Architecture gallery: 435 full renders incl. Mamba/SSM mixers
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- README.md +17 -5
- afmoe.diff.svg +223 -0
- afmoe.svg +225 -0
- aimv2.diff.svg +219 -0
- aimv2.svg +207 -0
- albert.svg +180 -0
- align.svg +207 -0
- altclip.diff.svg +245 -0
- altclip.svg +227 -0
- apertus.diff.svg +181 -0
- apertus.svg +181 -0
- arcee.diff.svg +167 -0
- arcee.svg +179 -0
- aria.diff.svg +271 -0
- aria.svg +233 -0
- audio_spectrogram_transformer.diff.svg +175 -0
- audio_spectrogram_transformer.svg +172 -0
- audioflamingo3.diff.svg +204 -0
- audioflamingo3.svg +209 -0
- autoformer.svg +169 -0
- aya_vision.diff.svg +246 -0
- aya_vision.svg +263 -0
- bamba.diff.svg +182 -0
- bamba.svg +179 -0
- bark.svg +177 -0
- bart.svg +170 -0
- beit.diff.svg +214 -0
- beit.svg +172 -0
- bert.svg +175 -0
- bert_generation.svg +175 -0
- big_bird.svg +174 -0
- bigbird_pegasus.svg +171 -0
- biogpt.diff.svg +177 -0
- biogpt.svg +171 -0
- bit.svg +155 -0
- bitnet.diff.svg +176 -0
- bitnet.svg +182 -0
- blenderbot.svg +170 -0
- blenderbot_small.svg +170 -0
- blip.svg +203 -0
- blip_2.svg +199 -0
- bloom.svg +176 -0
- blt.diff.svg +198 -0
- blt.svg +180 -0
- bridgetower.svg +264 -0
- bros.svg +175 -0
- camembert.diff.svg +182 -0
- camembert.svg +176 -0
- canine.svg +179 -0
- chameleon.svg +182 -0
README.md
CHANGED
|
@@ -1,10 +1,22 @@
|
|
| 1 |
---
|
| 2 |
-
title: Model Architectures
|
| 3 |
-
emoji:
|
| 4 |
-
colorFrom:
|
| 5 |
-
colorTo:
|
| 6 |
sdk: static
|
| 7 |
pinned: false
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Transformers Model Architectures
|
| 3 |
+
emoji: 📐
|
| 4 |
+
colorFrom: indigo
|
| 5 |
+
colorTo: blue
|
| 6 |
sdk: static
|
| 7 |
pinned: false
|
| 8 |
+
license: apache-2.0
|
| 9 |
---
|
| 10 |
|
| 11 |
+
# 🤗 Transformers — Model Architecture Gallery
|
| 12 |
+
|
| 13 |
+
Static, dependency-free architecture diagrams for every model in
|
| 14 |
+
[`huggingface/transformers`](https://github.com/huggingface/transformers), generated by
|
| 15 |
+
introspecting each model built on the `meta` device (no weights, no network).
|
| 16 |
+
|
| 17 |
+
- **full** view — the decoder/encoder block decomposed top-to-bottom (attention or SSM mixer,
|
| 18 |
+
Q/K/V/O projections with real dims, norms, MLP / Sparse MoE, residual stream, mask grids).
|
| 19 |
+
- **diff** view — for models with a `modular_*.py`, what each model overrides / adds / removes
|
| 20 |
+
versus the parent it inherits from.
|
| 21 |
+
|
| 22 |
+
Open `index.html` for the browsable gallery.
|
afmoe.diff.svg
ADDED
|
|
afmoe.svg
ADDED
|
|
aimv2.diff.svg
ADDED
|
|
aimv2.svg
ADDED
|
|
albert.svg
ADDED
|
|
align.svg
ADDED
|
|
altclip.diff.svg
ADDED
|
|
altclip.svg
ADDED
|
|
apertus.diff.svg
ADDED
|
|
apertus.svg
ADDED
|
|
arcee.diff.svg
ADDED
|
|
arcee.svg
ADDED
|
|
aria.diff.svg
ADDED
|
|
aria.svg
ADDED
|
|
audio_spectrogram_transformer.diff.svg
ADDED
|
|
audio_spectrogram_transformer.svg
ADDED
|
|
audioflamingo3.diff.svg
ADDED
|
|
audioflamingo3.svg
ADDED
|
|
autoformer.svg
ADDED
|
|
aya_vision.diff.svg
ADDED
|
|
aya_vision.svg
ADDED
|
|
bamba.diff.svg
ADDED
|
|
bamba.svg
ADDED
|
|
bark.svg
ADDED
|
|
bart.svg
ADDED
|
|
beit.diff.svg
ADDED
|
|
beit.svg
ADDED
|
|
bert.svg
ADDED
|
|
bert_generation.svg
ADDED
|
|
big_bird.svg
ADDED
|
|
bigbird_pegasus.svg
ADDED
|
|
biogpt.diff.svg
ADDED
|
|
biogpt.svg
ADDED
|
|
bit.svg
ADDED
|
|
bitnet.diff.svg
ADDED
|
|
bitnet.svg
ADDED
|
|
blenderbot.svg
ADDED
|
|
blenderbot_small.svg
ADDED
|
|
blip.svg
ADDED
|
|
blip_2.svg
ADDED
|
|
bloom.svg
ADDED
|
|
blt.diff.svg
ADDED
|
|
blt.svg
ADDED
|
|
bridgetower.svg
ADDED
|
|
bros.svg
ADDED
|
|
camembert.diff.svg
ADDED
|
|
camembert.svg
ADDED
|
|
canine.svg
ADDED
|
|
chameleon.svg
ADDED
|
|