0.00.001.219 I common_init_result: (for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on)
2026-07-27T13:10:48.905Z INFO uvicorn.access pid:73791 role:local address:0.0.0.0:9997 node:MacBook-Pro-3.local 127.0.0.1:58067 - "GET /v1/models/gemma-4/progress HTTP/1.1" 200
2026-07-27T13:10:48.906Z INFO uvicorn.access pid:73791 role:local address:0.0.0.0:9997 node:MacBook-Pro-3.local 127.0.0.1:58066 - "GET /v1/models/gemma-4/replicas HTTP/1.1" 200
0.00.490.421 W load: control-looking token: 50 '<|tool_response>' was not control-type; this is probably a bug in the model. its type will be overridden
0.00.492.586 W load: control-looking token: 212 '</s>' was not control-type; this is probably a bug in the model. its type will be overridden
0.00.495.427 W load: control-looking token: 1 '<eos>' was not control-type; this is probably a bug in the model. its type will be overridden
0.00.498.273 W load: special_eog_ids contains '<|tool_response>', removing '</s>' token from EOG list
2026-07-27T13:10:49.908Z INFO uvicorn.access pid:73791 role:local address:0.0.0.0:9997 node:MacBook-Pro-3.local 127.0.0.1:58066 - "GET /v1/models/gemma-4/progress HTTP/1.1" 200
2026-07-27T13:10:49.909Z INFO uvicorn.access pid:73791 role:local address:0.0.0.0:9997 node:MacBook-Pro-3.local 127.0.0.1:58067 - "GET /v1/models/gemma-4/replicas HTTP/1.1" 200
0.01.517.539 W llama_context: n_ctx_seq (32768) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
2026-07-27T13:10:50.906Z INFO uvicorn.access pid:73791 role:local address:0.0.0.0:9997 node:MacBook-Pro-3.local 127.0.0.1:58067 - "GET /v1/models/gemma-4/progress HTTP/1.1" 200
2026-07-27T13:10:50.907Z INFO uvicorn.access pid:73791 role:local address:0.0.0.0:9997 node:MacBook-Pro-3.local 127.0.0.1:58066 - "GET /v1/models/gemma-4/replicas HTTP/1.1" 200
0.02.265.762 I common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)
0.02.275.073 E ggml_metal_library_compile_pipeline: failed to compile pipeline: base = 'kernel_mul_mv_ext_bf16_f32_r1_2', name = 'kernel_mul_mv_ext_bf16_f32_r1_2_nsg=2_nxpsg=16_ne12=1_r2=1_r3=1'
0.02.275.115 E ggml_metal_library_compile_pipeline: Error Domain=MTLLibraryErrorDomain Code=5 "Function kernel_mul_mv_ext_bf16_f32_r1_2 was not found in the library" UserInfo={NSLocalizedDescription=Function kernel_mul_mv_ext_bf16_f32_r1_2 was not found in the library}
2026-07-27T13:10:50.987Z ERROR xinference.core.worker pid:73843 role:local address:0.0.0.0:52188 node:MacBook-Pro-3.local Failed to load model gemma-4-rep0