@algorisys/tinyfly 0.118.0 → 0.120.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +6 -6
- package/docs/editor-guide.md +1 -1
- package/docs/examples.md +1 -1
- package/docs/getting-started.md +1 -1
- package/docs/gsap-compat.md +1 -1
- package/docs/llm-guide.md +5 -1
- package/docs/maps.md +2 -2
- package/docs/math.md +89 -7
- package/docs/teaching.md +1 -1
- package/lib/addons/characters.js +1852 -1844
- package/lib/addons/math.js +2229 -1533
- package/lib/browser/tinyfly-math.iife.js +4 -4
- package/lib/browser/tinyfly.iife.js +11 -11
- package/lib/browser/tinyfly.js +438 -430
- package/lib/browser/tinyfly.umd.js +11 -11
- package/lib/types/math/figure/display-list.d.ts +2 -0
- package/lib/types/math/figure/figure-spec.d.ts +101 -1
- package/lib/types/math/figure/items/ml.d.ts +47 -0
- package/lib/types/math/index.d.ts +8 -0
- package/lib/types/math/ml/classify.d.ts +30 -0
- package/lib/types/math/ml/colormap.d.ts +14 -0
- package/lib/types/math/ml/contours.d.ts +15 -0
- package/lib/types/math/ml/descent.d.ts +11 -0
- package/lib/types/math/ml/kmeans.d.ts +11 -0
- package/lib/types/math/ml/network.d.ts +19 -0
- package/lib/types/math/presets/ml.d.ts +50 -0
- package/lib/types/math/surface/math-surface.d.ts +69 -0
- package/llms-full.txt +110 -16
- package/llms.txt +31 -31
- package/package.json +1 -1
- package/skills/tinyfly/SKILL.md +4 -1
package/README.md
CHANGED
|
@@ -47,7 +47,7 @@ A lightweight, API-driven animation engine and visual editor for creating high-p
|
|
|
47
47
|
- **Full screen** - `data-fullscreen="true"` adds a button that fills the screen with the figure, through the Fullscreen API or, on iPhone, an overlay; the drawing scales to the screen whatever the page's figure CSS says
|
|
48
48
|
- **Scenarios** - Several timelines on one figure, and the reader chooses which plays: a group of options, a stepped slider, or clickable parts of the SVG (`data-tinyfly-choose`). Switching undoes the previous scenario's drawing and keeps the reader at the same step
|
|
49
49
|
- **Authoring kit** - `lesson()` step builder and diagram primitives (array cells, pointer, stack, queue / channel, table, pipeline) in `@algorisys/tinyfly/teach`
|
|
50
|
-
- **Maths animations** - `@algorisys/tinyfly/math`: figures as JSON (a grid a matrix warps, î and ĵ, vectors, function plots that break at asymptotes, polygons whose area shows the determinant, eigenlines, TeX labels; calculus: tangents and secants with exact slopes, Riemann sums, signed areas, Taylor polynomials from symbolic derivatives, limits; probability: seeded histograms, distributions with shaded probabilities, a Galton board, the central limit theorem, Monte Carlo π, regression and Bayes as areas; discrete: graphs searched step by step (BFS, DFS, Dijkstra) with their queue or stack, recursion trees, Venn diagrams of any set expression, clock arithmetic), a built-in TeX subset (fractions, roots, sums and integrals with limits, matrices, Greek, live values) whose equations rewrite step by step with terms gliding, and ready linear algebra, calculus, probability
|
|
50
|
+
- **Maths animations** - `@algorisys/tinyfly/math`: figures as JSON (a grid a matrix warps, î and ĵ, vectors, function plots that break at asymptotes, polygons whose area shows the determinant, eigenlines, TeX labels; calculus: tangents and secants with exact slopes, Riemann sums, signed areas, Taylor polynomials from symbolic derivatives, limits; probability: seeded histograms, distributions with shaded probabilities, a Galton board, the central limit theorem, Monte Carlo π, regression and Bayes as areas; discrete: graphs searched step by step (BFS, DFS, Dijkstra) with their queue or stack, recursion trees, Venn diagrams of any set expression, clock arithmetic; AI/ML: neural networks with activations flowing, gradient descent on contoured loss surfaces, attention heatmaps, logistic regression finding its decision boundary, k-means), a built-in TeX subset (fractions, roots, sums and integrals with limits, matrices, Greek, live values) whose equations rewrite step by step with terms gliding, and ready linear algebra, calculus, probability, discrete maths and AI/ML figures. A figure is also a surface (`mathSurface`): characters point at `tip:basis.j` or `term:eq:2x` and change the figure as they speak, with `surfaceScript`. One layout feeds lesson-page embeds (`<figure data-tinyfly-math>`, stepped with captions and questions), explorables where readers drag î and ĵ, and 9:16 or 16:9 videos ([guide](docs/math.md))
|
|
51
51
|
- **Build tools** - `npx @algorisys/tinyfly validate` catches broken figures in CI; `render` writes a frame to static SVG for RSS, email and print
|
|
52
52
|
|
|
53
53
|
### Video from Code
|
|
@@ -284,7 +284,7 @@ GSAP-shaped functions at the top level. Teaching embeds are included: add
|
|
|
284
284
|
itself.
