whisper_init_state: ggml_metal_init() failed

[ToSeven]

./stream -m models/ggml-base.en.bin
init: found 3 capture devices:
init: - Capture device #0: 'MacBook Pro麦克风'
init: - Capture device #1: 'Infoflow Audio Device'
init: - Capture device #2: 'ZoomAudioDevice'
init: attempt to open default capture device ...
init: obtained spec for input device (SDL Id = 2):
init: - sample rate: 16000
init: - format: 33056 (required: 33056)
init: - channels: 1 (required: 1)
init: - samples per frame: 1024
whisper_init_from_file_no_state: loading model from 'models/ggml-base.en.bin'
whisper_model_load: loading model
whisper_model_load: n_vocab = 51864
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
whisper_model_load: n_text_state = 512
whisper_model_load: n_text_head = 8
whisper_model_load: n_text_layer = 6
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 2
whisper_model_load: adding 1607 extra tokens
whisper_model_load: model ctx = 140.66 MB
whisper_model_load: model size = 140.54 MB
whisper_init_state: kv self size = 5.25 MB
whisper_init_state: kv cross size = 17.58 MB
whisper_init_state: compute buffer (conv) = 14.10 MB
whisper_init_state: compute buffer (encode) = 81.85 MB
whisper_init_state: compute buffer (cross) = 4.40 MB
whisper_init_state: compute buffer (decode) = 24.61 MB
ggml_metal_init: allocating
ggml_metal_init: found device: Intel(R) Iris(TM) Plus Graphics 655
ggml_metal_init: picking default device: Intel(R) Iris(TM) Plus Graphics 655
ggml_metal_init: loading '/Volumes/extapple/projects/whisper.cpp/ggml-metal.metal'
ggml_metal_init: loaded kernel_add 0x7fc89c80d670 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_add_row 0x7fc89cd04c10 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_mul 0x7fc89cd05640 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_mul_row 0x7fc89cd05a10 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_scale 0x7fc89bf06be0 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_silu 0x7fc89bf072e0 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_relu 0x7fc89bf07b30 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_gelu 0x7fc89ca04990 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_soft_max 0x7fc89ce040c0 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_soft_max_4 0x7fc89cb04ab0 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_diag_mask_inf 0x7fc89cb05280 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_diag_mask_inf_8 0x7fc89cb059e0 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_f32 0x7fc89cb06010 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_f16 0x7fc89cb06700 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_q4_0 0x7fc89ce047c0 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_q4_1 0x7fc89ce05470 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_q8_0 0x7fc89ce05b60 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_q2_K 0x7fc89ce06250 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_q3_K 0x7fc89c80e3e0 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_q4_K 0x7fc89ce067a0 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_q5_K 0x7fc89ce06e90 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_get_rows_q6_K 0x7fc89ca0bc60 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_rms_norm 0x7fc89ca0d820 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_norm 0x7fc89ca0de00 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_mul_mat_f32_f32 0x7fc89c90a980 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_mul_mat_f16_f32 0x7fc89bf080c0 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_mul_mat_f16_f32_1row 0x7fc89bf08610 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_mul_mat_f16_f32_l4 0x7fc89bf08ba0 | th_max = 1024 | th_width = 32
ggml_metal_init: loaded kernel_mul_mat_q4_0_f32 0x7fc89cd05f90 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_mul_mat_q4_1_f32 0x7fc89cd06520 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_mul_mat_q8_0_f32 0x7fc89c80f150 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_mul_mat_q2_K_f32 0x7fc89c80f6e0 | th_max = 448 | th_width = 8
ggml_metal_init: loaded kernel_mul_mat_q3_K_f32 0x7fc89c80fad0 | th_max = 448 | th_width = 8
ggml_metal_init: loaded kernel_mul_mat_q4_K_f32 0x7fc89c90b270 | th_max = 448 | th_width = 8
ggml_metal_init: loaded kernel_mul_mat_q5_K_f32 0x7fc89c90b800 | th_max = 448 | th_width = 8
ggml_metal_init: loaded kernel_mul_mat_q6_K_f32 0x7fc89c90bd90 | th_max = 896 | th_width = 16
ggml_metal_init: loaded kernel_mul_mm_f32_f32 0x0 | th_max = 0 | th_width = 0
ggml_metal_init: load pipeline error: Error Domain=CompilerError Code=2 "AIR builtin function was called but no definition was found." UserInfo={NSLocalizedDescription=AIR builtin function was called but no definition was found.}
whisper_init_state: ggml_metal_init() failed

zsh: segmentation fault ./stream -m models/ggml-base.en.bin

[bobqianic]

Are you using Intel based Macbook Pro?

If so, disable Metal can solve this. See #3129 for more details

[fbernhard]

For others reading @bobqianic 's comment. The environment variable to disable Metal is called WHISPER_NO_METAL:

make clean
WHISPER_NO_METAL=1 make
WHISPER_NO_METAL=1 make base

[ToSeven]

got it. Thx all!