rename the app to offlinepage.app
Browse files
README.md
CHANGED
|
@@ -72,18 +72,18 @@ model-index:
|
|
| 72 |
<p align="center">
|
| 73 |
<a href="https://huggingface.co/spaces/cfahlgren1/enron-autocomplete"><b>Try the demo →</b></a>
|
| 74 |
·
|
| 75 |
-
<a href="https://
|
| 76 |
</p>
|
| 77 |
|
| 78 |
<br>
|
| 79 |
|
| 80 |
-
tinynext is the autocomplete in [
|
| 81 |
|
| 82 |
## v2: the next few words, with context
|
| 83 |
|
| 84 |
<sub>In [`v2/`](https://huggingface.co/paperweight/tinynext/tree/main/v2). The files at the root of this repo are still v1.</sub>
|
| 85 |
|
| 86 |
-
v1 guesses one letter at a time, so its suggestions are short word endings. v2 is a **12.5M-parameter transformer that predicts word pieces**, so one step can be a whole word, and it reads up to about 1,500 characters before the cursor (and, when there is any, text after it). It's the autocomplete in
|
| 87 |
|
| 88 |
**How much typing it saves.** Every model is scored by [`v2/eval.py`](v2/eval.py) on the same held-out text: type it character by character, pause every 12 characters, and count a suggestion only if the text really continues that way. Same decoding rule and 50% cutoff as the app, suggestions of 4+ characters.
|
| 89 |
|
|
@@ -141,7 +141,7 @@ The browser computes in float32 on purpose: onnxruntime-web's int8 kernels were
|
|
| 141 |
|
| 142 |
**Limits.**
|
| 143 |
|
| 144 |
-
-
|
| 145 |
- It still leans toward office email, and it only sees keyboard characters.
|
| 146 |
- It's English only.
|
| 147 |
|
|
@@ -202,10 +202,10 @@ The model knows how sure it is of each guess. You pick a cutoff, and it only sug
|
|
| 202 |
|---|---|---|
|
| 203 |
| 20% | 30% of the time | 48% |
|
| 204 |
| 35% | 18% of the time | 66% |
|
| 205 |
-
| **50%** *(
|
| 206 |
| 70% | 7% of the time | 90% |
|
| 207 |
|
| 208 |
-
The demo uses 40%, and
|
| 209 |
|
| 210 |
These numbers come from 5,000 places in emails from people the model never saw in training. “Right” means the suggestion exactly matched what the person typed next.
|
| 211 |
|
|
@@ -277,5 +277,5 @@ hf jobs uv run --flavor a10g-large --secrets HF_TOKEN train.py --repo your-name/
|
|
| 277 |
<br>
|
| 278 |
|
| 279 |
<p align="center">
|
| 280 |
-
<sub>tinynext is one of two tiny models behind <a href="https://
|
| 281 |
</p>
|
|
|
|
| 72 |
<p align="center">
|
| 73 |
<a href="https://huggingface.co/spaces/cfahlgren1/enron-autocomplete"><b>Try the demo →</b></a>
|
| 74 |
·
|
| 75 |
+
<a href="https://offlinepage.app"><b>Write with it in offlinepage.app →</b></a>
|
| 76 |
</p>
|
| 77 |
|
| 78 |
<br>
|
| 79 |
|
| 80 |
+
tinynext is the autocomplete in [offlinepage.app](https://offlinepage.app). The app uses **v2**, in [`v2/`](https://huggingface.co/paperweight/tinynext/tree/main/v2). The files at the root are v1, the original character model, the same as [`cfahlgren1/enron-autocomplete`](https://huggingface.co/cfahlgren1/enron-autocomplete).
|
| 81 |
|
| 82 |
## v2: the next few words, with context
|
| 83 |
|
| 84 |
<sub>In [`v2/`](https://huggingface.co/paperweight/tinynext/tree/main/v2). The files at the root of this repo are still v1.</sub>
|
| 85 |
|
| 86 |
+
v1 guesses one letter at a time, so its suggestions are short word endings. v2 is a **12.5M-parameter transformer that predicts word pieces**, so one step can be a whole word, and it reads up to about 1,500 characters before the cursor (and, when there is any, text after it). It's the autocomplete in offlinepage.app, running in a Web Worker.
|
| 87 |
|
| 88 |
**How much typing it saves.** Every model is scored by [`v2/eval.py`](v2/eval.py) on the same held-out text: type it character by character, pause every 12 characters, and count a suggestion only if the text really continues that way. Same decoding rule and 50% cutoff as the app, suggestions of 4+ characters.
|
| 89 |
|
|
|
|
| 141 |
|
| 142 |
**Limits.**
|
| 143 |
|
| 144 |
+
- offlinepage.app only suggests at the end of a paragraph, so the text after the cursor is the next paragraph. Mid-line suggestions, where fill-in-the-middle should help most, aren't built yet.
|
| 145 |
- It still leans toward office email, and it only sees keyboard characters.
|
| 146 |
- It's English only.
|
| 147 |
|
|
|
|
| 202 |
|---|---|---|
|
| 203 |
| 20% | 30% of the time | 48% |
|
| 204 |
| 35% | 18% of the time | 66% |
|
| 205 |
+
| **50%** *(offlinepage.app)* | **11% of the time** | **80%** |
|
| 206 |
| 70% | 7% of the time | 90% |
|
| 207 |
|
| 208 |
+
The demo uses 40%, and offlinepage.app uses 50%.
|
| 209 |
|
| 210 |
These numbers come from 5,000 places in emails from people the model never saw in training. “Right” means the suggestion exactly matched what the person typed next.
|
| 211 |
|
|
|
|
| 277 |
<br>
|
| 278 |
|
| 279 |
<p align="center">
|
| 280 |
+
<sub>tinynext is one of two tiny models behind <a href="https://offlinepage.app">offlinepage.app</a>, a writing app that works in airplane mode.<br>Its partner, <a href="https://huggingface.co/paperweight/tinyfix">tinyfix</a>, fixes typos from context.<br>Both are made by <a href="https://huggingface.co/paperweight">paperweight</a>, purpose-built small models.</sub>
|
| 281 |
</p>
|