|
|
285
285
|
|
|
286
286
|
```html
|
|
287
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
287
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly.iife.js"></script>
|
|
288
288
|
<script>
|
|
289
289
|
tinyfly.to('.box', { x: 200, rotate: 90, duration: 1, ease: 'power2.out' })
|
|
290
290
|
|
|
@@ -321,7 +321,7 @@ A teaching figure needs no code at all:
|
|
|
321
321
|
<script type="application/json" data-tinyfly-timeline>{ …timeline JSON with markers… }</script>
|
|
322
322
|
<figcaption>Appending to a full slice</figcaption>
|
|
323
323
|
</figure>
|
|
324
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
324
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly.iife.js" data-tinyfly-auto></script>
|
|
325
325
|
```
|
|
326
326
|
|
|
327
327
|
See [Teaching Animations](docs/teaching.md).
|
|
@@ -420,7 +420,7 @@ Without a build step, the player bundle puts the same functions on a `tinyfly` g
|
|
|
420
420
|
<div data-tinyfly="box" style="width: 60px; height: 60px; background: #4a9eff;"></div>
|
|
421
421
|
</div>
|
|
422
422
|
|
|
423
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
423
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly-player.iife.js"></script>
|
|
424
424
|
<script>
|
|
425
425
|
tinyfly.play('#animation', './animation.json', { loop: -1 })
|
|
426
426
|
</script>
|
|
@@ -529,7 +529,7 @@ Or skip the code entirely with declarative embeds (see [Teaching Animations](doc
|
|
|
529
529
|
<svg viewBox="0 0 720 200">…</svg>
|
|
530
530
|
<script type="application/json" data-tinyfly-timeline>{ …timeline JSON… }</script>
|
|
531
531
|
</figure>
|
|
532
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
532
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly-embed.iife.js" data-tinyfly-auto></script>
|
|
533
533
|
```
|
|
534
534
|
|
|
535
535
|
### Audio / Video Sync
|
|
@@ -839,7 +839,7 @@ Characters browser suite checks 74 desktop/mobile scenarios per browser.
|
|
|
839
839
|
| Docs viewer and `llms.txt` (`src/docs`) | 18 |
|
|
840
840
|
| Embeds, teaching kit, framework hooks, landing (`src/embed`, `src/teach`, `src/frameworks`, `src/landing`) | 33 |
|
|
841
841
|
| Headless rendering and audio (`src/headless`) | 63 |
|
|
842
|
-
| Maths figures, TeX, calculus, probability, discrete maths, equation steps, embeds and explorables (`src/math`) |
|
|
842
|
+
| Maths figures, TeX, calculus, probability, discrete maths, AI/ML, presenter surfaces, equation steps, embeds and explorables (`src/math`) | 125 |
|
|
843
843
|
| Entry boundaries | 3 |
|
|
844
844
|
|
|
845
845
|
## Contributing
|
package/docs/editor-guide.md
CHANGED
|
@@ -1143,7 +1143,7 @@ The dialog shows the generated HTML/JavaScript code with a **Copy Code** button.
