Hugging Face's logo

Spaces:
koura718
/
ideogram4
Runtime error

App Files Community
ideogram4

Commit History

  1. Fix CSS and remove queue

    multimodalart HF Staff
    Verified signature
    b6b3100

  2. Dynamic duration: 60s floor (keep 30% margin)

    multimodalart HF Staff
    Verified signature
    8a4d61f

  3. Dynamic GPU duration (per-step model); move remote upsample off-GPU

    multimodalart HF Staff
    Verified signature
    d00ab3b

  4. Update app.py

    multimodalart HF Staff
    Verified signature
    223d5e9

  5. Update app.py

    multimodalart HF Staff
    Verified signature
    672caae

  6. Update app.py

    multimodalart HF Staff
    Verified signature
    12442ca

  7. Add cached gr.Examples (lazy) + generate defaults + fp8 link; remove stale tmp/ and __pycache__/

    multimodalart HF Staff
    Verified signature
    19d7eb5

  8. Release: pip diffusers (PR #13860) + public model ideogram-ai/ideogram-4-nf4; drop bundled diffusers_src

    multimodalart HF Staff
    Verified signature
    de324ea

  9. Update app.py

    multimodalart HF Staff
    Verified signature
    888c872

  10. Re-enable AOTI (PR#6 artifact) + model-card blurb + title

    multimodalart HF Staff
    Verified signature
    d3ed6b5

  11. Update README.md

    multimodalart HF Staff
    Verified signature
    31d862b

  12. Update README.md

    multimodalart HF Staff
    Verified signature
    7387ea2

  13. RC header: 'first open-weights Ideogram model', 9.3B params; + dynamo-disable warning fix

    multimodalart HF Staff
    Verified signature
    4ef43f7

  14. Disable dynamo (TORCHDYNAMO_DISABLE) to silence outlines_core WON'T CONVERT spam

    multimodalart HF Staff
    Verified signature
    c6731f9

  15. Drop dequantize (AOTI off -> run nf4 directly)

    multimodalart HF Staff
    Verified signature
    332e802

  16. Bundle exactly PR #6 (df4eb9b) as the diffusers source

    multimodalart HF Staff
    Verified signature
    a1f24dc

  17. Remote Ideogram magic-prompt (default) + local Qwen fallback radio; lazy enhancer; AOTI off (recompiling)

    multimodalart HF Staff
    Verified signature
    4c728e2

  18. Pre-warm vec-isa probe in parent so aoti_blocks_load is instant per worker (was ~22s)

    multimodalart HF Staff
    Verified signature
    e04df5f

  19. Move bnb Params4bit.shape fix to a runtime shim in app.py; restore pristine bundled diffusers

    multimodalart HF Staff
    Verified signature
    4e809bd

  20. Fix bnb Params4bit.shape tuple (.numel via math.prod); unpin bitsandbytes for cu130

    multimodalart HF Staff
    Verified signature
    6cd5923

  21. Pin ML stack to validated versions (bitsandbytes 0.47.0 fixes Params4bit.shape tuple)

    multimodalart HF Staff
    Verified signature
    e283ee7

  22. Demo: native pipe.upsample_prompt + spaces.aoti_blocks_load (overlapped) + timing

    multimodalart HF Staff
    Verified signature
    87fa87f

  23. Bundle PR diffusers (yiyi-refactor-fused + native prompt upsampling)

    multimodalart HF Staff
    Verified signature
    168566a

  24. Apply precompiled AOTI block (dynamic L=16k) to both transformers; ~1.28x on 1024 turbo

    multimodalart HF Staff
    Verified signature
    e99a766

  25. Dequant nf4->bf16 in parent/CPU context (persists across ZeroGPU forks) + streamed upsampler progress

    multimodalart HF Staff
    Verified signature
    6ed28ce

  26. bf16 dequant only in warmup (no generate fallback) — surface warmup failures

    multimodalart HF Staff
    Verified signature
    013c185

  27. xlarge GPU + bf16 transformers + upsampling default-on + gr.JSON caption + seed-in-field + non-fatal warmup

    multimodalart HF Staff
    Verified signature
    392bb31

  28. Update app.py

    multimodalart HF Staff
    Verified signature
    cc19c6a

  29. Preload upsampler + pipeline on GPU via @spaces.GPU startup warmup

    multimodalart HF Staff
    Verified signature
    66efa02

  30. Update app.py

    multimodalart HF Staff
    Verified signature
    bc8eac2

  31. Update app.py

    multimodalart HF Staff
    Verified signature
    51f5d95

  32. Update app.py

    multimodalart HF Staff
    Verified signature
    5873472

  33. Update app.py

    multimodalart HF Staff
    Verified signature
    b0161b5

  34. Update app.py

    multimodalart HF Staff
    Verified signature
    04b66f9

  35. Update requirements.txt

    multimodalart HF Staff
    Verified signature
    0536f5a

  36. Graft hosted lm_head onto pipe encoder for upsampling (no second model)

    multimodalart HF Staff
    Verified signature
    a1f5c28

  37. Add prompt upsampling (Qwen3-VL+Outlines), 3 modes, v2 checkpoint, Citrus theme: v6.txt

    multimodalart HF Staff
    Verified signature
    2436f41

  38. Add prompt upsampling (Qwen3-VL+Outlines), 3 modes, v2 checkpoint, Citrus theme: requirements.txt

    multimodalart HF Staff
    Verified signature
    ffc2d42

  39. Add prompt upsampling (Qwen3-VL+Outlines), 3 modes, v2 checkpoint, Citrus theme: app.py

    multimodalart HF Staff
    Verified signature
    49e5d58

  1. Rebuild guidance_schedule to match step count

    multimodalart HF Staff
    Verified signature
    60f16ec

  2. Upload /tmp/ideogram4_space/app.py with huggingface_hub

    multimodalart HF Staff
    Verified signature
    5246183

  3. bnb: math.prod(shape) so tuple works

    multimodalart HF Staff
    Verified signature
    290c73c

  4. Drop ./diffusers_src from requirements

    multimodalart HF Staff
    Verified signature
    002edd6

  5. Use sys.path for bundled diffusers

    multimodalart HF Staff
    Verified signature
    4e05e87

  6. Embed diffusers PR source; install locally

    multimodalart HF Staff
    Verified signature
    b8c861f

  7. Initial Ideogram 4 NF4 demo

    multimodalart HF Staff
    Verified signature
    3cd11ff

  8. initial commit

    multimodalart HF Staff
    Verified signature
    3940a23