Refresh optimized model status documentation

This commit is contained in:
Morpheus
2026-10-28 16:56:28 +00:00
parent 6d0db5dea7
commit 333c01a7b4
5 changed files with 49 additions and 5 deletions
+5 -1
View File
@@ -1,6 +1,6 @@
# I2I workflows
Use [`flux/`](flux/) or [`qwen/`](qwen/) for one-reference-image transformations. The workflows infer the subject from the input image instead of naming a specific character or scene.
Use [`flux/`](flux/), [`qwen/`](qwen/), or [`flux3klein/`](flux3klein/) for one-reference-image transformations. The workflows infer the subject from the input image instead of naming a specific character or scene.
The shared reference asset and static variables are recorded in [`variables.json`](variables.json) and `character_reference.png`. The test plans live in each model README.
@@ -9,4 +9,8 @@ The shared reference asset and static variables are recorded in [`variables.json
Qwen I2I: [`qwen/`](qwen/) — Q-I2I-34, Euler CFG++, beta scheduler, 50 steps,
CFG 3.5, approximately 68.3 seconds.
FLUX.3 Klein: [`flux3klein/`](flux3klein/) — planned against the same
`character_reference.png` and shared I2I prompt at a fixed 1080×1920 output;
not yet run.
Store each I2I test under the selected model's `results/<test-id>.png` and `results/<test-id>.json` using the table ID as the label.
+6 -2
View File
@@ -1,7 +1,11 @@
# MultipleRef2I workflows
Use [`flux/`](flux/) for the current best multi-reference image composition workflow. It accepts a primary subject reference and an environment or scene reference. A Qwen MultipleRef2I workflow has not yet been created.
Use [`flux/`](flux/) or [`flux3klein/`](flux3klein/) for multi-reference image composition. Both use the same person and scene fixtures for controlled comparison.
The fixed reference assets and static variables are recorded in [`variables.json`](variables.json), `person_reference.png`, and `scene_reference.png`. The test plan lives in the Flux README.
Store each MultipleRef2I test under `flux/results/<test-id>.png` and `flux/results/<test-id>.json` using the table ID as the label.
Store each MultipleRef2I test under the selected model's `results/<test-id>.png` and `results/<test-id>.json` using the table ID as the label.
Current comparison status: Flux.2 Klein has an untested optimization matrix;
FLUX.3 Klein has a planned matrix at fixed 1920×1080; no cross-model winner
has been declared.
+7 -1
View File
@@ -7,12 +7,18 @@ This directory contains the best tested workflows for the available hardware. Ch
- [`T2I/`](T2I/) — text-to-image generation.
- [`flux/`](T2I/flux/) — Flux.2 Klein best baseline.
- [`qwen/`](T2I/qwen/) — Qwen Image 2.1 best baseline.
- [`flux3image/`](T2I/flux3image/) — FLUX.3 Klein local T2I matrix.
- [`ideogram/`](T2I/ideogram/) — Ideogram 4 text and layout matrix.
- [`I2I/`](I2I/) — one-reference-image transformation or editing.
- [`flux/`](I2I/flux/) — Flux.2 Klein photorealistic conversion.
- [`qwen/`](I2I/qwen/) — Qwen Image 2.1 photorealistic conversion.
- [`flux3klein/`](I2I/flux3klein/) — FLUX.3 Klein reference-image tests.
- [`MultipleRef2I/`](MultipleRef2I/) — multi-reference image composition.
- [`flux/`](MultipleRef2I/flux/) — Flux.2 Klein best multi-reference workflow.
- Qwen MultipleRef2I is not yet available.
- [`flux3klein/`](MultipleRef2I/flux3klein/) — FLUX.3 Klein multi-reference tests.
See [`STATUS.md`](STATUS.md) for the authoritative model/category inventory,
test status, canonical inputs, and comparison baselines.
