Kromoss
Kromoss is a full fine-tune of MOSS-Image-Base at 1024 px. It uses the same architecture, inference pipeline, text encoder (Qwen3.5-9B) and VAE (FLUX.2 small decoder) as MOSS-Image-SFT, and this repo has the same layout, so it is a drop-in replacement.
Params
5.1 B
Context
—
Downloads 30d
57 K
Likes
318
Download history
tracking started — chart appears after 7 days of snapshots (4 recorded)57 K downloads in the last 30 days
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| model.safetensors | bf16 | 80.3 GB | 89.6 GB | ❌ Won’t fit |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Specifications
- Parameters
- 5.1 B
- Tensor type
- BF16
- Licence
- apache-2.0
- First seen on the Hub
- 2026-10-19
- Training datasets
- undisclosed
- Added to our catalog
- 2026-11-07
Compare with any text-to-image model