Document the bundled qevi engine; drop the pip install claim

README.md

@@ -37,7 +37,8 @@ restricted to the allowed answer tokens. If you call `.generate()` and parse the
 reproduce them, because that is not the path the model was fine-tuned on.
 
 The readout is ~30 lines of ordinary transformers code and needs no `trust_remote_code`. A
-self-contained version is below; the [`qevi`](#the-qevi-package) package adds question packing.
+self-contained version is below; the [`qevi` engine bundled in this repo](#the-qevi-engine-bundled-in-this-repo)
+adds question packing.
 
 ---
 
@@ -77,7 +78,7 @@ probs = torch.softmax(logits[cand].float() / 2.45, dim=-1)   # 2.45 = calibratio
 print(dict(zip(answers, probs.tolist())))           # {'Yes': 0.97, 'No': 0.03}
 ```
 
-### With the `qevi` package (many questions, one forward pass)
+### Many questions, one forward pass (using the bundled engine)
 
 ```python
 from qevi import build_readouts, forward_logits, softmax, noul_question, choice_question_generic
@@ -267,14 +268,64 @@ documented above (1 epoch, lr 2e-5, ~5 GPU-hours on two consumer cards).
   has not been validated for moderation use and should not be deployed for it without its own
   evaluation.
 
-## The `qevi` package
+## The `qevi` engine (bundled in this repo)
 
-Packing, prompt construction and the readout live in a small Apache-2.0 package
-(~350 lines, torch only):
+Packing, prompt construction and the readout live in `qevi/` inside **this repository** — three
+files, ~350 lines, torch only, Apache-2.0. Nothing to pip install, and no `trust_remote_code`:
+it is ordinary Python you can read before you run it.
 
-```bash
-pip install qevi
 ```
+qevi/__init__.py   public API
+qevi/core.py       prompt templates, build_readouts(), forward_logits()
+qevi/pack.py       block-diagonal mask + M-RoPE position construction
+```
+
+### Using it
+
+`snapshot_download` fetches the whole repo (weights and engine together), then put it on the path:
+
+```python
+import sys, torch
+from PIL import Image
+from huggingface_hub import snapshot_download
+from transformers import AutoProcessor, Qwen3VLForConditionalGeneration
+
+REPO = "MeerDevelopment/Qevi-2B"
+local = snapshot_download(REPO)      # weights + qevi/ in one download
+sys.path.insert(0, local)            # makes `import qevi` work
+
+from qevi import (build_readouts, forward_logits, softmax,
+                  noul_question, choice_question_generic)
+
+model     = Qwen3VLForConditionalGeneration.from_pretrained(local, dtype=torch.bfloat16).cuda().eval()
+processor = AutoProcessor.from_pretrained(local)
+tok       = processor.tokenizer
+image     = Image.open("photo.jpg").convert("RGB")
+
+specs = [
+    (noul_question("There is a ladder in this image."),  ["Yes", "No"]),
+    (noul_question("There is a bicycle in this image."), ["Yes", "No"]),
+    (choice_question_generic("Which is shown?",
+         ["ladder", "bicycle", "car", "person"]),       ["A", "B", "C", "D"]),
+]
+
+readouts   = build_readouts(tok, specs)
+logit_sets = forward_logits(model, processor, image, readouts)   # ONE forward pass
+
+for (question, answers), lg in zip(specs, logit_sets):
+    probs = softmax(lg, T=2.45)                                  # see temperatures above
+    print(dict(zip(answers, [round(p, 4) for p in probs])))
+```
+
+All three questions share a single encoding of the image and are isolated from one another by the
+block-diagonal mask, so the answers are identical to asking them separately — verified bitwise,
+see [Order invariance](#order-invariance). Asking 30 questions instead of 3 costs about 5 ms more
+per question rather than another full forward pass each.
+
+### If you would rather vendor it
+
+The three files have no dependency on this repo's layout. Copy `qevi/` into your own project and
+`import qevi` directly; only `torch`, `transformers` and `pillow` are required.
 
 ## Citation