File size: 2,445 Bytes
c645213
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ac51c98
c645213
 
 
 
 
 
 
 
 
 
 
 
 
 
ac51c98
c645213
 
 
 
 
 
 
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
---
license: cc-by-nc-sa-4.0
base_model: Qwen/Qwen3.5-4B
library_name: peft
pipeline_tag: image-text-to-text
tags:
- medical
- radiology
- ct
- report-generation
- ct-rate
- lora
---

# Label-to-report assembler (stage 2)

Part of the master's thesis *From Scan to Report: Detecting, Segmenting and Reporting Findings in Chest CT* (code: https://github.com/Switchharsh/Switchharsh-From_Scan_to_Report_Master_Thesis; thesis: https://github.com/Switchharsh/Switchharsh-From_Scan_to_Report_Master_Thesis/tree/main/thesis).

Stage 2 of the two-stage pipeline. It writes a report from a list of findings and does not see the scan. Prompt: `Write the findings and impression for this chest CT.` + `Findings present: ... Findings absent: ...` (the 18 CT-RATE labels).

| labels given | CRG | macro-F1 |
|---|---|---|
| reference labels (upper bound) | 0.8099 | 0.8948 |
| FlexiCT attention classifier, t=0.60, raw text | 0.4968 | 0.4990 |
| same, clipped + coverage post-processing | 0.5176 | 0.5298 |
| another patient's classifier labels | 0.3494 | 0.2021 |

## How it fits together
- Input: the 8x8x8 FlexiCT token grid of a chest CT (`flexict/tokens8`, 864-dim per token) from CT-RATE.
- `projector.pt`: MLP (864 -> 2560 -> 2560, LayerNorm) that maps the visual tokens into the LLM embedding space.
- `adapter_model.safetensors` + `adapter_config.json`: LoRA (r=16, alpha=32) on [Qwen/Qwen3.5-4B](https://huggingface.co/Qwen/Qwen3.5-4B).
- Loading and generation code (model wrapper, data, evaluation): https://github.com/Switchharsh/Switchharsh-From_Scan_to_Report_Master_Thesis (`ctbench/rgen/model.py`, `evaluate.py`). The adapter is not a stand-alone chat model: the projector and the FlexiCT tokens are required.
- Task instruction used in every run: `Write the findings and impression for this chest CT.` Greedy decoding, at most 512 new tokens.

## Evaluation protocol
1,000 patient-sampled CT-RATE validation cases; CRG (micro) and macro-F1 with the RadBERT label reader (`suitch/radbert-english-ctrate-classifier`, threshold 0.5). Control = another patient's scan tokens (mismatched). Numbers are from the thesis *From Scan to Report: Detecting, Segmenting and Reporting Findings in Chest CT*.

## Limitations / licence
Research use only, not for clinical use. Trained on CT-RATE (CC BY-NC-SA 4.0), so the weights are released under the same non-commercial terms. Reports can contain false findings; see the thesis for the error analysis.