ggml_new_object: not enough space in the context's memory pool

[bobqianic]

C:\Users\qianp\Downloads\whisper.cpp_build-fix\bin\Release>C:\Users\qianp\Downloads\whisper.cpp_build-fix\bin\Release\bench.exe -m C:\Users\qianp\Downloads\whisper.cpp_build-fix\bin\Release\ggml-model-whisper-base.bin -w 2
ggml_new_object: not enough space in the context's memory pool (needed 301991760, available 268437504)

I just ran some tests and noticed that bench.cpp is corrupted. It seems this problem started with the commit 59a3d0c on September 5th. I'm guessing the ggml updates weren't properly reflected in bench.cpp whisper.cpp.

[dereklll]

same problem on mac os.

[dereklll]

./talk-llama -mw ~/work/whisper.cpp/ggml-model-whisper-base.bin -ml ~/work/7B/chinese-alpaca-2-7b-16k-hf/ggml-model-f16.bin -p "You" -t 12 -l zh -c 0
whisper_init_from_file_no_state: loading model from '/Users/lll/work/whisper.cpp/ggml-model-whisper-base.bin'
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
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: mem required = 310.00 MB (+ 6.00 MB per decoder)
whisper_model_load: adding 1608 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
llama.cpp: loading model from /Users/lll/work/7B/chinese-alpaca-2-7b-16k-hf/ggml-model-f16.bin
llama_model_load_internal: format = ggjt v1 (pre #1405)
llama_model_load_internal: n_vocab = 55296
llama_model_load_internal: n_ctx = 2048
llama_model_load_internal: n_embd = 4096
llama_model_load_internal: n_mult = 256
llama_model_load_internal: n_head = 32
llama_model_load_internal: n_layer = 32
llama_model_load_internal: n_rot = 128
llama_model_load_internal: ftype = 1 (mostly F16)
llama_model_load_internal: n_ff = 11008
llama_model_load_internal: n_parts = 1
llama_model_load_internal: model size = 7B
llama_model_load_internal: ggml ctx size = 0.09 MB
llama_model_load_internal: mem required = 15009.11 MB (+ 1026.00 MB per state)
.
llama_init_from_file: kv self size = 1024.00 MB

main: processing, 12 threads, lang = zh, task = transcribe, timestamps = 0 ...

init: found 1 capture devices:
init: - Capture device #0: 'Logitech USB Headset H340'
init: attempt to open capture device 0 : 'Logitech USB Headset H340' ...
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

main : initializing - please wait ...
ggml_new_object: not enough space in the context's memory pool (needed 910471808, available 805306368)
zsh: trace trap ./talk-llama -mw ~/work/whisper.cpp/ggml-model-whisper-base.bin -ml -p "You"

[dereklll]

@bobqianic
Do you still have this problem?
Did it（whisper : fix bench regression #1275） really solve the problem? I pull the latest code and still get the same problem.

[bobqianic]

@bobqianic
Do you still have this problem?
Did it（whisper : fix bench regression #1275） really solve the problem? I pull the latest code and still get the same problem.

It only fixes the issue in bench.cpp. You might need to start a separate issue for this.