← Docs Home · Model Catalog · HTTP API · Troubleshooting

Flex.2-preview LoRA Inference with Diffusers (AI Toolkit-trained)

API model id: flex2
URL slug: flex2

This page documents the reference Diffusers inference pipeline for flex2 (Flex.2-preview). It is designed for running LoRAs trained with ostris/ai-toolkit while minimizing training preview vs inference mismatch. If you are trying to reproduce AI Toolkit sample previews, treat the code linked below as the source of truth (scheduler wiring, resolution snapping, LoRA application, and conditioning).

Run in the cloud (optional): If you want to reproduce the examples on this page in a pinned runtime without local CUDA/driver setup (and reduce preview‑vs‑inference drift), run it via RunComfy’s Cloud AI Toolkit (Train + Inference). 👉 You can open it here: Cloud AI Toolkit (Train + Inference)

Quick facts

Field	Value
Pipeline	`src/pipelines/flex2.py`
Base checkpoint	`ostris/Flex.2-preview`
Defaults	`sample_steps=25`, `guidance_scale=4.0`, `seed=42`
Resolution snapping	Floors width/height to a multiple of 32
Control image	No
LoRA scale behavior	Uses `fuse_lora()` (weights merged). Scale is fixed after load; changing `loras[].network_multiplier` triggers a reload.
Needs AI Toolkit	No

Reference implementation (source of truth)

Pipeline implementation: src/pipelines/flex2.py
Shared behaviors (snapping, seeding, LoRA base logic): src/pipelines/base.py
Request schema (parameter names): src/schemas/request.py
Model ids (enum): src/schemas/models.py
Pipeline registry (model → class): src/pipelines/__init__.py
Executor (prompt processing + caching): src/tasks/executor.py

Minimal API request

{
  "model": "flex2",
  "trigger_word": "sks",
  "prompts": [
    {
      "prompt": "[trigger] a photo of a person",
      "width": 1024,
      "height": 1024,
      "seed": 42,
      "sample_steps": 25,
      "guidance_scale": 4.0,
      "neg": ""
    }
  ],
  "loras": [
    {
      "path": "my_lora_job/my_lora.safetensors",
      "network_multiplier": 1.0
    }
  ]
}

Pipeline behavior that matters

Instantiates an AutoPipelineForText2Image custom pipeline (custom_pipeline="ostris/Flex.2-preview").
Overrides the scheduler to FlowMatchEulerDiscreteScheduler with shift=3.0 and dynamic shifting enabled.
Uses fuse_lora() + unload_lora_weights() to match the intended inference path for this checkpoint.

Preview-matching notes (training preview vs inference mismatch)

Flex.2 is sensitive to scheduler wiring. If you run with a different scheduler (or shift), outputs will drift quickly.
Because LoRA is fused, changing loras[].network_multiplier requires a pipeline reload.
Width/height are floored to a multiple of 32.

What to compare when debugging mismatch

Confirm the effective width/height after snapping (the server floors to the divisor).
Match sample_steps and the scheduler family (FlowMatch / UniPC / DDPM differences matter).
Match guidance_scale semantics (some pipelines map it to a different internal parameter).
Match loras[].network_multiplier and whether LoRA scale is dynamic vs fused.

Flex.1-alpha