Image-to-Text
English
custom
ocr
document-ai
information-extraction
nameplate
nameplate-ocr
equipment-nameplate
data-center
data-centre
datacenter
mission-critical
mep
electrical
equipment-schedule
schedule-verification
asset-register
commissioning
quality-assurance
ups
pdu
switchgear
generator
ats
transformer
busway
battery
crah
crac
chiller
cooling
construction
aec
edge-ai
on-device
privacy-preserving
tesseract
browser
Upload folder using huggingface_hub
Browse files
README.md
CHANGED
|
@@ -169,6 +169,16 @@ harder: cast shadows, embossed lettering, print over brushed metal, a plate half
|
|
| 169 |
degradation suite, not the clean row, and treat every low-confidence field as one a person must confirm.
|
| 170 |
|
| 171 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 172 |
## Intended use
|
| 173 |
|
| 174 |
- **Commissioning and QA walks:** photograph each unit as it is installed; confirm the delivered unit matches
|
|
|
|
| 169 |
degradation suite, not the clean row, and treat every low-confidence field as one a person must confirm.
|
| 170 |
|
| 171 |
|
| 172 |
+
## Fine-tuning — the higher-accuracy server model
|
| 173 |
+
|
| 174 |
+
The engine above is the on-device path. Alongside it, a **Donut** model
|
| 175 |
+
([`naver-clova-ix/donut-base`](https://huggingface.co/naver-clova-ix/donut-base)) is fine-tuned to read a plate
|
| 176 |
+
**straight into fields** (`<s_nameplate><s_type>ups</s_type><s_manufacturer>…`) for a server / endpoint where
|
| 177 |
+
the browser engine falls short. Training data is synthetic plates with real-photo degradations, plus corrected
|
| 178 |
+
real plates from the field. Once trained, the weights are published in a sibling repository and linked here.
|
| 179 |
+
|
| 180 |
+
Nothing on this page claims the accuracy of that model yet — it is the roadmap, not a result.
|
| 181 |
+
|
| 182 |
## Intended use
|
| 183 |
|
| 184 |
- **Commissioning and QA walks:** photograph each unit as it is installed; confirm the delivered unit matches
|