Unverified50% confidenceFactExact time
After packaging Whistle into an APK in a Flutter project, the entire app is about 34MB; the model .cact file is placed in the assets directory and declared in pubspec.yaml.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/21/2026
First Seen
Valid until: 1/19/2027
Sources
17MB Speech Transcription Model Whistle: A New Option for On-Device Deployment
youtubeNeuralNine10/21/2026
Related Entities
Related Claims
UnverifiedTerminal command output exceeding 30,000 characters will be automatically saved to a file59% similarVerifiedThe pipeline ultimately packaged 39 finished products and 856 MB of output, backed by approximately 19,000 lines of parent project code and more than 300 pages of research material, all verified on a machine with 1 core, 2 GB of RAM, and no GPU.58% similarUnverifiedFor a 100MB Parquet file with 3,000 rows added near the beginning, nearly the entire file must be re-uploaded with default settings, whereas enabling content-defined chunking requires uploading only approximately 6 chunks.56% similarUnverifiedA previous approach based on Shen Sanshu's publicly available technical documentation achieved decoding speeds approaching 400 tokens/s and prefill of approximately 2500 tokens/s in a local 16GB VRAM environment.56% similarUnverifiedThe 256GB memory in this build runs stably at CL30 timings via EXPO overclocking.56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1082576API
curl https://kongchang.com/api/v1/knowledge/claims/1082576MCP
get_claim(id=1082576)