Ethanfel
0e9e99b8b2
Qwen3.5/3.6 are reasoning models — they 'think out loud' in markdown and never
reach the JSON, then get cut off at the token limit -> '(no parseable judgement)'.
Fixes: apply_chat_template(enable_thinking=False) + strip <think> blocks; hardened
'output ONLY JSON, do not think out loud' instruction; default max_new_tokens
1024->2048 (max 8192); and a markdown fallback parser (_parse_markdown_verdicts /
_parse_axes) that extracts per-axis {verdict,ref,gen} from the prose the model
reliably emits. describe falls back to using the raw text as the caption.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
146 lines
8.8 KiB
Markdown
146 lines
8.8 KiB
Markdown
# ComfyUI-Prompt-Calibratror
|
|
|
|
A **fully local** prompt calibration loop for ComfyUI. A vision-language model
|
|
(Qwen3-VL) judges how close a *generated* image is to a *reference* image and
|
|
returns a structured score + per-axis difference analysis, which is used to
|
|
**calibrate the prompt-generation method** ([ComfyUI-Prompt-Builder](../ComfyUI-Prompt-Builder))
|
|
until the generated image matches the reference.
|
|
|
|
> Full design rationale, controller options, and VLM-as-judge variance mitigations
|
|
> are in **[docs/METHODOLOGY.md](docs/METHODOLOGY.md)**. The controller is an **external
|
|
> CLI agent** that drives ComfyUI via its HTTP API — see **[docs/AGENT_LOOP.md](docs/AGENT_LOOP.md)**.
|
|
|
|
## Nodes & tools
|
|
|
|
| Component | What it is |
|
|
|---|---|
|
|
| `Qwen3-VL Image Judge (Calibrator)` | scores generated vs reference, writes analysis to disk for the agent |
|
|
| `SxCP External Prompt (Receptor)` | stable injection point; the agent sets `prompt/negative/seed` here per queue |
|
|
| `agent_bridge.py` | one CLI call = one iteration (inject → `POST /prompt` → wait → print analysis JSON) |
|
|
|
|
## The "vllm node": `Qwen3-VL Image Judge (Calibrator)`
|
|
|
|
The core node (`nodes/qwen_judge.py`). It reuses the standard transformers Qwen3-VL
|
|
inference plumbing (same approach as
|
|
[ComfyUI-QwenVL-MultiImage](https://github.com/hardik-uppal/ComfyUI-QwenVL-MultiImage)
|
|
— the recommended reuse base) but **forces strict JSON output** so an automated loop
|
|
can act on it.
|
|
|
|
**Inputs**
|
|
|
|
| name | type | default | notes |
|
|
|---|---|---|---|
|
|
| `reference_image` | IMAGE | — | the target |
|
|
| `mode` | compare / describe / chat | compare | `compare` = score ref vs generated. `describe` = first pass over the reference → caption + target spec. `chat` = **general VLM**: your `system_prompt` + `user_prompt` over the image(s) → raw text |
|
|
| `profile` | general / oral / penetration / handjob / solo | general | **analysis profile** — act-specialized axis set; the act-critical axes are distance/proximity-aware (e.g. `mouth_genital_distance`) so magnitude isn't hidden behind a coarse label |
|
|
| `generated_image` | IMAGE (optional) | — | the candidate to score (required for `compare`, ignored for `describe`) |
|
|
| `model_select` | dropdown (model name) | 4B local | **which judge** (transformers/safetensors, auto-downloaded): Qwen3-VL 4B/8B/30B-A3B, **Qwen3.5-9B**, **Qwen3.6-27B/35B-A3B** (newer, natively multimodal). Param size shown in the label |
|
|
| `precision` | bf16 / fp8 / nf4 | bf16 | **the quant** — applies to the selected model (VRAM table below) |
|
|
| `model_path` | STRING | "" (empty) | **manual override** of the dropdown — local dir, HF repo id, or alias (`8b`/`30b-a3b`/`3.5-9b`/`3.6-27b`/`3.6-35b`). Empty = use `model_select` |
|
|
| `axes` | STRING | "" (empty) | **override** the profile's axis set with a custom comma/newline list; empty = use `profile` |
|
|
| `max_new_tokens` | INT | 2048 | raise it if a reasoning model (Qwen3.5/3.6) gets cut off before finishing |
|
|
| `temperature` | FLOAT | 0.0 | 0 = greedy/repeatable |
|
|
| `swap_eval` | BOOL | true | run twice with images swapped, average → cuts position bias |
|
|
| `keep_loaded` | BOOL | true | cache weights across loop iterations |
|
|
| `auto_download` | BOOL | true | if `model_path` is a repo id/alias and not local, fetch it from HF into `models/prompt_generator/` |
|
|
| `system_prompt` | STRING | "" | **chat mode**: your system prompt |
|
|
| `user_prompt` | STRING | "Describe this image." | **chat mode**: your instruction over the image(s) |
|
|
|
|
**Auto-download:** set `model_path` to `30b-a3b` (alias) or any `org/name` repo id and leave
|
|
`auto_download` on — the node snapshot-downloads it on first run (into ComfyUI's
|
|
`models/prompt_generator/<name>`) and reuses the local copy afterward. Local paths and the
|
|
default skip download entirely.