## Reusable variables
+28
View File
@@ -0,0 +1,28 @@
# Optimized model/category status
This is the single inventory for the maintained `optimized/` workflows. A
model is not marked as a cross-model winner merely because it has a completed
matrix; visual review must use the same category fixtures and resolution.
| Category | Model/workflow | Canonical inputs | Resolution | Test state | Current position |
|---|---|---|---:|---|---|
| T2I | Flux.2 Klein | shared T2I prompt | 1088×1920 comparison target | Matrix/baseline recorded | Existing baseline |
| T2I | Qwen Image 2.1 | shared T2I prompt | 1088×1920 comparison target | 88/88 recorded | Existing best recorded in category README |
| T2I | FLUX.3 Klein | shared T2I prompt | 1088×1920 comparison target | 432-cell sampler matrix recorded | Visual cross-model ranking pending |
| T2I | Ideogram 4 | structured magazine prompt | varies by magazine matrix | 24-row matrix complete | Separate prompt family; not directly comparable to portrait rows |
| I2I | Flux.2 Klein | `character_reference.png` + shared I2I prompt | 1080×1920 category target | Matrix planned/untested | Candidate baseline |
| I2I | Qwen Image 2.1 | `character_reference.png` + shared I2I prompt | 1080×1920 category target | 34-row matrix recorded | Existing best recorded in category README |
| I2I | FLUX.3 Klein | same canonical I2I fixtures | 1080×1920 | Planned; scheduled | New candidate |
| MultipleRef2I | Flux.2 Klein | `person_reference.png` + `scene_reference.png` + shared prompt | 1920×1080 | Matrix planned/untested | Candidate baseline |
| MultipleRef2I | FLUX.3 Klein | same canonical multi-reference fixtures | 1920×1080 | Planned; scheduled | New candidate |
| R2V | MiniMax H3 Continuum | category-specific fixed prompt/references | 960×544 baseline | Mixed historical/planned matrix | Current video baseline |
## Comparison rules
- Keep prompts and reference assets fixed within a category.
- Keep the category resolution fixed for cross-model rows.
- Vary settings only in a labelled optimization or scaling study.
- Store one metadata record per test ID and keep the README table synchronized
with the actual result state.
- Historical results remain valid evidence but must identify their model and
input contract; they are not silently reclassified as current runs.
+3 -1
View File
@@ -1,6 +1,6 @@
# T2I workflows
Use [`flux/`](flux/), [`qwen/`](qwen/), or [`ideogram/`](ideogram/) for the best tested text-to-image baseline. Prompts, seeds, and dimensions are intended to be replaced per task; sampler, scheduler, CFG, and steps are the optimization controls.
Use [`flux/`](flux/), [`qwen/`](qwen/), [`flux3image/`](flux3image/), or [`ideogram/`](ideogram/) for text-to-image baselines. Prompts and seeds are fixed for cross-model comparison; sampler, scheduler, CFG, steps, and explicitly labelled scaling-study resolutions are the optimization controls.
Static variables are recorded in [`variables.json`](variables.json). The test plans live in each model README.
@@ -14,5 +14,7 @@ This summary is maintained across the model-specific test plans. A result is onl
|---|---|---:|---|---|
| Flux.2 Klein | F-T2I-03 | 220.3 s | [F-T2I-03.png](flux/results/F-T2I-03.png) | Current best; 1080×1920 target |
| Qwen Image 2.1 | Q-T2I-57 | 110.2 s | [Q-T2I-57.png](qwen/results/Q-T2I-57.png) | Current best; direct 1088×1920 latent; res_multistep, beta, 25 steps, CFG 2.5 |
| FLUX.3 Klein | F3I-T2I-01 / sampler matrix | recorded | [best5](flux3image/results/best5/) | Complete local matrix; canonical comparison resolution 1088×1920; visual cross-model ranking pending |
| Ideogram 4 | MAG-MATRIX-24 | 30.154 s | [MAG-MATRIX-24.png](ideogram/results/magazine-matrix/MAG-MATRIX-24.png) | Complete magazine-cover matrix; different structured prompt, not directly comparable to the shared portrait prompt |
Store each T2I test under the selected model's `results/<test-id>.png` and `results/<test-id>.json` using the table ID as the label.