github-starred/ollama-ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-04-27 16:36:54 -05:00
Code Issues 2.2k Packages Projects Releases 100 Wiki Activity

Error: an error was encountered while running the model: CUDA error: unspecified launch failure #6030

New Issue
Open
opened 2025-11-12 13:19:38 -06:00 by GiteaMirror · 2 comments
GiteaMirror commented 2025-11-12 13:19:38 -06:00
Owner
Copy Link

Originally created by @CJJ-Michael on GitHub (Feb 20, 2025).

What is the issue?

When I ask a question to LLM, it alaways give the error as the title shows. So, I have to use the CPU ONLY mode to run the model.

The error message is:

Error: an error was encountered while running the model: CUDA error: unspecified launch failure
current device: 0, in function ggml_backend_cuda_synchronize at llama/ggml-cuda/ggml-cuda.cu:2317
cudaStreamSynchronize(cuda_ctx->stream())
llama/ggml-cuda/ggml-cuda.cu:96: CUDA error

I tried changing different versions of model (size), Ollama, Nvidia driver and CUDA, but it's useless.

Relevant log output

ollama run deepseek-r1:7b
>>> 解释一下勾股定理
<think>
嗯，今天我在数学课上听老师讲到了勾股定理，但是我不太明白它具体是什么意思。勾股定理是什么呢Error: an error was encountered while running the model: CUDA error: unspecified launch failure
  current device: 0, in function ggml_backend_cuda_synchronize at llama/ggml-cuda/ggml-cuda.cu:2317
  cudaStreamSynchronize(cuda_ctx->stream())
llama/ggml-cuda/ggml-cuda.cu:96: CUDA error

OS

Linux

GPU

Nvidia

CPU

Intel

Ollama version

0.5.7 and 0.5.11

Originally created by @CJJ-Michael on GitHub (Feb 20, 2025). ### What is the issue? When I ask a question to LLM, it alaways give the error as the title shows. So, I have to use the CPU ONLY mode to run the model. The error message is: Error: an error was encountered while running the model: CUDA error: unspecified launch failure current device: 0, in function ggml_backend_cuda_synchronize at llama/ggml-cuda/ggml-cuda.cu:2317 cudaStreamSynchronize(cuda_ctx->stream()) llama/ggml-cuda/ggml-cuda.cu:96: CUDA error I tried changing different versions of model (size), Ollama, Nvidia driver and CUDA, but it's useless. ### Relevant log output ```shell ollama run deepseek-r1:7b >>> 解释一下勾股定理 <think> 嗯，今天我在数学课上听老师讲到了勾股定理，但是我不太明白它具体是什么意思。勾股定理是什么呢Error: an error was encountered while running the model: CUDA error: unspecified launch failure current device: 0, in function ggml_backend_cuda_synchronize at llama/ggml-cuda/ggml-cuda.cu:2317 cudaStreamSynchronize(cuda_ctx->stream()) llama/ggml-cuda/ggml-cuda.cu:96: CUDA error ``` ### OS Linux ### GPU Nvidia ### CPU Intel ### Ollama version 0.5.7 and 0.5.11
GiteaMirror added the bug label 2025-11-12 13:19:38 -06:00
GiteaMirror commented 2025-11-12 13:19:39 -06:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 20, 2025):

Server logs may help in debugging.

@rick-github commented on GitHub (Feb 20, 2025): [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) may help in debugging.
GiteaMirror commented 2025-11-12 13:19:39 -06:00
Author
Owner
Copy Link

@CJJ-Michael commented on GitHub (Feb 21, 2025):

Server logs may help in debugging.

Logs below belong to ollama server after the bug happend:

