MOSS-Image Base & heal on Turing (INT8 + FP16)
MOSS-Image-Base and MOSS-Image-heal don't runs at full speed on a Turing GPU (Ex: RTX 2060 SUPER) because they don't have native BF16. On Turing, BF16 is emulated via FP32, and ComfyUI's default loader keeps INT8 linears on INT8 but runs everything else (attention, norms, RoPE, residual) in FP32.
I collected the info and weights needed to run it as INT8 weights + FP16 everything else, and what to expect in terms of visual diff.
Big thanks to KBlueLeaf for KBlueLeaf/MOSS-Image-FP16-fixed & KBlueLeaf/MOSS-Image-heal.
Files involved
| File | Origin | Contents | Role |
|---|---|---|---|
moss_image_base_int8_convrot.safetensors |
Comfy-Org | 168Γ I8 (layers.N.proj.{qkv,o,ffn_up_gate,ffn_down}), rest BF16 |
official base INT8 (unfolded) |
moss_image_base_fp16_fixed.safetensors |
KBlueLeaf/MOSS-Image-FP16-fixed | all F16 (base weights folded Γc = 2β»β΄) | base-folded weights; outputs identical to the official base up to FP16 rounding |
moss_image_base_int8_fp16.safetensors |
local splice | 168Γ I8 official + 210 F16 from FP16-fixed | official base (folded) INT8 + FP16 β stream peak ~3.9k |
moss_image_heal_int8_convrot.safetensors |
KBlueLeaf/MOSS-Image-heal | 168Γ I8, rest BF16 | heal INT8 |
moss_image_heal_fp16.safetensors |
KBlueLeaf/MOSS-Image-heal | all F16 (no fold; global max abs 12.69) | heal weights |
moss_image_heal_int8_fp16.safetensors |
local splice | 168Γ I8 from heal INT8 + 210 F16 from heal FP16 | heal INT8 + FP16 β stream peak ~3.1k |
Local splice (recipe)
INT8 official + non-quantized part taken from the FP16 of the same model β byte-level offset copy (no re-quantization):
I8 codes + weight_scale F32 + comfy_quant U8 + comfy_attention.config U8: from official INT8
210 BF16 β F16 + rope.bands F32: from the corresponding FP16
Technical notes:
- The fold (c=2β»β΄) does not touch the 168 INT8 linears β the official INT8 codes are exactly the codes of the folded base.
rope.bandsis bit-identical in all variants.- The INT8 tensors are rotated by ConvRot (Hadamard 256,
H^4ofH4/2, entries Β±1/16). The correct comparison is against the dense weight with the same rotation applied; with that rotation, INT8-vs-dense error is 0.85β0.94 % rel. L2 β the expected INT8 quantization error. (A direct, unrotated comparison gives ~1.37 and is not an error.) video_embedder.weightdrops to 0.111 in the base splice vs 1.77 in the official INT8 β confirms the resulting model carries the folded weights (stream peak ~3.9k).
Benchmarks (same test for all runs)
| Model | Loader | s/it | Comment |
|---|---|---|---|
moss_image_heal_int8_fp16 |
MOSS Image Loader (FP16) node | 4.76 | full fast path |
moss_image_heal_int8_convrot |
default | 6.58 | native INT8 linears, rest in FP32 |
moss_image_base_int8_convrot |
default | 6.64 | official base, compute FP32 |
moss_image_base_int8_convrot |
MOSS Image Loader (FP16) node + clamp ON | 4.80 | official (unfolded) weights in FP16, residual clamp saturates at 65504 |
moss_image_base_int8_fp16 |
MOSS Image Loader (FP16) node + clamp ON | 4.78 | folded weights (stream peak ~3.9k), clamp not needed |
Visual diff
Common setup for all tests: 25 steps Β· seed 614001632771110 Β· 1936Γ1088 (16:9, 2 MP) Β· CFG 3 Β· scheduler simple Β· sampler ER-SDE (max-stage 3, eta 1.0, s_noise 1.10) Β· SLG off Β· no LoRAs Β· text encoder qwen3.5_9b_w4a8_gptq Β· VAE flux2-vae_small.
Prompt A (dog - day)
Prompt B (ducks - night)
MOSS Loader (FP16) node
Node: https://github.com/mkrgg/ComfyUI-MossFP16
The node ComfyUI-MossFP16 replaces MOSSImage.supported_inference_dtypes during load (equivalent to the dtypes half of ComfyUI #17684, without editing files) and registers its own cached_patcher_init so reloads keep the override. Optional clamp_FP16 toggle that adds a per-block hook doing x.clamp_(-65504, 65504) β a safety net for the unfolded base weights (not needed for FP16-fixed / heal weights).
When does MOSS go NaN (black image) in FP16?
Only with moss_image_base_int8_convrot in FP16 mode. Three measured points:
| Resolution | MP | Stream peak | Status |
|---|---|---|---|
| 1024Β² | 1.05 | 61.9k | ok (288 runs, none non-finite) |
| 1536Β² | 2.36 | 62.6k | ok |
| 1936Γ1088 | 2.11 | ~62.6k (fit) | ok |
| 2048Β² | 4.19 | 64.3k | ok (1.8% margin) |
| 2240Β² | 5.02 | 64.9k (fit) | ok (1%) |
| 2368Β² | 5.61 | 65.3k (fit) | risky depending on seed |
| 2496Β² | 6.23 | 65.8k (fit) | OVERFLOW |
| 2560Β² | 6.55 | 65,504 β black at step 29 of 50 (44/100 non-finite forwards) | OVERFLOW |
| 3072Β² | 9.44 | 68.3k (fit) | OVERFLOW |
Least-squares linear fit: peak β 60.97k + 0.773k Γ MP (residuals Β±0.2k). Critical point 5.86 MP (2420Β²). The first non-finite forward lands at roughly 56β58% of the schedule (29/50 β 58%, 56/100 β 56%) β with 25 ER-SDE steps expect it around step 14β15.
Caveats:
- Variance of Β±0.5β1% by seed/prompt/stochasticity β the boundary moves by a few dozen pixels per side.
- CFG barely affects it (<0.5%); the area is what matters. Long prompts also don't matter (Β±1%).
- With clamp: no NaN β the stream saturates at Β±65,504 β artifacts, not black.
- LoRA / finetunes raise the stream: the 2048Β² margin is just 1.8%, a LoRA adding even +5β10% can push it to black; LoRA + large images β go with the folded base or the heal.
- With the folded base (
moss_image_base_INT8_FP16) there is ~16Γ headroom β this failure mode does not exist.
References
- KBlueLeaf/MOSS-Image-heal β heal
- KBlueLeaf/MOSS-Image-FP16-fixed β FP16-fixed (folded base)
- KohakuBlueleaf/KohakuMossTrainer β KohakuMoss (concepts: numerics, massive channels, stream healing)
- Comfy-Org/ComfyUI #17522 β FP16 for Qwen-Image 2.1 / FIBO2 (MOSS excluded: stream peaks at 62β64k)
- Comfy-Org/ComfyUI #17684 β FP16 for MOSS + clamp (unmerged)
- OpenMOSS-Team/MOSS-Image-Base β official base
AI
The custom node and the README was created using Inklink 2 Small.
Model tree for mkrgg/MOSS-Image-Base-INT8-FP16-Turing
Base model
OpenMOSS-Team/MOSS-Image-Base








