sand pies

sandpies's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

New activity in FusionCow/Sulphur-3 6 days ago

Add the ComfyUI H3 corrector node, including a tunable apply node. The stock Apply node is unchanged and still loads the corrector with H3. Apply H3 Corrector (Tunable) can unload H3 for the corrector forward so a 32 GB card that already runs H3 can run the corrector without holding both networks at once.

#6 opened 6 days ago by

sandpies

New activity in FusionCow/Sulphur-3 7 days ago

Corrector on a 32 GB card: step window, budget, and a swap loader

1

#5 opened 7 days ago by

sandpies

Add rank-256 LoRA extracted from s3 (vs FL2VA bf16 base)

#3 opened 7 days ago by

sandpies

New activity in FusionCow/Sulphur-3 8 days ago

Add Ref2VA hybrid of s3 (bf16 + int8)

#2 opened 8 days ago by

sandpies

Add weight-only int8 quant of s3

#1 opened 8 days ago by

sandpies

New activity in sandpies/FastH3-OmniRef-V2-Hybrid-int8-ComfyUI 11 days ago

w4a8 ?

4

#1 opened 11 days ago by

davoodkharmanzar

New activity in TenStrip/LTX2.3-10Eros_Version-Testing 15 days ago

Add beta7 test2 W4A8 conservative GPTQ (activation-calibrated)

#39 opened 15 days ago by

sandpies

New activity in TenStrip/LTX2.3-10Eros_Version-Testing 17 days ago

Add beta7 W4A8 conservative quant (by cicalooo)

#31 opened 17 days ago by

sandpies

Add beta7 int8 (learned rounding, ConvRot)

#30 opened 17 days ago by

sandpies

New activity in sandpies/Wuli-MiniMax-H3-Turbo-ComfyUI 27 days ago

Reference image issue.

2

#2 opened 27 days ago by

dio011

Please provide an unpruned int8 model.

1

#1 opened 27 days ago by

dio011

New activity in Alissonerdx/h3upscale-4k about 1 month ago

Trained it end to end: h3upscale x4 / x2 LoRAs + a ComfyUI guide node

❤️ 1

1

#2 opened about 1 month ago by

sandpies