2月 21 08:52:10 GBITDeepSeek ollama[698750]: [GIN] 2025/02/21 - 08:52:10 | 200 | 3.857530363s | 127.0.0.1 | POST "/api/chat"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: [GIN] 2025/02/21 - 08:55:16 | 200 | 27.206µs | 172.17.0.2 | GET "/"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.213+08:00 level=INFO source=sched.go:507 msg="updated VRAM based on existing loaded models" gpu=GPU-99e4c40f-5e53-642f-aa9b-9c3b203cfebb library=cuda total="23.6 GiB" available="18.1 GiB"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.213+08:00 level=INFO source=sched.go:714 msg="new model will fit in available VRAM in single GPU, loading" model=/home/gbit/ollama/models/blobs/sha256-970aa74c0a90ef7482477cf803618e776e173c007bf957f635f1015bfcfef0e6 gpu=GPU-99e4c40f-5e53-642f-aa9b-9c3b203cfebb parallel=1 available=19391528960 required="809.9 MiB"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.309+08:00 level=INFO source=server.go:105 msg="system memory" total="251.5 GiB" free="238.0 GiB" free_swap="2.0 GiB"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.309+08:00 level=INFO source=memory.go:356 msg="offload to cuda" layers.requested=-1 layers.model=13 layers.offload=13 layers.split="" memory.available="[18.1 GiB]" memory.gpu_overhead="0 B" memory.required.full="809.9 MiB" memory.required.partial="809.9 MiB" memory.required.kv="24.0 MiB" memory.required.allocations="[809.9 MiB]" memory.weights.total="240.1 MiB" memory.weights.repeating="195.4 MiB" memory.weights.nonrepeating="44.7 MiB" memory.graph.full="48.0 MiB" memory.graph.partial="48.0 MiB"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.311+08:00 level=INFO source=server.go:397 msg="starting llama server" cmd="/tmp/ollama3857967831/runners/cuda_v12/ollama_llama_server --model /home/gbit/ollama/models/blobs/sha256-970aa74c0a90ef7482477cf803618e776e173c007bf957f635f1015bfcfef0e6 --ctx-size 8192 --batch-size 512 --n-gpu-layers 13 --threads 52 --parallel 1 --port 42335"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.312+08:00 level=INFO source=sched.go:449 msg="loaded runners" count=2
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.312+08:00 level=INFO source=server.go:576 msg="waiting for llama runner to start responding"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.312+08:00 level=INFO source=server.go:610 msg="waiting for server to become available" status="llm server error"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.372+08:00 level=INFO source=runner.go:941 msg="starting go runner"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.372+08:00 level=INFO source=runner.go:942 msg=system info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | RISCV_VECT = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | cgo(gcc)" threads=52
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.372+08:00 level=INFO source=.:0 msg="Server listening on 127.0.0.1:42335"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: loaded meta data with 24 key-value pairs and 112 tensors from /home/gbit/ollama/models/blobs/sha256-970aa74c0a90ef7482477cf803618e776e173c007bf957f635f1015bfcfef0e6 (version GGUF V3 (latest))
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 0: general.architecture str = nomic-bert
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 1: general.name str = nomic-embed-text-v1.5
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 2: nomic-bert.block_count u32 = 12
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 3: nomic-bert.context_length u32 = 2048
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 4: nomic-bert.embedding_length u32 = 768
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 5: nomic-bert.feed_forward_length u32 = 3072
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 6: nomic-bert.attention.head_count u32 = 12
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 7: nomic-bert.attention.layer_norm_epsilon f32 = 0.000000
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 8: general.file_type u32 = 1
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 9: nomic-bert.attention.causal bool = false
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 10: nomic-bert.pooling_type u32 = 1
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 11: nomic-bert.rope.freq_base f32 = 1000.000000
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 12: tokenizer.ggml.token_type_count u32 = 2
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 13: tokenizer.ggml.bos_token_id u32 = 101
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 14: tokenizer.ggml.eos_token_id u32 = 102
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 15: tokenizer.ggml.model str = bert
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 16: tokenizer.ggml.tokens arr[str,30522] = ["[PAD]", "[unused0]", "[unused1]", "...
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 17: tokenizer.ggml.scores arr[f32,30522] = [-1000.000000, -1000.000000, -1000.00...
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 18: tokenizer.ggml.token_type arr[i32,30522] = [3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 19: tokenizer.ggml.unknown_token_id u32 = 100
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 20: tokenizer.ggml.seperator_token_id u32 = 102
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 22: tokenizer.ggml.cls_token_id u32 = 101
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 23: tokenizer.ggml.mask_token_id u32 = 103
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - type f32: 51 tensors
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - type f16: 61 tensors
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_vocab: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_vocab: special tokens cache size = 5
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_vocab: token to piece cache size = 0.2032 MB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: format = GGUF V3 (latest)
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: arch = nomic-bert
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab type = WPM
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_vocab = 30522
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_merges = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab_only = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ctx_train = 2048
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd = 768
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_layer = 12
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_head = 12
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_head_kv = 12
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_rot = 64
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_swa = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_head_k = 64
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_head_v = 64
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_gqa = 1
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_k_gqa = 768
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_v_gqa = 768
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_norm_eps = 1.0e-12
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_norm_rms_eps = 0.0e+00
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_clamp_kqv = 0.0e+00
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_max_alibi_bias = 0.0e+00
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_logit_scale = 0.0e+00
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ff = 3072
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_expert = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_expert_used = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: causal attn = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: pooling type = 1
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope type = 2
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope scaling = linear
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: freq_base_train = 1000.0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: freq_scale_train = 1
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ctx_orig_yarn = 2048
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope_finetuned = unknown
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_conv = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_inner = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_state = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_dt_rank = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_dt_b_c_rms = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: model type = 137M
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: model ftype = F16
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: model params = 136.73 M
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: model size = 260.86 MiB (16.00 BPW)
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: general.name = nomic-embed-text-v1.5
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: BOS token = 101 '[CLS]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOS token = 102 '[SEP]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: UNK token = 100 '[UNK]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: SEP token = 102 '[SEP]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: PAD token = 0 '[PAD]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: CLS token = 101 '[CLS]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: MASK token = 103 '[MASK]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: LF token = 0 '[PAD]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOG token = 102 '[SEP]'
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: max token length = 21
2月 21 08:55:16 GBITDeepSeek ollama[698750]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
2月 21 08:55:16 GBITDeepSeek ollama[698750]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
2月 21 08:55:16 GBITDeepSeek ollama[698750]: ggml_cuda_init: found 1 CUDA devices:
2月 21 08:55:16 GBITDeepSeek ollama[698750]: Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: ggml ctx size = 0.10 MiB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: offloading 12 repeating layers to GPU
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: offloading non-repeating layers to GPU
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: offloaded 13/13 layers to GPU
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: CPU buffer size = 44.72 MiB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: CUDA0 buffer size = 216.15 MiB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.565+08:00 level=INFO source=server.go:610 msg="waiting for server to become available" status="llm server loading model"
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_ctx = 8192
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_batch = 512
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_ubatch = 512
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: flash_attn = 0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: freq_base = 1000.0
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: freq_scale = 1