|
|
|
1143
1143
|
**Steps to embed:**
|
|
1144
1144
|
1. Build the player: `npm run build:player`
|
|
1145
1145
|
2. Copy `lib/player/tinyfly-player.iife.js` to your project (or load
|
|
1146
|
-
`https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
1146
|
+
`https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly-player.iife.js`
|
|
1147
1147
|
instead; see [Deployment](DEPLOYMENT.md))
|
|
1148
1148
|
3. Copy the generated code into your HTML
|
|
1149
1149
|
4. Adjust the script `src` path if needed
|
package/docs/examples.md
CHANGED
|
@@ -743,7 +743,7 @@ For simple website embedding without a build system.
|
|
|
743
743
|
</html>
|
|
744
744
|
```
|
|
745
745
|
|
|
746
|
-
Build the player with `npm run build:player` to generate `lib/player/tinyfly-player.iife.js`, or load it from the CDN: `https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
746
|
+
Build the player with `npm run build:player` to generate `lib/player/tinyfly-player.iife.js`, or load it from the CDN: `https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly-player.iife.js`.
|
|
747
747
|
|
|
748
748
|
---
|
|
749
749
|
|
package/docs/getting-started.md
CHANGED
|
@@ -185,7 +185,7 @@ with a GSAP-style call (durations are in seconds here):
|
|
|
185
185
|
```html
|
|
186
186
|
<div class="box" style="width:60px;height:60px;background:#4a9eff"></div>
|
|
187
187
|
|
|
188
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
188
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly.iife.js"></script>
|
|
189
189
|
<script>
|
|
190
190
|
tinyfly.to('.box', { x: 200, duration: 1 })
|
|
191
191
|
</script>
|
package/docs/gsap-compat.md
CHANGED
|
@@ -74,7 +74,7 @@ Without a build step, the all-in-one bundle exposes the same functions on a
|
|
|
74
74
|
global — `tinyfly.to()`, `tinyfly.timeline()` and so on:
|
|
75
75
|
|
|
76
76
|
```html
|
|
77
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
77
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly.iife.js"></script>
|
|
78
78
|
<script>
|
|
79
79
|
tinyfly.to('.box', { x: 200, duration: 1 })
|
|
80
80
|
</script>
|
package/docs/llm-guide.md
CHANGED
|
@@ -276,7 +276,11 @@ lesson page (`<figure data-tinyfly-math>`), an explorable and a video
|
|
|
276
276
|
(`histogram`, `distribution`, `galton`, `scatter`, `bayes`) take a distribution as data
|
|
277
277
|
(`{ kind: 'normal', mean: 'mu', sd: 1 }`) and draw seeded samples, never `Math.random`. Discrete:
|
|
278
278
|
`graph` (nodes with `at`, or `tree: { root }`; `search` stepped by `step`), `venn` (`highlight: "(A | B)'"`),
|
|
279
|
-
`clock` (`modulus`, `value`).
|
|
279
|
+
`clock` (`modulus`, `value`). AI/ML: `network` (`flow`), `descent` (`steps`, exact gradient),
|
|
280
|
+
`contours`, `heatmap` (`softmax: true` for attention), `classifier` (`epochs`), `kmeans`
|
|
281
|
+
(`iteration`); all seeded. For a presenter, `mathSurface(figure, { x, y })` gives the
|
|
282
|
+
figure anchors (`tip:basis.i`, `item:ID`, `term:ID:TEXT`, `node:GRAPH:NODE`) and edits
|
|
283
|
+
(`set` with `value`, `step` with `to`, `drawOn`) for `surfaceScript` beats. See `docs/math.md`.
