Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
@@ -1,10 +1,10 @@
|
|
1 |
---
|
2 |
library_name: pytorch
|
3 |
license: mit
|
4 |
-
pipeline_tag: automatic-speech-recognition
|
5 |
tags:
|
6 |
- foundation
|
7 |
- android
|
|
|
8 |
|
9 |
---
|
10 |
|
@@ -38,34 +38,55 @@ More details on model performance across various devices, can be found
|
|
38 |
|
39 |
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
|
40 |
|---|---|---|---|---|---|---|---|---|
|
41 |
-
| WhisperDecoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE |
|
42 |
-
| WhisperDecoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 |
|
43 |
-
| WhisperDecoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 24.
|
44 |
-
| WhisperDecoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 |
|
45 |
-
| WhisperDecoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE |
|
46 |
-
| WhisperDecoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite |
|
47 |
-
| WhisperDecoder |
|
48 |
-
| WhisperDecoder | SA7255P ADP | SA7255P |
|
49 |
-
| WhisperDecoder | SA8255 (Proxy) | SA8255P Proxy | TFLITE |
|
50 |
-
| WhisperDecoder |
|
51 |
-
| WhisperDecoder |
|
52 |
-
| WhisperDecoder |
|
53 |
-
| WhisperDecoder |
|
54 |
-
| WhisperDecoder |
|
55 |
-
|
|
56 |
-
|
|
57 |
-
|
|
58 |
-
|
|
59 |
-
|
|
60 |
-
|
|
61 |
-
|
|
62 |
-
|
|
63 |
-
|
|
64 |
-
|
|
65 |
-
|
|
66 |
-
| WhisperEncoder |
|
67 |
-
| WhisperEncoder |
|
68 |
-
| WhisperEncoder |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
69 |
|
70 |
|
71 |
|
@@ -129,8 +150,8 @@ Profiling Results
|
|
129 |
WhisperDecoder
|
130 |
Device : Samsung Galaxy S23 (13)
|
131 |
Runtime : TFLITE
|
132 |
-
Estimated inference time (ms) :
|
133 |
-
Estimated peak memory usage (MB): [16,
|
134 |
Total # Ops : 2573
|
135 |
Compute Unit(s) : NPU (2573 ops)
|
136 |
|
@@ -138,8 +159,8 @@ Compute Unit(s) : NPU (2573 ops)
|
|
138 |
WhisperEncoder
|
139 |
Device : Samsung Galaxy S23 (13)
|
140 |
Runtime : TFLITE
|
141 |
-
Estimated inference time (ms) :
|
142 |
-
Estimated peak memory usage (MB): [110,
|
143 |
Total # Ops : 911
|
144 |
Compute Unit(s) : GPU (900 ops) CPU (11 ops)
|
145 |
```
|
|
|
1 |
---
|
2 |
library_name: pytorch
|
3 |
license: mit
|
|
|
4 |
tags:
|
5 |
- foundation
|
6 |
- android
|
7 |
+
pipeline_tag: automatic-speech-recognition
|
8 |
|
9 |
---
|
10 |
|
|
|
38 |
|
39 |
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
|
40 |
|---|---|---|---|---|---|---|---|---|
|
41 |
+
| WhisperDecoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 29.043 ms | 16 - 98 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
42 |
+
| WhisperDecoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 12.146 ms | 61 - 64 MB | FP16 | NPU | [Whisper-Small-En.so](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.so) |
|
43 |
+
| WhisperDecoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 24.349 ms | 13 - 145 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
44 |
+
| WhisperDecoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 9.735 ms | 61 - 76 MB | FP16 | NPU | [Whisper-Small-En.so](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.so) |
|
45 |
+
| WhisperDecoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 19.988 ms | 15 - 176 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
46 |
+
| WhisperDecoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 8.385 ms | 61 - 197 MB | FP16 | NPU | Use Export Script |
|
47 |
+
| WhisperDecoder | SA7255P ADP | SA7255P | TFLITE | 101.114 ms | 16 - 175 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
48 |
+
| WhisperDecoder | SA7255P ADP | SA7255P | QNN | 75.251 ms | 56 - 63 MB | FP16 | NPU | Use Export Script |
|
49 |
+
| WhisperDecoder | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 29.44 ms | 16 - 101 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
50 |
+
| WhisperDecoder | SA8255 (Proxy) | SA8255P Proxy | QNN | 11.752 ms | 64 - 66 MB | FP16 | NPU | Use Export Script |
|
51 |
+
| WhisperDecoder | SA8295P ADP | SA8295P | TFLITE | 30.886 ms | 16 - 163 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
52 |
+
| WhisperDecoder | SA8295P ADP | SA8295P | QNN | 14.568 ms | 57 - 67 MB | FP16 | NPU | Use Export Script |
|
53 |
+
| WhisperDecoder | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 29.578 ms | 16 - 101 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
54 |
+
| WhisperDecoder | SA8650 (Proxy) | SA8650P Proxy | QNN | 12.319 ms | 55 - 57 MB | FP16 | NPU | Use Export Script |
|
55 |
+
| WhisperDecoder | SA8775P ADP | SA8775P | TFLITE | 33.115 ms | 16 - 174 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
56 |
+
| WhisperDecoder | SA8775P ADP | SA8775P | QNN | 14.713 ms | 59 - 67 MB | FP16 | NPU | Use Export Script |
|
57 |
+
| WhisperDecoder | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 101.114 ms | 16 - 175 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
58 |
+
| WhisperDecoder | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 75.251 ms | 56 - 63 MB | FP16 | NPU | Use Export Script |
|
59 |
+