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_kv_cache_init: CUDA0 KV buffer size = 288.00 MiB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: KV self size = 288.00 MiB, K (f16): 144.00 MiB, V (f16): 144.00 MiB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CPU output buffer size = 0.00 MiB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA0 compute buffer size = 23.00 MiB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA_Host compute buffer size = 3.50 MiB
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: graph nodes = 453
2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: graph splits = 2
2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.816+08:00 level=INFO source=server.go:615 msg="llama runner started in 0.50 seconds"
2月 21 08:55:17 GBITDeepSeek ollama[698750]: [GIN] 2025/02/21 - 08:55:17 | 200 | 919.068568ms | 172.17.0.2 | POST "/api/embeddings"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.487+08:00 level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.095851007 model=/home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.701+08:00 level=INFO source=sched.go:507 msg="updated VRAM based on existing loaded models" gpu=GPU-99e4c40f-5e53-642f-aa9b-9c3b203cfebb library=cuda total="23.6 GiB" available="21.4 GiB"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.701+08:00 level=INFO source=sched.go:714 msg="new model will fit in available VRAM in single GPU, loading" model=/home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 gpu=GPU-99e4c40f-5e53-642f-aa9b-9c3b203cfebb parallel=4 available=22992453632 required="6.5 GiB"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.737+08:00 level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.345946677 model=/home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.789+08:00 level=INFO source=server.go:105 msg="system memory" total="251.5 GiB" free="237.6 GiB" free_swap="2.0 GiB"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.789+08:00 level=INFO source=memory.go:356 msg="offload to cuda" layers.requested=-1 layers.model=29 layers.offload=29 layers.split="" memory.available="[21.4 GiB]" memory.gpu_overhead="0 B" memory.required.full="6.5 GiB" memory.required.partial="6.5 GiB" memory.required.kv="896.0 MiB" memory.required.allocations="[6.5 GiB]" memory.weights.total="4.5 GiB" memory.weights.repeating="4.1 GiB" memory.weights.nonrepeating="426.4 MiB" memory.graph.full="942.0 MiB" memory.graph.partial="1.1 GiB"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.791+08:00 level=INFO source=server.go:397 msg="starting llama server" cmd="/tmp/ollama3857967831/runners/cuda_v12/ollama_llama_server --model /home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 --ctx-size 16384 --batch-size 512 --n-gpu-layers 29 --threads 52 --mlock --parallel 4 --port 35995"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.792+08:00 level=INFO source=sched.go:449 msg="loaded runners" count=2
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.792+08:00 level=INFO source=server.go:576 msg="waiting for llama runner to start responding"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.792+08:00 level=INFO source=server.go:610 msg="waiting for server to become available" status="llm server error"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.849+08:00 level=INFO source=runner.go:941 msg="starting go runner"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.850+08:00 level=INFO source=runner.go:942 msg=system info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | RISCV_VECT = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | cgo(gcc)" threads=52
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.850+08:00 level=INFO source=.:0 msg="Server listening on 127.0.0.1:35995"
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: loaded meta data with 26 key-value pairs and 339 tensors from /home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 (version GGUF V3 (latest))
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 0: general.architecture str = qwen2
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 1: general.type str = model
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 2: general.name str = DeepSeek R1 Distill Qwen 7B
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 3: general.basename str = DeepSeek-R1-Distill-Qwen
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 4: general.size_label str = 7B
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 5: qwen2.block_count u32 = 28
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 6: qwen2.context_length u32 = 131072
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 7: qwen2.embedding_length u32 = 3584
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 8: qwen2.feed_forward_length u32 = 18944
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 9: qwen2.attention.head_count u32 = 28
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 10: qwen2.attention.head_count_kv u32 = 4
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 11: qwen2.rope.freq_base f32 = 10000.000000
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 12: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 13: general.file_type u32 = 15
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 14: tokenizer.ggml.model str = gpt2
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 15: tokenizer.ggml.pre str = qwen2
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 16: tokenizer.ggml.tokens arr[str,152064] = ["!", """, "#", "$", "%", "&", "'", ...
2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 17: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.986+08:00 level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.59510752 model=/home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 18: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 19: tokenizer.ggml.bos_token_id u32 = 151646
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 151643
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 151643
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 22: tokenizer.ggml.add_bos_token bool = true
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 23: tokenizer.ggml.add_eos_token bool = false
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 24: tokenizer.chat_template str = {% if not add_generation_prompt is de...
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 25: general.quantization_version u32 = 2
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - type f32: 141 tensors
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - type q4_K: 169 tensors
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - type q6_K: 29 tensors
2月 21 08:55:23 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:23.044+08:00 level=INFO source=server.go:610 msg="waiting for server to become available" status="llm server loading model"
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_vocab: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_vocab: special tokens cache size = 22
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_vocab: token to piece cache size = 0.9310 MB
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: format = GGUF V3 (latest)
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: arch = qwen2
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab type = BPE
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_vocab = 152064
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_merges = 151387
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab_only = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ctx_train = 131072
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd = 3584
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_layer = 28
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_head = 28
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_head_kv = 4
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_rot = 128
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_swa = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_head_k = 128
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_head_v = 128
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_gqa = 7
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_k_gqa = 512
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_v_gqa = 512
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_norm_eps = 0.0e+00
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_norm_rms_eps = 1.0e-06
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_clamp_kqv = 0.0e+00
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_max_alibi_bias = 0.0e+00
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_logit_scale = 0.0e+00
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ff = 18944
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_expert = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_expert_used = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: causal attn = 1
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: pooling type = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope type = 2
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope scaling = linear
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: freq_base_train = 10000.0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: freq_scale_train = 1
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ctx_orig_yarn = 131072
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope_finetuned = unknown
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_conv = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_inner = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_state = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_dt_rank = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_dt_b_c_rms = 0
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: model type = ?B
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: model ftype = Q4_K - Medium
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: model params = 7.62 B
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: model size = 4.36 GiB (4.91 BPW)
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: general.name = DeepSeek R1 Distill Qwen 7B
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: BOS token = 151646 '<｜begin▁of▁sentence｜>'
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOS token = 151643 '<｜end▁of▁sentence｜>'
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: PAD token = 151643 '<｜end▁of▁sentence｜>'
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: LF token = 148848 'ÄĬ'
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOG token = 151643 '<｜end▁of▁sentence｜>'
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: max token length = 256
2月 21 08:55:23 GBITDeepSeek ollama[698750]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
2月 21 08:55:23 GBITDeepSeek ollama[698750]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
2月 21 08:55:23 GBITDeepSeek ollama[698750]: ggml_cuda_init: found 1 CUDA devices:
2月 21 08:55:23 GBITDeepSeek ollama[698750]: Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: ggml ctx size = 0.30 MiB
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: offloading 28 repeating layers to GPU
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: offloading non-repeating layers to GPU
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: offloaded 29/29 layers to GPU
