Error: llama runner process has terminated: signal: aborted (core dumped)详细错误日志：
2024-05-21T06:24:45.266916811Z time=2024-05-21T06:24:45.266Z level=INFO source=memory.go:127 msg="offload to gpu" layers.real=-1 layers.estimate=41 memory.available="23.2 GiB" memory.required.full="8.3 GiB" memory.required.partial="8.3 GiB" memory.required.kv="800.0 MiB" memory.weights.total="6.8 GiB" memory.weights.repeating="6.1 GiB" memory.weights.nonrepeating="741.9 MiB" memory.graph.full="301.8 MiB" memory.graph.partial="606.0 MiB" 2024-05-21T06:24:45.267462488Z time=2024-05-21T06:24:45.267Z level=INFO source=memory.go:127 msg="offload to gpu" layers.real=-1 layers.estimate=41 memory.available="23.2 GiB" memory.required.full="8.3 GiB" memory.required.partial="8.3 GiB" memory.required.kv="800.0 MiB" memory.weights.total="6.8 GiB" memory.weights.repeating="6.1 GiB" memory.weights.nonrepeating="741.9 MiB" memory.graph.full="301.8 MiB" memory.graph.partial="606.0 MiB" 2024-05-21T06:24:45.267705557Z time=2024-05-21T06:24:45.267Z level=INFO source=server.go:318 msg="starting llama server" cmd="/tmp/ollama3703814779/runners/cuda_v11/ollama_llama_server --model /root/.ollama/models/blobs/sha256-1851048f09798b44b41e905f66e0a00be8f0145133931de67aac1d424117862c --ctx-size 2048 --batch-size 512 --embedding --log-disable --n-gpu-layers 41 --parallel 1 --port 35511" 2024-05-21T06:24:45.267841042Z time=2024-05-21T06:24:45.267Z level=INFO source=sched.go:333 msg="loaded runners" count=1 2024-05-21T06:24:45.267857398Z time=2024-05-21T06:24:45.267Z level=INFO source=server.go:488 msg="waiting for llama runner to start responding" 2024-05-21T06:24:45.268153238Z time=2024-05-21T06:24:45.267Z level=INFO source=server.go:524 msg="waiting for server to become available" status="llm server error" 2024-05-21T06:24:45.283630907Z INFO [main] build info | build=1 commit="952d03d" tid="140597682634752" timestamp=1716272685 2024-05-21T06:24:45.283677293Z INFO [main] system info | n_threads=8 n_threads_batch=-1 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | FMA = 0 | NEON = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="140597682634752" timestamp=1716272685 total_threads=24 2024-05-21T06:24:45.283751767Z INFO [main] HTTP server listening | hostname="127.0.0.1" n_threads_http="23" port="35511" tid="140597682634752" timestamp=1716272685 2024-05-21T06:24:45.310000256Z llama_model_loader: loaded meta data with 21 key-value pairs and 483 tensors from /root/.ollama/models/blobs/sha256-1851048f09798b44b41e905f66e0a00be8f0145133931de67aac1d424117862c (version GGUF V3 (latest)) 2024-05-21T06:24:45.310015701Z llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. 2024-05-21T06:24:45.310018241Z llama_model_loader: - kv 0: general.architecture str = qwen2 2024-05-21T06:24:45.310019996Z llama_model_loader: - kv 1: general.name str = MindChat-Qwen2-4B 2024-05-21T06:24:45.310021701Z llama_model_loader: - kv 2: qwen2.block_count u32 = 40 2024-05-21T06:24:45.310023255Z llama_model_loader: - kv 3: qwen2.context_length u32 = 32768 2024-05-21T06:24:45.310024815Z llama_model_loader: - kv 4: qwen2.embedding_length u32 = 2560 2024-05-21T06:24:45.310026428Z llama_model_loader: - kv 5: qwen2.feed_forward_length u32 = 6912 2024-05-21T06:24:45.310028051Z llama_model_loader: - kv 6: qwen2.attention.head_count u32 = 20 2024-05-21T06:24:45.310029594Z llama_model_loader: - kv 7: qwen2.attention.head_count_kv u32 = 20 2024-05-21T06:24:45.310031212Z llama_model_loader: - kv 8: qwen2.rope.freq_base f32 = 5000000.000000 2024-05-21T06:24:45.310032784Z llama_model_loader: - kv 9: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001 2024-05-21T06:24:45.310034441Z llama_model_loader: - kv 10: general.file_type u32 = 1 2024-05-21T06:24:45.310035972Z llama_model_loader: - kv 11: tokenizer.ggml.model str = gpt2 2024-05-21T06:24:45.310037561Z llama_model_loader: - kv 12: tokenizer.ggml.pre str = qwen2 2024-05-21T06:24:45.336203511Z llama_model_loader: - kv 13: tokenizer.ggml.tokens arr[str,151936] = ["!", "\"", "#", "$", "%", "&", "'", ... 2024-05-21T06:24:45.340927274Z llama_model_loader: - kv 14: tokenizer.ggml.token_type arr[i32,151936] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... 2024-05-21T06:24:45.367506788Z llama_model_loader: - kv 15: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... 2024-05-21T06:24:45.367513925Z llama_model_loader: - kv 16: tokenizer.ggml.eos_token_id u32 = 151645 2024-05-21T06:24:45.367515206Z llama_model_loader: - kv 17: tokenizer.ggml.padding_token_id u32 = 151643 2024-05-21T06:24:45.367516167Z llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 151643 2024-05-21T06:24:45.367524184Z llama_model_loader: - kv 19: tokenizer.chat_template str = {% for message in messages %}{{'<|im_... 2024-05-21T06:24:45.367525679Z llama_model_loader: - kv 20: general.quantization_version u32 = 2 2024-05-21T06:24:45.367526625Z llama_model_loader: - type f32: 201 tensors 2024-05-21T06:24:45.367527532Z llama_model_loader: - type f16: 282 tensors 2024-05-21T06:24:45.434814810Z llama_model_load: error loading model: error loading model vocabulary: unknown pre-tokenizer type: 'qwen2' 2024-05-21T06:24:45.434831117Z llama_load_model_from_file: exception loading model 2024-05-21T06:24:45.444744162Z terminate called after throwing an instance of 'std::runtime_error' 2024-05-21T06:24:45.444757722Z what(): error loading model vocabulary: unknown pre-tokenizer type: 'qwen2' 2024-05-21T06:24:45.519276490Z time=2024-05-21T06:24:45.518Z level=ERROR source=sched.go:339 msg="error loading llama server" error="llama runner process has terminated: signal: aborted (core dumped) " 2024-05-21T06:24:45.519423089Z [GIN] 2024/05/21 - 06:24:45 | 500 | 692.156464ms | 127.0.0.1 | POST "/api/chat" 总结
截至目前为止github上已经有人提出了解决了方案
cd llama.cpp git reset --hard 46e12c4692a37bdd31a0432fc5153d7d22bc7f72 