| WhisperDecoder | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 29.314 ms | 16 - 100 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
60 |
+
| WhisperDecoder | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 12.196 ms | 61 - 63 MB | FP16 | NPU | Use Export Script |
|
61 |
+
| WhisperDecoder | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 33.115 ms | 16 - 174 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
62 |
+
| WhisperDecoder | QCS9075 (Proxy) | QCS9075 Proxy | QNN | 14.713 ms | 59 - 67 MB | FP16 | NPU | Use Export Script |
|
63 |
+
| WhisperDecoder | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 32.534 ms | 16 - 137 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) |
|
64 |
+
| WhisperDecoder | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 16.208 ms | 53 - 167 MB | FP16 | NPU | Use Export Script |
|
65 |
+
| WhisperDecoder | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 10.845 ms | 61 - 61 MB | FP16 | NPU | Use Export Script |
|
66 |
+
| WhisperEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 714.232 ms | 110 - 177 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
67 |
+
| WhisperEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 683.763 ms | 0 - 3 MB | FP16 | NPU | [Whisper-Small-En.so](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.so) |
|
68 |
+
| WhisperEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 1134.657 ms | 29 - 121 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
69 |
+
| WhisperEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 454.493 ms | 1 - 17 MB | FP16 | NPU | [Whisper-Small-En.so](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.so) |
|
70 |
+
| WhisperEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 545.194 ms | 110 - 141 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
71 |
+
| WhisperEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 522.261 ms | 0 - 905 MB | FP16 | NPU | Use Export Script |
|
72 |
+
| WhisperEncoder | SA7255P ADP | SA7255P | TFLITE | 4461.436 ms | 108 - 141 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
73 |
+
| WhisperEncoder | SA7255P ADP | SA7255P | QNN | 3226.699 ms | 1 - 7 MB | FP16 | NPU | Use Export Script |
|
74 |
+
| WhisperEncoder | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 782.116 ms | 65 - 120 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
75 |
+
| WhisperEncoder | SA8255 (Proxy) | SA8255P Proxy | QNN | 662.251 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
|
76 |
+
| WhisperEncoder | SA8295P ADP | SA8295P | TFLITE | 657.203 ms | 108 - 139 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
77 |
+
| WhisperEncoder | SA8295P ADP | SA8295P | QNN | 700.971 ms | 0 - 11 MB | FP16 | NPU | Use Export Script |
|
78 |
+
| WhisperEncoder | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 716.632 ms | 49 - 222 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
79 |
+
| WhisperEncoder | SA8650 (Proxy) | SA8650P Proxy | QNN | 696.687 ms | 1 - 4 MB | FP16 | NPU | Use Export Script |
|
80 |
+
| WhisperEncoder | SA8775P ADP | SA8775P | TFLITE | 1291.081 ms | 104 - 137 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
81 |
+
| WhisperEncoder | SA8775P ADP | SA8775P | QNN | 603.281 ms | 1 - 8 MB | FP16 | NPU | Use Export Script |
|
82 |
+
| WhisperEncoder | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 4461.436 ms | 108 - 141 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
83 |
+
| WhisperEncoder | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 3226.699 ms | 1 - 7 MB | FP16 | NPU | Use Export Script |
|
84 |
+
| WhisperEncoder | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 737.787 ms | 102 - 187 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
85 |
+
| WhisperEncoder | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 645.71 ms | 1 - 3 MB | FP16 | NPU | Use Export Script |
|
86 |
+
| WhisperEncoder | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 1291.081 ms | 104 - 137 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
87 |
+
| WhisperEncoder | QCS9075 (Proxy) | QCS9075 Proxy | QNN | 603.281 ms | 1 - 8 MB | FP16 | NPU | Use Export Script |
|
88 |
+
| WhisperEncoder | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 1035.473 ms | 74 - 177 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) |
|
89 |
+
| WhisperEncoder | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 503.462 ms | 0 - 0 MB | FP16 | NPU | Use Export Script |
|
90 |
|
91 |
|
92 |
|
|
|
150 |
WhisperDecoder
|
151 |
Device : Samsung Galaxy S23 (13)
|
152 |
Runtime : TFLITE
|
153 |
+
Estimated inference time (ms) : 29.0
|
154 |
+
Estimated peak memory usage (MB): [16, 98]
|
155 |
Total # Ops : 2573
|
156 |
Compute Unit(s) : NPU (2573 ops)
|
157 |
|
|
|
159 |
WhisperEncoder
|
160 |
Device : Samsung Galaxy S23 (13)
|
161 |
Runtime : TFLITE
|
162 |
+
Estimated inference time (ms) : 714.2
|
163 |
+
Estimated peak memory usage (MB): [110, 177]
|
164 |
Total # Ops : 911
|
165 |
Compute Unit(s) : GPU (900 ops) CPU (11 ops)
|
166 |
```
|