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: CPU buffer size = 292.36 MiB
2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: CUDA0 buffer size = 4168.09 MiB
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_ctx = 16384
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_batch = 2048
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_ubatch = 512
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: flash_attn = 0
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: freq_base = 10000.0
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: freq_scale = 1
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_kv_cache_init: CUDA0 KV buffer size = 896.00 MiB
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: KV self size = 896.00 MiB, K (f16): 448.00 MiB, V (f16): 448.00 MiB
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA_Host output buffer size = 2.38 MiB
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA0 compute buffer size = 956.00 MiB
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA_Host compute buffer size = 39.01 MiB
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: graph nodes = 986
2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: graph splits = 2
2月 21 08:55:25 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:25.055+08:00 level=INFO source=server.go:615 msg="llama runner started in 2.26 seconds"
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: loaded meta data with 26 key-value pairs and 339 tensors from /home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 (version GGUF V3 (latest))
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 0: general.architecture str = qwen2
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 1: general.type str = model
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 2: general.name str = DeepSeek R1 Distill Qwen 7B
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 3: general.basename str = DeepSeek-R1-Distill-Qwen
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 4: general.size_label str = 7B
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 5: qwen2.block_count u32 = 28
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 6: qwen2.context_length u32 = 131072
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 7: qwen2.embedding_length u32 = 3584
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 8: qwen2.feed_forward_length u32 = 18944
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 9: qwen2.attention.head_count u32 = 28
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 10: qwen2.attention.head_count_kv u32 = 4
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 11: qwen2.rope.freq_base f32 = 10000.000000
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 12: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 13: general.file_type u32 = 15
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 14: tokenizer.ggml.model str = gpt2
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 15: tokenizer.ggml.pre str = qwen2
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 16: tokenizer.ggml.tokens arr[str,152064] = ["!", """, "#", "$", "%", "&", "'", ...
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 17: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 18: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 19: tokenizer.ggml.bos_token_id u32 = 151646
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 151643
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 151643
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 22: tokenizer.ggml.add_bos_token bool = true
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 23: tokenizer.ggml.add_eos_token bool = false
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 24: tokenizer.chat_template str = {% if not add_generation_prompt is de...
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 25: general.quantization_version u32 = 2
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - type f32: 141 tensors
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - type q4_K: 169 tensors
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - type q6_K: 29 tensors
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_vocab: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_vocab: special tokens cache size = 22
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_vocab: token to piece cache size = 0.9310 MB
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: format = GGUF V3 (latest)
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: arch = qwen2
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab type = BPE
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_vocab = 152064
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_merges = 151387
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab_only = 1
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: model type = ?B
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: model ftype = all F32
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: model params = 7.62 B
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: model size = 4.36 GiB (4.91 BPW)
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: general.name = DeepSeek R1 Distill Qwen 7B
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: BOS token = 151646 '<｜begin▁of▁sentence｜>'
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOS token = 151643 '<｜end▁of▁sentence｜>'
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: PAD token = 151643 '<｜end▁of▁sentence｜>'
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: LF token = 148848 'ÄĬ'
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOG token = 151643 '<｜end▁of▁sentence｜>'
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: max token length = 256
2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_load: vocab only - skipping tensors
2月 21 08:55:27 GBITDeepSeek ollama[698750]: CUDA error: misaligned address
2月 21 08:55:27 GBITDeepSeek ollama[698750]: current device: 0, in function ggml_backend_cuda_synchronize at ggml-cuda.cu:2508
2月 21 08:55:27 GBITDeepSeek ollama[698750]: cudaStreamSynchronize(cuda_ctx->stream())
2月 21 08:55:27 GBITDeepSeek ollama[698750]: ggml-cuda.cu:132: CUDA error
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 716998]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 716999]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717000]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717001]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717002]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717009]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717022]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717023]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717029]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [Thread debugging using libthread_db enabled]
2月 21 08:55:27 GBITDeepSeek ollama[717146]: Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
2月 21 08:55:27 GBITDeepSeek ollama[717146]: 0x00005ed082619083 in ?? ()
2月 21 08:55:27 GBITDeepSeek ollama[717146]: #0 0x00005ed082619083 in ?? ()
2月 21 08:55:27 GBITDeepSeek ollama[717146]: #1 0x00005ed0825de3d0 in _start ()
2月 21 08:55:27 GBITDeepSeek ollama[717146]: [Inferior 1 (process 716996) detached]
2月 21 08:55:27 GBITDeepSeek ollama[698750]: SIGABRT: abort
2月 21 08:55:27 GBITDeepSeek ollama[698750]: PC=0x7440f20969fc m=4 sigcode=18446744073709551610
2月 21 08:55:27 GBITDeepSeek ollama[698750]: signal arrived during cgo execution
2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 7 gp=0xc000252000 m=4 mp=0xc000139808 [syscall]:
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.cgocall(0x5ed08282d8c0, 0xc000147c60)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/cgocall.go:157 +0x4b fp=0xc000147c38 sp=0xc000147c00 pc=0x5ed0825ae8ab
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama._Cfunc_gpt_sampler_csample(0x744098008400, 0x74409c006450, 0x0)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: _cgo_gotypes.go:468 +0x4f fp=0xc000147c60 sp=0xc000147c38 pc=0x5ed0826ab98f
2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).processBatch.(*SamplingContext).Sample.func4(0x5ed082e32100?, 0x0?, 0x0)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/llama.go:699 +0x86 fp=0xc000147cb0 sp=0xc000147c60 pc=0x5ed082829346
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama.(*SamplingContext).Sample(...)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/llama.go:699
2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).processBatch(0xc000214120, 0xc0002ae000, 0xc000147f10)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:487 +0x6ab fp=0xc000147ed0 sp=0xc000147cb0 pc=0x5ed08282860b
2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).run(0xc000214120, {0x5ed082b7a9a0, 0xc00016a0a0})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:342 +0x1e5 fp=0xc000147fb8 sp=0xc000147ed0 pc=0x5ed082827c25
2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.main.gowrap2()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:980 +0x28 fp=0xc000147fe0 sp=0xc000147fb8 pc=0x5ed08282ca88
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc000147fe8 sp=0xc000147fe0 pc=0x5ed0826172c1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by main.main in goroutine 1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:980 +0xd3e
2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 1 gp=0xc0000061c0 m=nil [IO wait]:
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0x1?, 0xc00002b8e0?, 0xd4?, 0x52?, 0xc00002b8c0?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc00002b860 sp=0xc00002b840 pc=0x5ed0825e54ee
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.netpollblock(0x10?, 0x825ae006?, 0xd0?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/netpoll.go:573 +0xf7 fp=0xc00002b898 sp=0xc00002b860 pc=0x5ed0825dd737
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.runtime_pollWait(0x744117b88020, 0x72)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/netpoll.go:345 +0x85 fp=0xc00002b8b8 sp=0xc00002b898 pc=0x5ed082611f85
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*pollDesc).wait(0x3?, 0x744117bc3e88?, 0x0)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_poll_runtime.go:84 +0x27 fp=0xc00002b8e0 sp=0xc00002b8b8 pc=0x5ed082661ea7
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*pollDesc).waitRead(...)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_poll_runtime.go:89
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*FD).Accept(0xc00024c080)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_unix.go:611 +0x2ac fp=0xc00002b988 sp=0xc00002b8e0 pc=0x5ed08266336c
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*netFD).accept(0xc00024c080)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/fd_unix.go:172 +0x29 fp=0xc00002ba40 sp=0xc00002b988 pc=0x5ed0826d1fa9
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*TCPListener).accept(0xc0001521e0)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/tcpsock_posix.go:159 +0x1e fp=0xc00002ba68 sp=0xc00002ba40 pc=0x5ed0826e2cde
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*TCPListener).Accept(0xc0001521e0)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/tcpsock.go:327 +0x30 fp=0xc00002ba98 sp=0xc00002ba68 pc=0x5ed0826e2030
