Jul 25

1. 16:13

   transformer: int8 DiT (attn+FFN+modulation, group 64)

   xocialize
   Verified signature

   677924a
2. 16:08

   text_encoder: int8 Qwen2.5-VL-7B language model (no vision tower)

   xocialize
   Verified signature

   bbab999
3. 16:07

   vae: unquantized AutoencoderKLQwenImage

   xocialize
   Verified signature

   ceea953
4. 16:06

   model_index.json

   xocialize
   Verified signature

   d70fe39
5. 16:06

   scheduler: static shift-3 FlowMatchEuler

   xocialize
   Verified signature

   5bfdb20
6. 16:06

   tokenizer (incl. converted tokenizer.json)

   xocialize
   Verified signature

   ffa4ba8
7. 16:06

   LICENSE (Apache-2.0, additional information per upstream)

   xocialize
   Verified signature

   fd66959
8. 16:06

   NOTICE: NVIDIA OML attribution + statement of quantization changes

   xocialize
   Verified signature

   c431224
9. 16:06

   README: int8 tier — measured quality, footprints, and why there is no 4-bit tier

   xocialize
   Verified signature

   1b13810
10. 16:06

    initial commit

    xocialize
    Verified signature

    a4ee298