|
|
280
284
|
|
|
281
285
|
## Your own behaviours
|
|
282
286
|
|
package/docs/maps.md
CHANGED
|
@@ -14,8 +14,8 @@ On a page with no build step, load the maps add-on after the browser bundle; it
|
|
|
14
14
|
adds to the same `tinyfly` global:
|
|
15
15
|
|
|
16
16
|
```html
|
|
17
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
18
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
17
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly.iife.js"></script>
|
|
18
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly-maps.iife.js"></script>
|
|
19
19
|
```
|
|
20
20
|
|
|
21
21
|
It is a separate script because the offline world map it carries adds about
|
package/docs/math.md
CHANGED
|
@@ -86,6 +86,12 @@ Some names are reserved, because a track on them would move the whole figure:
|
|
|
86
86
|
| `graph` | `nodes: [{ id, at?, label? }]`, `edges` (`[from, to, weight?]` or objects), `directed?`, `tree?: { root, spacing?, at? }`, `search?: { kind: 'bfs' \| 'dfs' \| 'dijkstra', start }`, `step?`, `frontier?` (screen px), `weights?`, `radius?` |
|
|
87
87
|
| `venn` | `sets` (two or three `{ id, label?, color? }`), `highlight?` (a set expression), `highlightColor?`, `at?`, `radius?`, `universe?` |
|
|
88
88
|
| `clock` | `modulus`, `value`, `at?`, `radius?` |
|
|
89
|
+
| `network` | `layers` (`[3, 5, 4, 2]`), `activation?: 'sigmoid' \| 'tanh' \| 'relu'`, `inputs?`, `flow?`, `seed?`, `at?`, `size?`, `labels?: { inputs?, outputs? }`, `radius?` |
|
|
90
|
+
| `heatmap` | `cells` (rows, top first; `softmax?` per row) or `expr` in x and y with `region?`, `resolution?`; `range?`, `colormap?: 'viridis' \| 'heat' \| 'diverging'`, `rowLabels?`, `colLabels?`, `values?`, `at?`, `size?` |
|
|
91
|
+
| `contours` | `expr` in x and y, `levels?` (a count or the levels), `region?`, `resolution?`, `colormap?` |
|
|
92
|
+
| `descent` | `expr` in x and y, `start`, `rate`, `steps`, `momentum?` |
|
|
93
|
+
| `classifier` | `epochs`, `centres?`, `count?`, `spread?`, `rate?`, `start?: [w₁, w₂, b]`, `probability?`, `seed?` |
|
|
94
|
+
| `kmeans` | `centres` (the true clusters), `iteration`, `k?`, `count?`, `spread?`, `seed?` |
|
|
89
95
|
|
|
90
96
|
`point` also rides a curve: `{ on: 'sin(x)', x: 'p' }` is the point (p, sin p), and
|
|
91
97
|
`hollow: true` draws an open circle for a hole or a limit. `function` takes
|
|
@@ -113,6 +119,31 @@ Items work out numbers that TeX can show with `\val{…}`:
|
|
|
113
119
|
| `bayes` | `ID.prior`, `ID.posterior`, `ID.positive`, `ID.truePositive`, `ID.falsePositive` |
|
|
114
120
|
| `graph` with a search | `ID.visited`, `ID.steps`, and for Dijkstra `ID.d.NODE` (∞ until known) |
|
|
115
121
|
| `clock` | `ID.hour` (value mod modulus), `ID.turns` |
|
|
122
|
+
| `network` | `ID.out0`, `ID.out1` … (the outputs) |
|
|
123
|
+
| `descent` | `ID.x`, `ID.y`, `ID.loss`, `ID.step` |
|
|
124
|
+
| `classifier` | `ID.accuracy`, `ID.loss`, `ID.epoch` |
|
|
125
|
+
| `kmeans` | `ID.inertia`, `ID.iteration`, `ID.iterations` |
|
|
126
|
+
|
|
127
|
+
### Machine learning
|
|
128
|
+
|
|
129
|
+
Everything is seeded and worked out exactly, so a figure trains the same way
|
|
130
|
+
every time:
|
|
131
|
+
|
|
132
|
+
- **Networks** draw seeded weights: blue edges add, red ones subtract, and
|
|
133
|
+
thicker edges count more. `flow` moves activations layer by layer, with
|
|
134
|
+
pulses on the edges being fed.
|
|
135
|
+
- **Descent** takes exact partial derivatives of its expression, so the path
|
|
136
|
+
follows the true gradient. A step size that diverges stops cleanly.
|
|
137
|
+
- **Classifiers** train logistic regression by full-batch gradient descent,
|
|
138
|
+
starting from a deliberately wrong line that visibly turns into place.
|
|
139
|
+
- **k-means** runs Lloyd's algorithm from the points nearest the middle of the
|
|
140
|
+
data, so the centroids travel; a fractional `iteration` glides them.
|
|
141
|
+
- **Heatmaps and contours** colour through a colormap. Dimmed cells are
|
|
142
|
+
pre-mixed with the background, so neighbours never leave seams. Contours are
|
|
143
|
+
traced by marching squares and joined into whole curves.