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*onceCloseListener).Accept(0xc000298090?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: :1 +0x24 fp=0xc00002bab0 sp=0xc00002ba98 pc=0x5ed082809244
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*Server).Serve(0xc0000181e0, {0x5ed082b7a360, 0xc0001521e0})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:3260 +0x33e fp=0xc00002bbe0 sp=0xc00002bab0 pc=0x5ed08280005e
2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.main()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:1000 +0x10cd fp=0xc00002bf50 sp=0xc00002bbe0 pc=0x5ed08282c80d
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.main()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:271 +0x29d fp=0xc00002bfe0 sp=0xc00002bf50 pc=0x5ed0825e50bd
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc00002bfe8 sp=0xc00002bfe0 pc=0x5ed0826172c1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 2 gp=0xc000006c40 m=nil [force gc (idle)]:
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000132fa8 sp=0xc000132f88 pc=0x5ed0825e54ee
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goparkunlock(...)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:408
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.forcegchelper()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:326 +0xb8 fp=0xc000132fe0 sp=0xc000132fa8 pc=0x5ed0825e5378
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc000132fe8 sp=0xc000132fe0 pc=0x5ed0826172c1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by runtime.init.6 in goroutine 1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:314 +0x1a
2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 3 gp=0xc000007180 m=nil [GC sweep wait]:
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000133780 sp=0xc000133760 pc=0x5ed0825e54ee
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goparkunlock(...)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:408
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.bgsweep(0xc00015c000)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgcsweep.go:278 +0x94 fp=0xc0001337c8 sp=0xc000133780 pc=0x5ed0825d0034
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gcenable.gowrap1()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgc.go:203 +0x25 fp=0xc0001337e0 sp=0xc0001337c8 pc=0x5ed0825c4b65
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc0001337e8 sp=0xc0001337e0 pc=0x5ed0826172c1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by runtime.gcenable in goroutine 1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgc.go:203 +0x66
2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 4 gp=0xc000007340 m=nil [GC scavenge wait]:
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0xc00015c000?, 0x5ed082a774f0?, 0x1?, 0x0?, 0xc000007340?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000133f78 sp=0xc000133f58 pc=0x5ed0825e54ee
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goparkunlock(...)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:408
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.(*scavengerState).park(0x5ed082d49560)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgcscavenge.go:425 +0x49 fp=0xc000133fa8 sp=0xc000133f78 pc=0x5ed0825cda29
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.bgscavenge(0xc00015c000)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgcscavenge.go:653 +0x3c fp=0xc000133fc8 sp=0xc000133fa8 pc=0x5ed0825cdfbc
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gcenable.gowrap2()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgc.go:204 +0x25 fp=0xc000133fe0 sp=0xc000133fc8 pc=0x5ed0825c4b05
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc000133fe8 sp=0xc000133fe0 pc=0x5ed0826172c1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by runtime.gcenable in goroutine 1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgc.go:204 +0xa5
2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 5 gp=0xc000007c00 m=nil [finalizer wait]:
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0xc000132648?, 0x5ed0825b8465?, 0xa8?, 0x1?, 0xc0000061c0?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000132620 sp=0xc000132600 pc=0x5ed0825e54ee
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.runfinq()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mfinal.go:194 +0x107 fp=0xc0001327e0 sp=0xc000132620 pc=0x5ed0825c3ba7
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc0001327e8 sp=0xc0001327e0 pc=0x5ed0826172c1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by runtime.createfing in goroutine 1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mfinal.go:164 +0x3d
2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 35 gp=0xc0002521c0 m=nil [select]:
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0xc0001d5a28?, 0x2?, 0xe0?, 0x56?, 0xc0001d57ec?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc0001d5660 sp=0xc0001d5640 pc=0x5ed0825e54ee
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.selectgo(0xc0001d5a28, 0xc0001d57e8, 0xc00024cd00?, 0x0, 0x1?, 0x1)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/select.go:327 +0x725 fp=0xc0001d5780 sp=0xc0001d5660 pc=0x5ed0825f68c5
2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).completion(0xc000214120, {0x5ed082b7a510, 0xc0001a2380}, 0xc000190480)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:698 +0xa86 fp=0xc0001d5ab8 sp=0xc0001d5780 pc=0x5ed08282a006
2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).completion-fm({0x5ed082b7a510?, 0xc0001a2380?}, 0x5ed08280438d?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: :1 +0x36 fp=0xc0001d5ae8 sp=0xc0001d5ab8 pc=0x5ed08282d2b6
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.HandlerFunc.ServeHTTP(0xc00017ab60?, {0x5ed082b7a510?, 0xc0001a2380?}, 0x10?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:2171 +0x29 fp=0xc0001d5b10 sp=0xc0001d5ae8 pc=0x5ed0827fce29
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*ServeMux).ServeHTTP(0x5ed0825b8465?, {0x5ed082b7a510, 0xc0001a2380}, 0xc000190480)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:2688 +0x1ad fp=0xc0001d5b60 sp=0xc0001d5b10 pc=0x5ed0827fecad
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.serverHandler.ServeHTTP({0x5ed082b79860?}, {0x5ed082b7a510?, 0xc0001a2380?}, 0x6?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:3142 +0x8e fp=0xc0001d5b90 sp=0xc0001d5b60 pc=0x5ed0827ffcce
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*conn).serve(0xc000298090, {0x5ed082b7a968, 0xc000178db0})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:2044 +0x5e8 fp=0xc0001d5fb8 sp=0xc0001d5b90 pc=0x5ed0827fba68
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*Server).Serve.gowrap3()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:3290 +0x28 fp=0xc0001d5fe0 sp=0xc0001d5fb8 pc=0x5ed082800448
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc0001d5fe8 sp=0xc0001d5fe0 pc=0x5ed0826172c1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by net/http.(*Server).Serve in goroutine 1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:3290 +0x4b4
2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 22 gp=0xc000252700 m=nil [IO wait]:
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0x10?, 0x10?, 0xf0?, 0x5d?, 0xb?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000135da8 sp=0xc000135d88 pc=0x5ed0825e54ee
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.netpollblock(0x5ed08264ba38?, 0x825ae006?, 0xd0?)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/netpoll.go:573 +0xf7 fp=0xc000135de0 sp=0xc000135da8 pc=0x5ed0825dd737
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.runtime_pollWait(0x744117b87f28, 0x72)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/netpoll.go:345 +0x85 fp=0xc000135e00 sp=0xc000135de0 pc=0x5ed082611f85
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*pollDesc).wait(0xc0002b4000?, 0xc0002921f1?, 0x0)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_poll_runtime.go:84 +0x27 fp=0xc000135e28 sp=0xc000135e00 pc=0x5ed082661ea7
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*pollDesc).waitRead(...)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_poll_runtime.go:89
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*FD).Read(0xc0002b4000, {0xc0002921f1, 0x1, 0x1})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_unix.go:164 +0x27a fp=0xc000135ec0 sp=0xc000135e28 pc=0x5ed0826629fa
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*netFD).Read(0xc0002b4000, {0xc0002921f1?, 0xc000135f48?, 0x5ed082613bb0?})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/fd_posix.go:55 +0x25 fp=0xc000135f08 sp=0xc000135ec0 pc=0x5ed0826d0ea5
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*conn).Read(0xc0002b6000, {0xc0002921f1?, 0x0?, 0x5ed082e32100?})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/net.go:185 +0x45 fp=0xc000135f50 sp=0xc000135f08 pc=0x5ed0826db165
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*TCPConn).Read(0x5ed082d0a890?, {0xc0002921f1?, 0x0?, 0x0?})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: :1 +0x25 fp=0xc000135f80 sp=0xc000135f50 pc=0x5ed0826e6b45
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*connReader).backgroundRead(0xc0002921e0)
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:681 +0x37 fp=0xc000135fc8 sp=0xc000135f80 pc=0x5ed0827f59d7
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*connReader).startBackgroundRead.gowrap2()
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:677 +0x25 fp=0xc000135fe0 sp=0xc000135fc8 pc=0x5ed0827f5905
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({})
2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc000135fe8 sp=0xc000135fe0 pc=0x5ed0826172c1
2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by net/http.(*connReader).startBackgroundRead in goroutine 35
2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:677 +0xba
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rax 0x0
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rbx 0x7440aa200000
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rcx 0x7440f20969fc
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rdx 0x6
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rdi 0xaf0c4
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rsi 0xaf0c8
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rbp 0xaf0c8
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rsp 0x7440aa1de3e0
2月 21 08:55:27 GBITDeepSeek ollama[698750]: r8 0x7440aa1de4b0
2月 21 08:55:27 GBITDeepSeek ollama[698750]: r9 0x7440aa1de480
2月 21 08:55:27 GBITDeepSeek ollama[698750]: r10 0x8
2月 21 08:55:27 GBITDeepSeek ollama[698750]: r11 0x246
2月 21 08:55:27 GBITDeepSeek ollama[698750]: r12 0x6
2月 21 08:55:27 GBITDeepSeek ollama[698750]: r13 0x16
2月 21 08:55:27 GBITDeepSeek ollama[698750]: r14 0x74411788cd30
2月 21 08:55:27 GBITDeepSeek ollama[698750]: r15 0x0
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rip 0x7440f20969fc
2月 21 08:55:27 GBITDeepSeek ollama[698750]: rflags 0x246
2月 21 08:55:27 GBITDeepSeek ollama[698750]: cs 0x33
2月 21 08:55:27 GBITDeepSeek ollama[698750]: fs 0x0
2月 21 08:55:27 GBITDeepSeek ollama[698750]: gs 0x0
2月 21 08:55:27 GBITDeepSeek ollama[698750]: [GIN] 2025/02/21 - 08:55:27 | 200 | 10.421581803s | 172.17.0.2 | POST "/api/chat"