|
|
|
|
**General VLM (chat mode):** set `mode=chat` and the node becomes a plain vision-language
|
|
node — feed an image (and optionally a second), write your own `system_prompt`/`user_prompt`,
|
|
and read the model's text from the `analysis` output. Reuses the same model dropdown, quant,
|
|
and auto-download as the judge, so it's a one-node abliterated VLM for captioning, tagging,
|
|
Q&A, prompt-from-image, etc. (CLI: `agent_bridge.py --mode chat --user-prompt "..."`).
|
|
|
|
**Outputs**
|
|
|
|
| name | type | use |
|
|
|---|---|---|
|
|
| `overall_score` | FLOAT 0..1 | compare: mean verdict (computed here, not by the model). describe: `1.0` placeholder |
|
|
| `axis_scores_json` | STRING (JSON) | compare: per-axis `{verdict, ref, gen}` (verdict = match/partial/mismatch). describe: `{axis: value}` |
|
|
| `analysis` | STRING | compare: header (`overall, N mismatches`) + axes worst-first (`VERDICT ref:[…] gen:[…]`). describe: the `caption`. chat: the model's response |
|
|
| `raw` | STRING | raw model output (both passes if `swap_eval`) |
|
|
| `report_path` | STRING | path to the written `calib_<tag>.json` (carries `mismatch_count`) |
|
|
|
|
## Install
|
|
|
|
```bash
|
|
cd /media/p5/Comfyui/custom_nodes
|
|
ln -s /media/p5/ComfyUI-Prompt-Calibratror . # or git clone
|
|
/media/p5/Comfyui/venv/bin/pip install -r /media/p5/ComfyUI-Prompt-Calibratror/requirements.txt
|
|
```
|
|
|
|
The node defaults to the **huihui-ai Qwen3-VL-4B-Instruct abliterated** weights already
|
|
converted at `/media/p5/qwen3vl_4b_abliterated_comfy_convert/` so it runs out of the box
|
|
(the abliterated/uncensored variant won't refuse to analyze adult imagery, which would
|
|
otherwise break the loop).
|
|
|
|
**Pick a model in `model_select` and a quant in `precision`.** All are abliterated,
|
|
multimodal **safetensors** (transformers), auto-downloaded. The newer **Qwen3.5/3.6** are
|
|
natively multimodal (need a recent transformers — they load via `AutoModelForMultimodalLM`).
