File size: 1,591 Bytes
e56d379
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
---
language:
- en
license: other
license_name: pleius-internal
tags:
- onnx
- conditional-text-generation
- writing-feedback
- distillation
- creator-tools
---

# cortexa-write-feedback (distilled student)

A ~4.4M-parameter conditional decoder distilled from
`M725/cortexa-write-scorer` (the worker-side TF-IDF/lexical stub).
Takes MiniLM text features (384-d) + the 4 Write pillar scores and
emits a creator-vernacular phrase chain about the draft:

```
"first line hooks | ending sticks"
"tight middle | shareable"
"wall of text | no reason to read"
"drags | no payoff"
```

## Files

| file | purpose |
|---|---|
| `student_int8.onnx` | TinyTransformer decoder, 4 layers / 256-dim / 4 heads, INT8 dynamic-quantized. 6.8 MB. |
| `tokenizer.json`    | Whole-phrase tokenizer (vocab ~120; specials `<pad>`, `<bos>`, `<eos>`, `<sep>`). |
| `config.json`       | Encoder dim (384), pillar names, vocab size, special-token ids. |

## Inference shape

```
inputs:
  encoder_feats   (1, 384)  float32   # sentence-transformers/all-MiniLM-L6-v2 mean-pooled, L2-normalized
  scores          (1, 4)    float32   # [read_likelihood, hold, structure, score] in [0,1]
  scores_present  (1,)      float32   # 1.0 anchored, 0.0 fast-mode
  input_ids       (1, T)    int64
outputs:
  logits          (1, T, V) float32
```

## Training

See `research/distill_students/train_write.py` in the app repo. Teacher
is `score_write_for_rules()` — the Python port of the cortexa-proxy
worker's deterministic TF-IDF write scorer.

## License

Pleius internal — see https://pleius.com. Not for redistribution.