https://ggml.ai/whisper.cpp cant transcribe any audio file

[Piloair]

I tried mp3,waw,m4a, using tiny,base
In Windows 10, Brave Browser.

js: Running...
js: 
loadRemote: storage quota: 2147483648 bytes
loadRemote: storage usage: 713255784 bytes
loadRemote: "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.bin" is already in the IndexedDB
storeFS: stored model: whisper.bin size: 147951465
js: loading audio: CR-480 317.wav, size: 132974710 bytes
js: please wait ...
js: audio loaded, size: 22162444
whisper_init_from_file_with_params_no_state: loading model from 'whisper.bin'
whisper_init_with_params_no_state: use gpu    = 1
whisper_init_with_params_no_state: flash attn = 1
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw        = 0
whisper_init_with_params_no_state: devices    = 1
whisper_init_with_params_no_state: backends   = 1
whisper_model_load: loading model
whisper_model_load: n_vocab       = 51865
whisper_model_load: n_audio_ctx   = 1500
whisper_model_load: n_audio_state = 512
whisper_model_load: n_audio_head  = 8
whisper_model_load: n_audio_layer = 6
whisper_model_load: n_text_ctx    = 448
helpers.js:14 whisper_model_load: n_text_state  = 512
helpers.js:14 whisper_model_load: n_text_head   = 8
helpers.js:14 whisper_model_load: n_text_layer  = 6
helpers.js:14 whisper_model_load: n_mels        = 80
helpers.js:14 whisper_model_load: ftype         = 1
helpers.js:14 whisper_model_load: qntvr         = 0
helpers.js:14 whisper_model_load: type          = 2 (base)
helpers.js:14 whisper_model_load: adding 1608 extra tokens
helpers.js:14 whisper_model_load: n_langs       = 99
helpers.js:14 whisper_model_load:          CPU total size =   147.37 MB
helpers.js:14 whisper_model_load: model size    =  147.37 MB
helpers.js:14 whisper_backend_init_gpu: device 0: CPU (type: 0)
helpers.js:14 whisper_backend_init_gpu: no GPU found
helpers.js:14 whisper_init_state: kv self size  =    6.29 MB
helpers.js:14 whisper_init_state: kv cross size =   18.87 MB
helpers.js:14 whisper_init_state: kv pad  size  =    3.15 MB
helpers.js:14 whisper_init_state: compute buffer (conv)   =   15.87 MB
helpers.js:14 whisper_init_state: compute buffer (encode) =   22.68 MB
helpers.js:14 whisper_init_state: compute buffer (cross)  =    4.25 MB
helpers.js:14 whisper_init_state: compute buffer (decode) =   95.96 MB
helpers.js:14 js: whisper initialized, instance: 1
helpers.js:14 
helpers.js:14 js: processing - this might take a while ...
helpers.js:14 
VM26:4 Uncaught RangeError: Invalid typed array length: 22162444
   at new Float32Array (<anonymous>)
   at methodCaller<(emscripten::val,unsigned long,int) => emscripten::val> (eval at __emval_create_invoker (https://ggml.ai/whisper.cpp/main.js:1:1241566), <anonymous>:4:58)
   at __emval_invoke (https://ggml.ai/whisper.cpp/main.js:1:1242202)
   at wasm://wasm/00450d76:wasm-function[1138]:0xed09a
   at wasm://wasm/00450d76:wasm-function[1141]:0xed748
   at Object.full_default (eval at createJsInvoker (https://ggml.ai/whisper.cpp/main.js:1:1), <anonymous>:9:10)
   at https://ggml.ai/whisper.cpp/:648:42