@CJJ-Michael commented on GitHub (Feb 21, 2025): > [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) may help in debugging. Logs below belong to ollama server after the bug happend: 2月 21 08:52:10 GBITDeepSeek ollama[698750]: [GIN] 2025/02/21 - 08:52:10 | 200 | 3.857530363s | 127.0.0.1 | POST "/api/chat" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: [GIN] 2025/02/21 - 08:55:16 | 200 | 27.206µs | 172.17.0.2 | GET "/" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.213+08:00 level=INFO source=sched.go:507 msg="updated VRAM based on existing loaded models" gpu=GPU-99e4c40f-5e53-642f-aa9b-9c3b203cfebb library=cuda total="23.6 GiB" available="18.1 GiB" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.213+08:00 level=INFO source=sched.go:714 msg="new model will fit in available VRAM in single GPU, loading" model=/home/gbit/ollama/models/blobs/sha256-970aa74c0a90ef7482477cf803618e776e173c007bf957f635f1015bfcfef0e6 gpu=GPU-99e4c40f-5e53-642f-aa9b-9c3b203cfebb parallel=1 available=19391528960 required="809.9 MiB" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.309+08:00 level=INFO source=server.go:105 msg="system memory" total="251.5 GiB" free="238.0 GiB" free_swap="2.0 GiB" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.309+08:00 level=INFO source=memory.go:356 msg="offload to cuda" layers.requested=-1 layers.model=13 layers.offload=13 layers.split="" memory.available="[18.1 GiB]" memory.gpu_overhead="0 B" memory.required.full="809.9 MiB" memory.required.partial="809.9 MiB" memory.required.kv="24.0 MiB" memory.required.allocations="[809.9 MiB]" memory.weights.total="240.1 MiB" memory.weights.repeating="195.4 MiB" memory.weights.nonrepeating="44.7 MiB" memory.graph.full="48.0 MiB" memory.graph.partial="48.0 MiB" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.311+08:00 level=INFO source=server.go:397 msg="starting llama server" cmd="/tmp/ollama3857967831/runners/cuda_v12/ollama_llama_server --model /home/gbit/ollama/models/blobs/sha256-970aa74c0a90ef7482477cf803618e776e173c007bf957f635f1015bfcfef0e6 --ctx-size 8192 --batch-size 512 --n-gpu-layers 13 --threads 52 --parallel 1 --port 42335" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.312+08:00 level=INFO source=sched.go:449 msg="loaded runners" count=2 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.312+08:00 level=INFO source=server.go:576 msg="waiting for llama runner to start responding" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.312+08:00 level=INFO source=server.go:610 msg="waiting for server to become available" status="llm server error" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.372+08:00 level=INFO source=runner.go:941 msg="starting go runner" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.372+08:00 level=INFO source=runner.go:942 msg=system info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | RISCV_VECT = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | cgo(gcc)" threads=52 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.372+08:00 level=INFO source=.:0 msg="Server listening on 127.0.0.1:42335" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: loaded meta data with 24 key-value pairs and 112 tensors from /home/gbit/ollama/models/blobs/sha256-970aa74c0a90ef7482477cf803618e776e173c007bf957f635f1015bfcfef0e6 (version GGUF V3 (latest)) 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 0: general.architecture str = nomic-bert 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 1: general.name str = nomic-embed-text-v1.5 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 2: nomic-bert.block_count u32 = 12 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 3: nomic-bert.context_length u32 = 2048 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 4: nomic-bert.embedding_length u32 = 768 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 5: nomic-bert.feed_forward_length u32 = 3072 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 6: nomic-bert.attention.head_count u32 = 12 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 7: nomic-bert.attention.layer_norm_epsilon f32 = 0.000000 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 8: general.file_type u32 = 1 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 9: nomic-bert.attention.causal bool = false 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 10: nomic-bert.pooling_type u32 = 1 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 11: nomic-bert.rope.freq_base f32 = 1000.000000 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 12: tokenizer.ggml.token_type_count u32 = 2 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 13: tokenizer.ggml.bos_token_id u32 = 101 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 14: tokenizer.ggml.eos_token_id u32 = 102 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 15: tokenizer.ggml.model str = bert 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 16: tokenizer.ggml.tokens arr[str,30522] = ["[PAD]", "[unused0]", "[unused1]", "... 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 17: tokenizer.ggml.scores arr[f32,30522] = [-1000.000000, -1000.000000, -1000.00... 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 18: tokenizer.ggml.token_type arr[i32,30522] = [3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 19: tokenizer.ggml.unknown_token_id u32 = 100 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 20: tokenizer.ggml.seperator_token_id u32 = 102 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 22: tokenizer.ggml.cls_token_id u32 = 101 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 23: tokenizer.ggml.mask_token_id u32 = 103 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - type f32: 51 tensors 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_model_loader: - type f16: 61 tensors 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_vocab: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_vocab: special tokens cache size = 5 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_vocab: token to piece cache size = 0.2032 MB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: format = GGUF V3 (latest) 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: arch = nomic-bert 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab type = WPM 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_vocab = 30522 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_merges = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab_only = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ctx_train = 2048 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd = 768 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_layer = 12 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_head = 12 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_head_kv = 12 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_rot = 64 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_swa = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_head_k = 64 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_head_v = 64 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_gqa = 1 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_k_gqa = 768 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_v_gqa = 768 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_norm_eps = 1.0e-12 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_norm_rms_eps = 0.0e+00 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_clamp_kqv = 0.0e+00 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_max_alibi_bias = 0.0e+00 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_logit_scale = 0.0e+00 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ff = 3072 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_expert = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_expert_used = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: causal attn = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: pooling type = 1 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope type = 2 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope scaling = linear 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: freq_base_train = 1000.0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: freq_scale_train = 1 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ctx_orig_yarn = 2048 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope_finetuned = unknown 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_conv = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_inner = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_state = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_dt_rank = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_dt_b_c_rms = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: model type = 137M 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: model ftype = F16 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: model params = 136.73 M 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: model size = 260.86 MiB (16.00 BPW) 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: general.name = nomic-embed-text-v1.5 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: BOS token = 101 '[CLS]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOS token = 102 '[SEP]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: UNK token = 100 '[UNK]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: SEP token = 102 '[SEP]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: PAD token = 0 '[PAD]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: CLS token = 101 '[CLS]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: MASK token = 103 '[MASK]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: LF token = 0 '[PAD]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOG token = 102 '[SEP]' 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_print_meta: max token length = 21 2月 21 08:55:16 GBITDeepSeek ollama[698750]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no 2月 21 08:55:16 GBITDeepSeek ollama[698750]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no 2月 21 08:55:16 GBITDeepSeek ollama[698750]: ggml_cuda_init: found 1 CUDA devices: 2月 21 08:55:16 GBITDeepSeek ollama[698750]: Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: ggml ctx size = 0.10 MiB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: offloading 12 repeating layers to GPU 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: offloading non-repeating layers to GPU 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: offloaded 13/13 layers to GPU 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: CPU buffer size = 44.72 MiB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llm_load_tensors: CUDA0 buffer size = 216.15 MiB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.565+08:00 level=INFO source=server.go:610 msg="waiting for server to become available" status="llm server loading model" 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_ctx = 8192 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_batch = 512 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_ubatch = 512 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: flash_attn = 0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: freq_base = 1000.0 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: freq_scale = 1 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_kv_cache_init: CUDA0 KV buffer size = 288.00 MiB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: KV self size = 288.00 MiB, K (f16): 144.00 MiB, V (f16): 144.00 MiB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CPU output buffer size = 0.00 MiB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA0 compute buffer size = 23.00 MiB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA_Host compute buffer size = 3.50 MiB 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: graph nodes = 453 2月 21 08:55:16 GBITDeepSeek ollama[698750]: llama_new_context_with_model: graph splits = 2 2月 21 08:55:16 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:16.816+08:00 level=INFO source=server.go:615 msg="llama runner started in 0.50 seconds" 2月 21 08:55:17 GBITDeepSeek ollama[698750]: [GIN] 2025/02/21 - 08:55:17 | 200 | 919.068568ms | 172.17.0.2 | POST "/api/embeddings" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.487+08:00 level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.095851007 model=/home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.701+08:00 level=INFO source=sched.go:507 msg="updated VRAM based on existing loaded models" gpu=GPU-99e4c40f-5e53-642f-aa9b-9c3b203cfebb library=cuda total="23.6 GiB" available="21.4 GiB" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.701+08:00 level=INFO source=sched.go:714 msg="new model will fit in available VRAM in single GPU, loading" model=/home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 gpu=GPU-99e4c40f-5e53-642f-aa9b-9c3b203cfebb parallel=4 available=22992453632 required="6.5 GiB" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.737+08:00 level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.345946677 model=/home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.789+08:00 level=INFO source=server.go:105 msg="system memory" total="251.5 GiB" free="237.6 GiB" free_swap="2.0 GiB" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.789+08:00 level=INFO source=memory.go:356 msg="offload to cuda" layers.requested=-1 layers.model=29 layers.offload=29 layers.split="" memory.available="[21.4 GiB]" memory.gpu_overhead="0 B" memory.required.full="6.5 GiB" memory.required.partial="6.5 GiB" memory.required.kv="896.0 MiB" memory.required.allocations="[6.5 GiB]" memory.weights.total="4.5 GiB" memory.weights.repeating="4.1 