner, called without a token
Verified signature63f7c91 -
Point at the public multimodalart/qwen3vl-conditioner, called without a token
Verified signature1f02c70 -
Fast canvas stays the default; the examples carry full-resolution canvases
Verified signature1aa28b5 -
Full 16:9 as the default canvas; match the conditioner's renamed fast label
Verified signature6237c6a