|
|
|
|
VRAM by quant on the RTX 5090 32 GB (✅ fits / ⚠ tight / ❌):
|
|
|
|
| model | bf16 | fp8 | nf4 | note |
|
|
|---|---|---|---|---|
|
|
| Qwen3-VL-4B (local) | ✅ ~9 | ✅ ~5 | ✅ ~3 | fast, weak |
|
|
| Qwen3-VL-8B | ✅ ~17 | ✅ ~9 | ✅ ~6 | solid, fast |
|
|
| **Qwen3.5-9B** | ✅ ~20 | ✅ ~10 | ✅ ~7 | **newer, fast — recommended** |
|
|
| Qwen3-VL-30B-A3B (MoE) | ❌ ~62 | ⚠ ~31 | ✅ ~18 | nf4 slow |
|
|
| Qwen3.6-27B (dense) | ❌ ~56 | ⚠ ~28 | ✅ ~16 | nf4 slow, strong |
|
|
| Qwen3.6-35B-A3B (MoE) | ❌ ~70 | ❌ | ✅ ~20 | nf4 slow, top quality |
|
|
|
|
`nf4` (bitsandbytes) fits the big ones but is **slow** (dequant overhead) — that's the
|
|
bottleneck, not the model. `fp8` is fast but only when a real fp8 checkpoint exists (the
|
|
local 4B has one; `precision=fp8` on a bf16-only repo won't quantize). For speed + recency,
|
|
**Qwen3.5-9B at bf16** is the sweet spot. See
|
|
[docs/METHODOLOGY.md](docs/METHODOLOGY.md#model-sizing-on-32-gb-rtx-5090--abliterated-latest-qwen-vl).
|
|
|
|
## Loop sketch
|
|
|
|
```
|
|
Prompt-Builder (SxCP) ──prompt──▶ T2I (SDXL/Flux/Krea2) ──image──▶ Qwen3-VL Image Judge
|
|
▲ │
|
|
└──────── knob overrides ◀── Controller ◀── overall_score + diff ┘
|
|
```
|
|
|
|
Use the Prompt-Builder **For-Loop Start/End + Accumulator** nodes to drive iterations and
|
|
route `overall_score` into the stop condition. Controller options (greedy hill-climb →
|
|
black-box optimizer → LLM-in-the-loop) are in the methodology doc.
|
|
|
|
## End-to-end loop
|
|
|
|
1. Run ComfyUI with `--listen`, install this node pack, put your reference at `ComfyUI/input/reference.png`.
|
|
2. **First pass (describe):** the judge looks at the reference alone and emits **one canonical
|
|
scene description** (coherent paragraph + per-axis target spec) to seed the prompt *and*
|
|
anchor the loop:
|
|
```bash
|
|
python agent_bridge.py --mode describe --workflow workflow/workflow_describe_api.json \
|
|
--run-tag seed --analysis-dir /media/p5/Comfyui/output/calibrator
|
|
```
|
|
3. **Compare loop:** load `workflow/workflow_api.json` (SDXL `waiIllustriousSDXL_v160` example —
|
|
swap the checkpoint for Flux/Krea as needed) and iterate, following `docs/CALIBRATION_POLICY.md`.
|
|
Pass `--ref-desc-file` so compare anchors on the canonical reference (the `ref` side stays
|
|
fixed; only the generated image is re-read each turn):
|
|
```bash
|
|
python agent_bridge.py --workflow workflow/workflow_api.json \
|
|
--prompt "<description from step 2, then calibrated>" \
|
|
--ref-desc-file /media/p5/Comfyui/output/calibrator/calib_seed.json \
|
|
--run-tag iter001 --analysis-dir /media/p5/Comfyui/output/calibrator
|
|
```
|
|
stdout = the analysis JSON (`{verdict, ref, gen}` per axis) → agent steers toward `ref` → next iteration.
|
|
|
|
## Status
|
|
|
|
- [x] Methodology + node selection (`docs/METHODOLOGY.md`)
|
|
- [x] Qwen3-VL Image Judge node — `describe` (first pass) + `compare` (scoring), swap-eval, file report
|
|
- [x] Agent-driven architecture (`docs/AGENT_LOOP.md`) — Receptor node + `agent_bridge.py` (`--mode`)
|
|
- [x] Example workflows: `workflow_describe_api.json` (first pass) + `workflow_api.json` (compare loop)
|
|
- [x] Agent calibration policy (`docs/CALIBRATION_POLICY.md`)
|
|
- [ ] Optional: structured-config receptor (carry Prompt-Builder knobs instead of a flat string)
|