GiB" memory.weights.nonrepeating="426.4 MiB" memory.graph.full="942.0 MiB" memory.graph.partial="1.1 GiB" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.791+08:00 level=INFO source=server.go:397 msg="starting llama server" cmd="/tmp/ollama3857967831/runners/cuda_v12/ollama_llama_server --model /home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 --ctx-size 16384 --batch-size 512 --n-gpu-layers 29 --threads 52 --mlock --parallel 4 --port 35995" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.792+08:00 level=INFO source=sched.go:449 msg="loaded runners" count=2 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.792+08:00 level=INFO source=server.go:576 msg="waiting for llama runner to start responding" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.792+08:00 level=INFO source=server.go:610 msg="waiting for server to become available" status="llm server error" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.849+08:00 level=INFO source=runner.go:941 msg="starting go runner" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.850+08:00 level=INFO source=runner.go:942 msg=system info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | RISCV_VECT = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | cgo(gcc)" threads=52 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.850+08:00 level=INFO source=.:0 msg="Server listening on 127.0.0.1:35995" 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: loaded meta data with 26 key-value pairs and 339 tensors from /home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 (version GGUF V3 (latest)) 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 0: general.architecture str = qwen2 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 1: general.type str = model 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 2: general.name str = DeepSeek R1 Distill Qwen 7B 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 3: general.basename str = DeepSeek-R1-Distill-Qwen 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 4: general.size_label str = 7B 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 5: qwen2.block_count u32 = 28 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 6: qwen2.context_length u32 = 131072 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 7: qwen2.embedding_length u32 = 3584 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 8: qwen2.feed_forward_length u32 = 18944 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 9: qwen2.attention.head_count u32 = 28 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 10: qwen2.attention.head_count_kv u32 = 4 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 11: qwen2.rope.freq_base f32 = 10000.000000 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 12: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 13: general.file_type u32 = 15 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 14: tokenizer.ggml.model str = gpt2 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 15: tokenizer.ggml.pre str = qwen2 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 16: tokenizer.ggml.tokens arr[str,152064] = ["!", "\"", "#", "$", "%", "&", "'", ... 2月 21 08:55:22 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 17: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... 2月 21 08:55:22 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:22.986+08:00 level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.59510752 model=/home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 18: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 19: tokenizer.ggml.bos_token_id u32 = 151646 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 151643 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 151643 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 22: tokenizer.ggml.add_bos_token bool = true 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 23: tokenizer.ggml.add_eos_token bool = false 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 24: tokenizer.chat_template str = {% if not add_generation_prompt is de... 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 25: general.quantization_version u32 = 2 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - type f32: 141 tensors 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - type q4_K: 169 tensors 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llama_model_loader: - type q6_K: 29 tensors 2月 21 08:55:23 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:23.044+08:00 level=INFO source=server.go:610 msg="waiting for server to become available" status="llm server loading model" 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_vocab: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_vocab: special tokens cache size = 22 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_vocab: token to piece cache size = 0.9310 MB 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: format = GGUF V3 (latest) 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: arch = qwen2 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab type = BPE 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_vocab = 152064 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_merges = 151387 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab_only = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ctx_train = 131072 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd = 3584 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_layer = 28 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_head = 28 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_head_kv = 4 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_rot = 128 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_swa = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_head_k = 128 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_head_v = 128 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_gqa = 7 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_k_gqa = 512 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_embd_v_gqa = 512 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_norm_eps = 0.0e+00 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_norm_rms_eps = 1.0e-06 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_clamp_kqv = 0.0e+00 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_max_alibi_bias = 0.0e+00 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: f_logit_scale = 0.0e+00 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ff = 18944 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_expert = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_expert_used = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: causal attn = 1 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: pooling type = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope type = 2 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope scaling = linear 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: freq_base_train = 10000.0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: freq_scale_train = 1 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_ctx_orig_yarn = 131072 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: rope_finetuned = unknown 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_conv = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_inner = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_d_state = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_dt_rank = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: ssm_dt_b_c_rms = 0 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: model type = ?B 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: model ftype = Q4_K - Medium 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: model params = 7.62 B 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: model size = 4.36 GiB (4.91 BPW) 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: general.name = DeepSeek R1 Distill Qwen 7B 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: BOS token = 151646 '<｜begin▁of▁sentence｜>' 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOS token = 151643 '<｜end▁of▁sentence｜>' 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: PAD token = 151643 '<｜end▁of▁sentence｜>' 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: LF token = 148848 'ÄĬ' 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOG token = 151643 '<｜end▁of▁sentence｜>' 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_print_meta: max token length = 256 2月 21 08:55:23 GBITDeepSeek ollama[698750]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no 2月 21 08:55:23 GBITDeepSeek ollama[698750]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no 2月 21 08:55:23 GBITDeepSeek ollama[698750]: ggml_cuda_init: found 1 CUDA devices: 2月 21 08:55:23 GBITDeepSeek ollama[698750]: Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: ggml ctx size = 0.30 MiB 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: offloading 28 repeating layers to GPU 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: offloading non-repeating layers to GPU 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: offloaded 29/29 layers to GPU 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: CPU buffer size = 292.36 MiB 2月 21 08:55:23 GBITDeepSeek ollama[698750]: llm_load_tensors: CUDA0 buffer size = 4168.09 MiB 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_ctx = 16384 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_batch = 2048 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: n_ubatch = 512 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: flash_attn = 0 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: freq_base = 10000.0 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: freq_scale = 1 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_kv_cache_init: CUDA0 KV buffer size = 896.00 MiB 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: KV self size = 896.00 MiB, K (f16): 448.00 MiB, V (f16): 448.00 MiB 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA_Host output buffer size = 2.38 MiB 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA0 compute buffer size = 956.00 MiB 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: CUDA_Host compute buffer size = 39.01 MiB 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: graph nodes = 986 2月 21 08:55:24 GBITDeepSeek ollama[698750]: llama_new_context_with_model: graph splits = 2 2月 21 08:55:25 GBITDeepSeek ollama[698750]: time=2025-02-21T08:55:25.055+08:00 level=INFO source=server.go:615 msg="llama runner started in 2.26 seconds" 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: loaded meta data with 26 key-value pairs and 339 tensors from /home/gbit/ollama/models/blobs/sha256-96c415656d377afbff962f6cdb2394ab092ccbcbaab4b82525bc4ca800fe8a49 (version GGUF V3 (latest)) 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 0: general.architecture str = qwen2 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 1: general.type str = model 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 2: general.name str = DeepSeek R1 Distill Qwen 7B 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 3: general.basename str = DeepSeek-R1-Distill-Qwen 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 4: general.size_label str = 7B 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 5: qwen2.block_count u32 = 28 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 6: qwen2.context_length u32 = 131072 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 7: qwen2.embedding_length u32 = 3584 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 8: qwen2.feed_forward_length u32 = 18944 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 9: qwen2.attention.head_count u32 = 28 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 10: qwen2.attention.head_count_kv u32 = 4 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 11: qwen2.rope.freq_base f32 = 10000.000000 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 12: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 13: general.file_type u32 = 15 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 14: tokenizer.ggml.model str = gpt2 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 15: tokenizer.ggml.pre str = qwen2 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 16: tokenizer.ggml.tokens arr[str,152064] = ["!", "\"", "#", "$", "%", "&", "'", ... 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 17: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 18: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 19: tokenizer.ggml.bos_token_id u32 = 151646 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 151643 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 151643 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 22: tokenizer.ggml.add_bos_token bool = true 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 23: tokenizer.ggml.add_eos_token bool = false 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 24: tokenizer.chat_template str = {% if not add_generation_prompt is de... 