|
|
144
|
+
|
|
145
|
+
The pieces are also available on their own: `seededNetwork`, `forward`,
|
|
146
|
+
`descentPath`, `trainLogistic`, `kmeansHistory`, `contourLines`, `colorAt`.
|
|
116
147
|
|
|
117
148
|
### Graphs, searches and sets
|
|
118
149
|
|
|
@@ -207,7 +238,7 @@ lesson's step options (`duration`, `easing`, `delay`).
|
|
|
207
238
|
<figcaption>The determinant is the area of the unit square after the map.</figcaption>
|
|
208
239
|
</figure>
|
|
209
240
|
|
|
210
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
241
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly-math.iife.js" data-tinyfly-auto></script>
|
|
211
242
|
```
|
|
212
243
|
|
|
213
244
|
The figure draws itself into an SVG and redraws from the timeline whenever the
|
|
@@ -265,6 +296,50 @@ figure's smaller side, so a figure reads the same as a 600 px card and as a
|
|
|
265
296
|
1080×1920 short. `mathTarget()` is a canvas custom target: register it on any
|
|
266
297
|
`CanvasAdapter` too, and `describeTarget` / `checkTracks` know its props.
|
|
267
298
|
|
|
299
|
+
## Presenting with characters
|
|
300
|
+
|
|
301
|
+
`mathSurface(figure, { x, y })` makes a figure a **surface**, the way a code
|
|
302
|
+
panel, a whiteboard or a chart is one. Its places have names, so a character's
|
|
303
|
+
beats can aim at them as plain data. Its edits change the figure when a beat
|
|
304
|
+
says so, recorded on the figure's own props:
|
|
305
|
+
|
|
306
|
+
```js
|
|
307
|
+
import { linearMapFigure, mathSurface } from '@algorisys/tinyfly/math'
|
|
308
|
+
import { surfaceScript } from '@algorisys/tinyfly/characters'
|
|
309
|
+
|
|
310
|
+
const fig = mathSurface(linearMapFigure({ width: 400, height: 400 }), { x: 20, y: 20 })
|
|
311
|
+
const on = (anchor) => ({ surface: 'fig', anchor })
|
|
312
|
+
const result = surfaceScript('teacher', { fig }, [
|
|
313
|
+
{ do: 'walk', to: 560 },
|
|
314
|
+
{ do: 'point', target: on('tip:basis.j'), say: 'Now slide it over…',
|
|
315
|
+
then: { surface: 'fig', edit: 'set', anchor: 'value:M', value: [1, 1, 0, 1] } },
|
|
316
|
+
{ do: 'point', target: on('item:square'), say: 'A shear: the area stays 1.' },
|
|
317
|
+
], { from: 740, ground: 420, height: 190 })
|
|
318
|
+
// result.tracks: the figure's and the teacher's tracks; draw fig.target with its props
|
|
319
|
+
```
|
|
320
|
+
|
|
321
|
+
| Anchor | Names |
|
|
322
|
+
|---|---|
|
|
323
|
+
| `box` | the whole figure |
|
|
324
|
+
| `item:ID` | everything an item draws |
|
|
325
|
+
| `tip:ID` | the tip of a vector (`tip:basis.i`, `tip:basis.j` for î and ĵ) |
|
|
326
|
+
| `point:ID` | a point, including one riding a curve |
|
|
327
|
+
| `node:GRAPH:NODE` | a node of a graph |
|
|
328
|
+
| `term:ID:TEXT` | glyphs of an equation or label by their text (`term:eq:2x`; `term:eq#2:x` for the second) |
|
|
329
|
+
| `value:NAME` | a value, for `set` (its box is the whole figure) |
|
|
330
|
+
|
|
331
|
+
| Edit | Does |
|
|
332
|
+
|---|---|
|
|
333
|
+
| `drawOn`, `erase` | draws an item on, or off (`item:ID`) |
|
|
334
|
+
| `show`, `hide` | fades an item in or out |
|
|
335
|
+
| `step` | moves an equation to step `to`, terms gliding |
|
|
336
|
+
| `set` | tweens a value to `value`: a number, `[x, y]` or `[a, b, c, d]` |
|
|
337
|
+
|
|
338
|
+
Anchors are worked out from the figure as it is at that moment, so after a
|
|
339
|
+
`set` warps the plane, `tip:basis.i` is where î has gone. Parts of a figure do
|
|
340
|
+
not come loose: `piece`, `fling`, `follow` and `move` explain that edits are
|
|
341
|
+
the way to change it.
