context.vn
Menu

Small Model Hub

3,618 efficient AI models across 9 domains · Charts · API

ScorePercentile rank within each task — comparable across domains
Showing 14 models
#ModelDomainTaskParams ↑GFLOPsVRAMContextSpeedArchSource
1
Silero-VAD
Speechvoice activity detection0.24M—1 MB———Qualcomm AI Hub
2
YamNet
SpeechImage Classification3.73M—9 MB———Qualcomm AI Hub
3
abr-ai/niagara-9m-batch.en
SpeechSpeech Recognition9.00M—22 MB—7158 RTFxState-space model/CTCOpen ASR
4
abr-ai/niagara-19m-batch.en
SpeechSpeech Recognition20.0M—48 MB—4324 RTFxState-space model/CTCOpen ASR
5
abr-ai/niagara-38m-batch.en
SpeechSpeech Recognition38.0M—91 MB—4015 RTFx—Open ASR
6
DeepSpeech2
SpeechSpeech94.6M—227 MB———Qualcomm AI Hub
7
HuggingFace-WavLM-Base-Plus
SpeechSpeech95.1M—228 MB———Qualcomm AI Hub
8
nvidia/canary-180m-flash
SpeechSpeech Recognition180.0M—432 MB—2489 RTFxConformer-based/TransformerOpen ASR
9
FunASR-Conformer-EN
SpeechSpeech220.0M—528 MB———Qualcomm AI Hub
10
soundsgoodai/Zipformer-cr-ctc-transducer-XL-290M
SpeechSpeech Recognition290.0M—696 MB—158 RTFxZipformer/Pruned Stateless Transducer with k2 modified_beam_searchOpen ASR
11
soundsgoodai/Zipformer-transducer-XL-290M
SpeechSpeech Recognition290.0M—696 MB—141 RTFxZipformer/RNN-TOpen ASR
12
AutoArk-AI/Audio8-ASR-0.1B
SpeechSpeech Recognition324.0M—778 MB—719 RTFxQwen3-ASR audio encoder/8-layer Qwen-style causal LMOpen ASR
13
ibm-granite/granite-speech-5.0-470m-turboctc
SpeechSpeech Recognition470.0M—1.1 GB—12946 RTFxStreaming-style Conformer/CTCOpen ASR
14
ibm-granite/granite-speech-5.0-470m-turboctc-nc
SpeechSpeech Recognition470.0M—1.1 GB—12762 RTFxStreaming-style Conformer/CTCOpen ASR