Fix canonical comparison resolutions

This commit is contained in:
Morpheus
2026-10-28 16:54:26 +00:00
parent 6dd16b8926
commit 6d0db5dea7
6 changed files with 19 additions and 6 deletions
+5
View File
@@ -105,3 +105,8 @@ I2I prompt; MultipleRef2I comparisons use
prompt. A model-specific matrix may vary settings such as resolution, steps, prompt. A model-specific matrix may vary settings such as resolution, steps,
CFG, sampler, scheduler, or seed, but it must not silently change the prompt CFG, sampler, scheduler, or seed, but it must not silently change the prompt
or reference inputs when claiming a cross-model comparison. or reference inputs when claiming a cross-model comparison.
Category comparison resolutions are also fixed: T2I uses `1088×1920`, I2I
uses a `1080×1920` output target, and MultipleRef2I uses `1920×1080`. A
separate resolution-scaling study may vary resolution, but it must be clearly
labelled as a scaling study rather than a cross-model comparison.
+5 -2
View File
@@ -7,9 +7,12 @@ canonical prompt is the shared I2I prompt in `../flux/workflow.json` (node
`74`). These inputs must remain fixed when comparing FLUX.3 Klein with other `74`). These inputs must remain fixed when comparing FLUX.3 Klein with other
models; only the tested generation settings may vary. models; only the tested generation settings may vary.
| ID | Encoder | Resolution | References | Sampler / scheduler | Purpose | The cross-model output target is fixed at **1080×1920** (the source fixture is
736×1308). Resolution changes are out of scope for this comparison row.
| ID | Encoder | Resolution | References | Sampler / scheduler | Status | Purpose |
|---|---|---:|---:|---|---| |---|---|---:|---:|---|---|
| F3K-I2I-01 | `qwen3-vl-4b-heretic_int8.safetensors` | 1024×1024 | 1 | Euler / simple | Single-reference conditioning | | F3K-I2I-01 | `qwen3-vl-4b-heretic_int8.safetensors` | 1080×1920 | 1 | Euler / simple | Planned | Single-reference conditioning |
Use `../character_reference.png`, the canonical I2I prompt, the installed Use `../character_reference.png`, the canonical I2I prompt, the installed
`flux3-image-klein-int8-convrot.safetensors`, seed `310101`, guidance `3.5`, `flux3-image-klein-int8-convrot.safetensors`, seed `310101`, guidance `3.5`,
+1 -1
View File
@@ -2,6 +2,6 @@
"model": "flux3-image-klein-int8-convrot.safetensors", "model": "flux3-image-klein-int8-convrot.safetensors",
"text_encoder": "qwen3-vl-4b-heretic_int8.safetensors", "text_encoder": "qwen3-vl-4b-heretic_int8.safetensors",
"vae": "flux2-vae.safetensors", "vae": "flux2-vae.safetensors",
"fixed": {"prompt_source":"../flux/workflow.json#74","reference_image":"../character_reference.png","seed":310101,"guidance":3.5,"steps":4}, "fixed": {"prompt_source":"../flux/workflow.json#74","reference_image":"../character_reference.png","width":1080,"height":1920,"seed":310101,"guidance":3.5,"steps":4},
"tests": [{"id":"F3K-I2I-01","mode":"i2i","width":1024,"height":1024,"references":1,"sampler":"euler","scheduler":"simple"}] "tests": [{"id":"F3K-I2I-01","mode":"i2i","width":1024,"height":1024,"references":1,"sampler":"euler","scheduler":"simple"}]
} }
+5 -2
View File
@@ -7,9 +7,12 @@ The canonical comparison fixtures are `../person_reference.png` and
MultipleRef2I prompt in `../flux/workflow.json` (node `74`). These inputs must MultipleRef2I prompt in `../flux/workflow.json` (node `74`). These inputs must
remain fixed for cross-model comparisons; only generation settings may vary. remain fixed for cross-model comparisons; only generation settings may vary.
| ID | Encoder | Resolution | References | Sampler / scheduler | Purpose | The cross-model output target is fixed at **1920×1080** (the source fixtures
are 1920×1072). Resolution changes are out of scope for this comparison row.
| ID | Encoder | Resolution | References | Sampler / scheduler | Status | Purpose |
|---|---|---:|---:|---|---| |---|---|---:|---:|---|---|
| F3K-MI2I-01 | `qwen3-vl-4b-heretic_int8.safetensors` | 1024×1024 | 2 | Euler / simple | Two-reference conditioning | | F3K-MI2I-01 | `qwen3-vl-4b-heretic_int8.safetensors` | 1920×1080 | 2 | Euler / simple | Planned | Two-reference conditioning |
Use both canonical reference images, the shared MultipleRef2I prompt, the Use both canonical reference images, the shared MultipleRef2I prompt, the
installed `flux3-image-klein-int8-convrot.safetensors`, seed `310101`, installed `flux3-image-klein-int8-convrot.safetensors`, seed `310101`,
@@ -2,6 +2,6 @@
"model": "flux3-image-klein-int8-convrot.safetensors", "model": "flux3-image-klein-int8-convrot.safetensors",
"text_encoder": "qwen3-vl-4b-heretic_int8.safetensors", "text_encoder": "qwen3-vl-4b-heretic_int8.safetensors",
"vae": "flux2-vae.safetensors", "vae": "flux2-vae.safetensors",
"fixed": {"prompt_source":"../flux/workflow.json#74","reference_images":["../person_reference.png","../scene_reference.png"],"seed":310101,"guidance":3.5,"steps":4}, "fixed": {"prompt_source":"../flux/workflow.json#74","reference_images":["../person_reference.png","../scene_reference.png"],"width":1920,"height":1080,"seed":310101,"guidance":3.5,"steps":4},
"tests": [{"id":"F3K-MI2I-01","mode":"mi2i","width":1024,"height":1024,"references":2,"sampler":"euler","scheduler":"simple"}] "tests": [{"id":"F3K-MI2I-01","mode":"mi2i","width":1024,"height":1024,"references":2,"sampler":"euler","scheduler":"simple"}]
} }
+2
View File
@@ -3,6 +3,8 @@
This is the canonical local FLUX.3 Klein text-to-image workflow and test This is the canonical local FLUX.3 Klein text-to-image workflow and test
matrix. New reference-based tests belong in `optimized/I2I/flux3klein/` and matrix. New reference-based tests belong in `optimized/I2I/flux3klein/` and
`optimized/MultipleRef2I/flux3klein/`; do not create a second T2I directory. `optimized/MultipleRef2I/flux3klein/`; do not create a second T2I directory.
The canonical cross-model comparison row is `1088×1920`; the other recorded
resolution rows are explicitly labelled scaling studies.
This directory contains the first controlled workflow and test plan for the This directory contains the first controlled workflow and test plan for the
server's locally installed FLUX 3 Image Klein model. The workflow does not use server's locally installed FLUX 3 Image Klein model. The workflow does not use