Spaces:
Running on Zero
Running on Zero
Commit History
11 days ago
-
Link to my other Space, Graphics Enhancer; Showcase picks up replaced files Running
Verified signature2e77cb7 -
Transcribe: use Whisper long-form decoding for recordings over 30 seconds
Verified signature59e1b02 -
Transcribe: GPU time follows the length of the audio; also show the text with times
Verified signaturea9da61f -
Narration: all 28 English Kokoro voices, British pronunciation for the British ones
Verified signature5adb4c2
13 days ago
-
Prompt improver: accept one-line and two-line answers
Verified signaturea7d947c -
Remove the Chat tab and Gemma: this is a graphics studio
Verified signaturee400aec -
Start-up report of where files are stored
Verified signaturefb18834 -
Local Gemma off again: the packs disk ran out of space
Verified signatureee4c483 -
Free each model's downloads after sealing its pack; load Gemma 4 31B (8-bit) first
Verified signature1c79a19 -
Local Gemma chat off by default (it exceeded the 150 GB disk limit)
Verified signature7aa1810 -
Chat: Gemma 4 31B in 8-bit in its own pack; fix prompt improver parsing
Verified signaturef951038 -
Image model: FLUX.2 klein 9B (FLUX.1 schnell as fallback); fix the prompt improver
Verified signature974b690 -
Remove the Enhance tabs; add a prompt improver before video generation
Verified signature4865030
14 days ago
-
GPU-time requests set from half-card measurements
Verified signature8a9bd05
15 days ago
-
One ZeroGPU pack per model; each button loads only what it needs
Verified signaturef9638d8 -
Selective GPU loading + Gemma chat (OAuth) + Florence-2 image / SmolVLM2 video / CLAP music describe tabs; Qwen2.5-VL removed; Qwen3-4B for narration & prompt rewrite; README credits updated
Verified signature6f9c479 -
Selective GPU loading + Gemma chat (OAuth) + Florence-2 image / SmolVLM2 video / CLAP music describe tabs; Qwen2.5-VL removed; Qwen3-4B for narration & prompt rewrite; README credits updated
Verified signature486c5dc -
Selective GPU loading + Gemma chat (OAuth) + Florence-2 image / SmolVLM2 video / CLAP music describe tabs; Qwen2.5-VL removed; Qwen3-4B for narration & prompt rewrite; README credits updated
Verified signature0643043 -
Selective GPU loading + Gemma chat (OAuth) + Florence-2 image / SmolVLM2 video / CLAP music describe tabs; Qwen2.5-VL removed; Qwen3-4B for narration & prompt rewrite; README credits updated
Verified signatured6a11cf -
Selective GPU loading + Gemma chat (OAuth) + Florence-2 image / SmolVLM2 video / CLAP music describe tabs; Qwen2.5-VL removed; Qwen3-4B for narration & prompt rewrite; README credits updated
Verified signaturebf492b7
16 days ago
-
Enhance: cut the chatty opening and Markdown marks from model answers
Verified signature09affcc
17 days ago
-
Run Kokoro in its own process; image request 12 s
Verified signature4b6c588 -
Use the bucket: running totals and a Showcase tab
Verified signatureed98cb4 -
Image 8 s, enhance image 17 s
Verified signature780d87e -
Halve every GPU-time estimate
Verified signature3b6ea25 -
Enhance video: optional music prompt and narration
Verified signature7f5eab8 -
Add an Enhance tab: describe, rewrite and regenerate images, videos and music
Verified signaturec21a6b0 -
Add a Chat tab (Qwen2.5 7B), which also replaces the small narration planner
Verified signatured004811 -
Add a Narration tab (Kokoro)
Verified signature27d0e4f -
Remove an option this Gradio version does not accept
Verified signature7943346 -
Add a Transcribe tab (Whisper large v3 turbo)
Verified signature405495a -
Add a Music tab (MusicGen medium)
Verified signature1cce6dc
18 days ago
-
Two tabs (Image, one-button Video), sharper 1152x768 clips, better narration planner
Verified signatureaff8ac3 -
Model cache back on the local disk: the bucket reads too slowly
Verified signaturef6e9e12 -
Keep models on attached storage when it is available
Verified signaturee82501e -
Remove the loop cross-fade
Verified signatureb140f0e -
Music and narration work alone or together, in any order
Verified signature0dd66b6 -
Narration button stays locked until music has been added
Verified signature2363cac -
Only the phrase "slow motion" allows slow motion
Verified signature2589955 -
Avoid slow motion unless the motion text asks for it
Verified signaturede8e9a8 -
Sharper, real-time motion: frame-rate conditioning, motion-first prompt, extended negative prompt
Verified signature10f02ef -
Loop cross-fade lengthened to 2 seconds
Verified signature45a5aa0 -
Video no longer autoplays
Verified signature46071a0 -
MusicGen medium; narration needs no hint box
Verified signature8cf9cb0 -
Add narration: Qwen2.5 plans the lines and timing, Kokoro speaks them
Verified signature4880086 -
Add MusicGen large: soundtrack step for the animated clip
Verified signature2746dab -
LTX-Video 0.9.8 13B distilled on the full GPU, with seamless looping
Verified signaturebab0308 -
Cross-fade the end of each clip into its start for a seamless loop
Verified signaturec545212 -
10 fps clips with the first 8 frames trimmed
Verified signaturef4f8d1a -
LTX-Video 0.9.5 with scene+motion prompt and start-image noise, to fix frozen clips
Verified signature527b28a