status 1 signal arrived during cgo execution GGML_ASSERT([rsets->data count] == 0) failed
#9
by jjido - opened
I use gemma 4 E2B in Ollama on MacOS.
I did not like a response, so I went to edit my question. But now I keep getting this:
an error was encountered while running the model: error: command buffer 1 failed with status 1 signal
arrived during cgo execution GGML_ASSERT([rsets->data count] == 0) failed
Do I need to restart the discussion from zero?
More info in server.log:
r0 0x0
r1 0x0
r2 0x0
r3 0x0
r4 0x184b9d8b7
r5 0x16d095cd0
r6 0x32
r7 0x0
r8 0x5fe7b05774696061
r9 0x5fe7b05619601061
r10 0x2
r11 0xfffffffd
r12 0x0
r13 0x0
r14 0x0
r15 0x0
r16 0x148
r17 0x1f20d5f20
r18 0x0
r19 0x6
r20 0x1a03
r21 0x16d0970e0
r22 0x1093e2c18
r23 0x2
r24 0x0
r25 0xf4493202150
r26 0xf44926f2750
r27 0x828
r28 0xf449258c1e0
r29 0x16d0965c0
lr 0x184c938d8
sp 0x16d0965a0
pc 0x184c585e8
fault 0x184c585e8
ggml-metal-device.m:608: GGML_ASSERT([rsets->data count] == 0) failed
WARNING: Using native backtrace. Set GGML_BACKTRACE_LLDB for more info.
WARNING: GGML_BACKTRACE_LLDB may cause native MacOS Terminal.app to crash.
See: https://github.com/ggml-org/llama.cpp/pull/17869
0 ollama 0x00000001055102ac ggml_print_backtrace + 276
1 ollama 0x0000000105510498 ggml_abort + 156
2 ollama 0x000000010575ee04 ggml_metal_device_init + 0
3 ollama 0x000000010575f6f8 ggml_metal_device_free + 28
4 ollama 0x00000001057628ec ggml_metal_device_get + 188
5 libsystem_c.dylib 0x0000000184b49e48 __cxa_finalize_ranges + 416
6 libsystem_c.dylib 0x0000000184b49c48 exit + 44
7 ollama 0x0000000104609548 _cgo_topofstack + 2568
8 ollama 0x00000001045ec7a8 ollama + 444328
9 ??? 0x00000f4492595978 0x0 + 16787187521912
panic during panic
SIGABRT: abort
PC=0x184c585e8 m=4 sigcode=0
Note that the model produces a few lines of reasoning before the error.
Oh well, it looks like I can switch the model midway. I selected gemma 4 E4B and got my answer, even though it is slow on my hardware. No error.