|
|
342
|
+
|
|
268
343
|
## Equations
|
|
269
344
|
|
|
270
345
|
Equations are typeset by a built-in TeX subset, with no KaTeX or MathJax:
|
|
@@ -318,6 +393,12 @@ their own.
|
|
|
318
393
|
| `recursionTreeFigure({ n })` | The call tree of naive recursive Fibonacci, in call order (`step`) |
|
|
319
394
|
| `vennFigure({ sets, highlight, label })` | Two or three sets with a region shaded |
|
|
320
395
|
| `clockFigure({ modulus, value })` | Clock arithmetic (`value`) |
|
|
396
|
+
| `neuralNetworkFigure({ layers, activation, inputs })` | A forward pass (`flow`, inputs `in0` …) |
|
|
397
|
+
| `gradientDescentFigure({ expr, start, rate, momentum })` | Descent on a narrow valley over its contours (`steps`, `rate`, `momentum`) |
|
|
398
|
+
| `attentionFigure({ tokens, scores })` | Attention weights, softmaxed per row, with token labels |
|
|
399
|
+
| `decisionBoundaryFigure({ rate, spread })` | Logistic regression finding the line between two clusters (`epochs`) |
|
|
400
|
+
| `kmeansFigure({ centres, k })` | k-means finding clusters (`iteration`) |
|
|
401
|
+
| `activationFigure()` | Sigmoid, tanh and ReLU |
|
|
321
402
|
|
|
322
403
|
Each returns a plain figure to change or extend. In each, `M` starts at the
|
|
323
404
|
identity unless `matrix` is given.
|
|
@@ -327,13 +408,14 @@ identity unless `matrix` is given.
|
|
|
327
408
|
- Gallery: **Linear Map**, **Eigenvectors**, **Equation Steps**, **Matrix Lesson**,
|
|
328
409
|
**Derivative**, **Riemann Sums**, **Taylor Series**, **Galton Board**, **Central Limit
|
|
329
410
|
Theorem**, **Monte Carlo π**, **Bayes: a Positive Test**, **Graph Search**, **Recursion
|
|
330
|
-
Tree**, **Set Operations**
|
|
331
|
-
|
|
332
|
-
|
|
333
|
-
`
|
|
411
|
+
Tree**, **Set Operations**, **Clock Arithmetic**, **Neural Network**, **Gradient Descent**,
|
|
412
|
+
**Decision Boundary**, **K-Means**, **Attention Weights** and **Maths Presenter**.
|
|
413
|
+
- Videos: `linear-map-short.mjs`, `riemann-short.mjs`, `galton-short.mjs`,
|
|
414
|
+
`graph-search-short.mjs` and `gradient-descent-short.mjs` (9:16); `eigen-explainer.mjs`,
|
|
415
|
+
`taylor-explainer.mjs`, `clt-explainer.mjs`, `de-morgan-explainer.mjs` and
|
|
416
|
+
`neural-network-explainer.mjs` (16:9), in `examples/headless-video/`.
|
|
334
417
|
|
|
335
418
|
## Next
|
|
336
419
|
|
|
337
|
-
|
|
338
|
-
decision boundaries) builds on the same figure model.
|
|
420
|
+
A 3D view of a loss surface (with the scene-3d renderer) is planned. See `todo.md`.
|
|
339
421
|
See `todo.md` for their status.
|
package/docs/teaching.md
CHANGED
|
@@ -25,7 +25,7 @@ path for all of that:
|
|
|
25
25
|
</figure>
|
|
26
26
|
|
|
27
27
|
<!-- once, anywhere on the page (a site-wide footer is fine) -->
|
|
28
|
-
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.
|
|
28
|
+
<script src="https://cdn.jsdelivr.net/gh/algorisys-oss/tinyfly@v0.120.0/cdn/tinyfly-embed.iife.js" data-tinyfly-auto></script>
|
|
29
29
|
```
|
|
30
30
|
|
|
31
31
|
Every `[data-tinyfly-embed]` mounts when the page is ready, with no per-post
|