Hugging Face's logo

kruatech
/
gigaam-v3-mlx

Automatic Speech Recognition
MLX
Russian
speech-recognition
russian
offline
apple
apple-silicon
ios
ipados
macos
rnnt
conformer
sentencepiece
native-apple
Model card Files
xet
 Community
kruatech commited on
Commit
5cb61cc
·
verified ·
1 Parent(s): 781755e

Upload folder using huggingface_hub

Browse files
Files changed (2) hide show
  1. README.md +6 -6
  2. checksums.sha256 +1 -1
README.md CHANGED
@@ -35,11 +35,11 @@ No additional training or fine-tuning. The native runtime produces identical dec
35
 
36
  ## Platforms
37
 
38
- | Platform | Target | Runtime |
39
- |----------|--------|---------|
40
- | iOS | iPhone | Apple MLX (native) |
41
- | iPadOS | iPad | Apple MLX (native) |
42
- | macOS | Apple Silicon Mac | Apple MLX (native) |
43
 
44
  ---
45
 
@@ -289,7 +289,7 @@ End-to-end: the native runtime and the Python reference produce **identical deco
289
  - Optimized for native Apple MLX runtimes; not intended for server or Python-based inference.
290
  - Long audio should be processed in chunks by the host runtime.
291
  - Word-level timestamps are not included in the bundle.
292
- - iOS/iPadOS memory pressure should be verified on target devices.
293
  - This repository contains **model assets only** — no application code, no Swift SDK source.
294
 
295
  ---
 
35
 
36
  ## Platforms
37
 
38
+ | Platform | Target | Status |
39
+ |----------|--------|--------|
40
+ | iOS | iPhone | ✅ verified on iPhone 15 Pro Max |
41
+ | iPadOS | iPad | beta / needs verification |
42
+ | macOS | Apple Silicon Mac | ✅ verified |
43
 
44
  ---
45
 
 
289
  - Optimized for native Apple MLX runtimes; not intended for server or Python-based inference.
290
  - Long audio should be processed in chunks by the host runtime.
291
  - Word-level timestamps are not included in the bundle.
292
+ - iPadOS not yet verified on device.
293
  - This repository contains **model assets only** — no application code, no Swift SDK source.
294
 
295
  ---
checksums.sha256 CHANGED
@@ -1,4 +1,4 @@
1
- f321086edd313c75a1d287f922ab6e851a9f3dabdc78a5820605230919543e87 README.md
2
  993c538b6dceffb7b018c633a7cc8e4034e7ca975b5c27ecc8ecd679fa574d16 manifest.json
3
  2ff6d955a61003e0d3cb2155be36c14aa3df69c879713f30880ddd07d71fde21 weights.fp16.safetensors
4
  828c12c991019eef952a960661f25a92d6ad279591e2ea466b4aeddf1d20a18a tokenizer.model
 
1
+ 03641007b1f215a9ca676d967fdd3ee8ee3abee424a9b1957827a9bb62ec58cb README.md
2
  993c538b6dceffb7b018c633a7cc8e4034e7ca975b5c27ecc8ecd679fa574d16 manifest.json
3
  2ff6d955a61003e0d3cb2155be36c14aa3df69c879713f30880ddd07d71fde21 weights.fp16.safetensors
4
  828c12c991019eef952a960661f25a92d6ad279591e2ea466b4aeddf1d20a18a tokenizer.model