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - kv 25: general.quantization_version u32 = 2 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - type f32: 141 tensors 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - type q4_K: 169 tensors 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_loader: - type q6_K: 29 tensors 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_vocab: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_vocab: special tokens cache size = 22 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_vocab: token to piece cache size = 0.9310 MB 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: format = GGUF V3 (latest) 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: arch = qwen2 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab type = BPE 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_vocab = 152064 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: n_merges = 151387 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: vocab_only = 1 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: model type = ?B 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: model ftype = all F32 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: model params = 7.62 B 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: model size = 4.36 GiB (4.91 BPW) 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: general.name = DeepSeek R1 Distill Qwen 7B 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: BOS token = 151646 '<｜begin▁of▁sentence｜>' 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOS token = 151643 '<｜end▁of▁sentence｜>' 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: PAD token = 151643 '<｜end▁of▁sentence｜>' 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: LF token = 148848 'ÄĬ' 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: EOG token = 151643 '<｜end▁of▁sentence｜>' 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llm_load_print_meta: max token length = 256 2月 21 08:55:25 GBITDeepSeek ollama[698750]: llama_model_load: vocab only - skipping tensors 2月 21 08:55:27 GBITDeepSeek ollama[698750]: CUDA error: misaligned address 2月 21 08:55:27 GBITDeepSeek ollama[698750]: current device: 0, in function ggml_backend_cuda_synchronize at ggml-cuda.cu:2508 2月 21 08:55:27 GBITDeepSeek ollama[698750]: cudaStreamSynchronize(cuda_ctx->stream()) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: ggml-cuda.cu:132: CUDA error 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 716998] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 716999] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717000] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717001] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717002] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717009] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717022] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717023] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [New LWP 717029] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [Thread debugging using libthread_db enabled] 2月 21 08:55:27 GBITDeepSeek ollama[717146]: Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". 2月 21 08:55:27 GBITDeepSeek ollama[717146]: 0x00005ed082619083 in ?? () 2月 21 08:55:27 GBITDeepSeek ollama[717146]: #0 0x00005ed082619083 in ?? () 2月 21 08:55:27 GBITDeepSeek ollama[717146]: #1 0x00005ed0825de3d0 in _start () 2月 21 08:55:27 GBITDeepSeek ollama[717146]: [Inferior 1 (process 716996) detached] 2月 21 08:55:27 GBITDeepSeek ollama[698750]: SIGABRT: abort 2月 21 08:55:27 GBITDeepSeek ollama[698750]: PC=0x7440f20969fc m=4 sigcode=18446744073709551610 2月 21 08:55:27 GBITDeepSeek ollama[698750]: signal arrived during cgo execution 2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 7 gp=0xc000252000 m=4 mp=0xc000139808 [syscall]: 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.cgocall(0x5ed08282d8c0, 0xc000147c60) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/cgocall.go:157 +0x4b fp=0xc000147c38 sp=0xc000147c00 pc=0x5ed0825ae8ab 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama._Cfunc_gpt_sampler_csample(0x744098008400, 0x74409c006450, 0x0) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: _cgo_gotypes.go:468 +0x4f fp=0xc000147c60 sp=0xc000147c38 pc=0x5ed0826ab98f 2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).processBatch.(*SamplingContext).Sample.func4(0x5ed082e32100?, 0x0?, 0x0) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/llama.go:699 +0x86 fp=0xc000147cb0 sp=0xc000147c60 pc=0x5ed082829346 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama.(*SamplingContext).Sample(...) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/llama.go:699 2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).processBatch(0xc000214120, 0xc0002ae000, 0xc000147f10) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:487 +0x6ab fp=0xc000147ed0 sp=0xc000147cb0 pc=0x5ed08282860b 2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).run(0xc000214120, {0x5ed082b7a9a0, 0xc00016a0a0}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:342 +0x1e5 fp=0xc000147fb8 sp=0xc000147ed0 pc=0x5ed082827c25 2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.main.gowrap2() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:980 +0x28 fp=0xc000147fe0 sp=0xc000147fb8 pc=0x5ed08282ca88 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc000147fe8 sp=0xc000147fe0 pc=0x5ed0826172c1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by main.main in goroutine 1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:980 +0xd3e 2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 1 gp=0xc0000061c0 m=nil [IO wait]: 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0x1?, 0xc00002b8e0?, 0xd4?, 0x52?, 0xc00002b8c0?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc00002b860 sp=0xc00002b840 pc=0x5ed0825e54ee 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.netpollblock(0x10?, 0x825ae006?, 0xd0?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/netpoll.go:573 +0xf7 fp=0xc00002b898 sp=0xc00002b860 pc=0x5ed0825dd737 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.runtime_pollWait(0x744117b88020, 0x72) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/netpoll.go:345 +0x85 fp=0xc00002b8b8 sp=0xc00002b898 pc=0x5ed082611f85 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*pollDesc).wait(0x3?, 0x744117bc3e88?, 0x0) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_poll_runtime.go:84 +0x27 fp=0xc00002b8e0 sp=0xc00002b8b8 pc=0x5ed082661ea7 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*pollDesc).waitRead(...) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_poll_runtime.go:89 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*FD).Accept(0xc00024c080) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_unix.go:611 +0x2ac fp=0xc00002b988 sp=0xc00002b8e0 pc=0x5ed08266336c 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*netFD).accept(0xc00024c080) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/fd_unix.go:172 +0x29 fp=0xc00002ba40 sp=0xc00002b988 pc=0x5ed0826d1fa9 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*TCPListener).accept(0xc0001521e0) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/tcpsock_posix.go:159 +0x1e fp=0xc00002ba68 sp=0xc00002ba40 pc=0x5ed0826e2cde 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*TCPListener).Accept(0xc0001521e0) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/tcpsock.go:327 +0x30 fp=0xc00002ba98 sp=0xc00002ba68 pc=0x5ed0826e2030 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*onceCloseListener).Accept(0xc000298090?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: <autogenerated>:1 +0x24 fp=0xc00002bab0 sp=0xc00002ba98 pc=0x5ed082809244 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*Server).Serve(0xc0000181e0, {0x5ed082b7a360, 0xc0001521e0}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:3260 +0x33e fp=0xc00002bbe0 sp=0xc00002bab0 pc=0x5ed08280005e 2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.main() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:1000 +0x10cd fp=0xc00002bf50 sp=0xc00002bbe0 pc=0x5ed08282c80d 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.main() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:271 +0x29d fp=0xc00002bfe0 sp=0xc00002bf50 pc=0x5ed0825e50bd 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc00002bfe8 sp=0xc00002bfe0 pc=0x5ed0826172c1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 2 gp=0xc000006c40 m=nil [force gc (idle)]: 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000132fa8 sp=0xc000132f88 pc=0x5ed0825e54ee 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goparkunlock(...) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:408 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.forcegchelper() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:326 +0xb8 fp=0xc000132fe0 sp=0xc000132fa8 pc=0x5ed0825e5378 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc000132fe8 sp=0xc000132fe0 pc=0x5ed0826172c1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by runtime.init.6 in goroutine 1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:314 +0x1a 2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 3 gp=0xc000007180 m=nil [GC sweep wait]: 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000133780 sp=0xc000133760 pc=0x5ed0825e54ee 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goparkunlock(...) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:408 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.bgsweep(0xc00015c000) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgcsweep.go:278 +0x94 fp=0xc0001337c8 sp=0xc000133780 pc=0x5ed0825d0034 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gcenable.gowrap1() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgc.go:203 +0x25 fp=0xc0001337e0 sp=0xc0001337c8 pc=0x5ed0825c4b65 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc0001337e8 sp=0xc0001337e0 pc=0x5ed0826172c1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by runtime.gcenable in goroutine 1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgc.go:203 +0x66 2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 4 gp=0xc000007340 m=nil [GC scavenge wait]: 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0xc00015c000?, 0x5ed082a774f0?, 0x1?, 0x0?, 0xc000007340?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000133f78 sp=0xc000133f58 pc=0x5ed0825e54ee 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goparkunlock(...) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:408 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.(*scavengerState).park(0x5ed082d49560) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgcscavenge.go:425 +0x49 fp=0xc000133fa8 sp=0xc000133f78 pc=0x5ed0825cda29 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.bgscavenge(0xc00015c000) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgcscavenge.go:653 +0x3c fp=0xc000133fc8 sp=0xc000133fa8 pc=0x5ed0825cdfbc 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gcenable.gowrap2() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgc.go:204 +0x25 fp=0xc000133fe0 sp=0xc000133fc8 pc=0x5ed0825c4b05 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc000133fe8 sp=0xc000133fe0 pc=0x5ed0826172c1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by runtime.gcenable in goroutine 1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mgc.go:204 +0xa5 2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 5 gp=0xc000007c00 m=nil [finalizer wait]: 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0xc000132648?, 0x5ed0825b8465?, 0xa8?, 0x1?, 0xc0000061c0?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000132620 sp=0xc000132600 pc=0x5ed0825e54ee 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.runfinq() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mfinal.go:194 +0x107 fp=0xc0001327e0 sp=0xc000132620 pc=0x5ed0825c3ba7 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc0001327e8 sp=0xc0001327e0 pc=0x5ed0826172c1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by runtime.createfing in goroutine 1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/mfinal.go:164 +0x3d 2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 35 gp=0xc0002521c0 m=nil [select]: 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0xc0001d5a28?, 0x2?, 0xe0?, 0x56?, 0xc0001d57ec?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc0001d5660 sp=0xc0001d5640 pc=0x5ed0825e54ee 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.selectgo(0xc0001d5a28, 0xc0001d57e8, 0xc00024cd00?, 0x0, 0x1?, 0x1) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/select.go:327 +0x725 fp=0xc0001d5780 sp=0xc0001d5660 pc=0x5ed0825f68c5 2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).completion(0xc000214120, {0x5ed082b7a510, 0xc0001a2380}, 0xc000190480) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: github.com/ollama/ollama/llama/runner/runner.go:698 +0xa86 fp=0xc0001d5ab8 sp=0xc0001d5780 pc=0x5ed08282a006 2月 21 08:55:27 GBITDeepSeek ollama[698750]: main.(*Server).completion-fm({0x5ed082b7a510?, 0xc0001a2380?}, 0x5ed08280438d?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: <autogenerated>:1 +0x36 fp=0xc0001d5ae8 sp=0xc0001d5ab8 pc=0x5ed08282d2b6 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.HandlerFunc.ServeHTTP(0xc00017ab60?, {0x5ed082b7a510?, 0xc0001a2380?}, 0x10?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:2171 +0x29 fp=0xc0001d5b10 sp=0xc0001d5ae8 pc=0x5ed0827fce29 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*ServeMux).ServeHTTP(0x5ed0825b8465?, {0x5ed082b7a510, 0xc0001a2380}, 0xc000190480) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:2688 +0x1ad fp=0xc0001d5b60 sp=0xc0001d5b10 pc=0x5ed0827fecad 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.serverHandler.ServeHTTP({0x5ed082b79860?}, {0x5ed082b7a510?, 0xc0001a2380?}, 0x6?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:3142 +0x8e fp=0xc0001d5b90 sp=0xc0001d5b60 pc=0x5ed0827ffcce 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*conn).serve(0xc000298090, {0x5ed082b7a968, 0xc000178db0}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:2044 +0x5e8 fp=0xc0001d5fb8 sp=0xc0001d5b90 pc=0x5ed0827fba68 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*Server).Serve.gowrap3() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:3290 +0x28 fp=0xc0001d5fe0 sp=0xc0001d5fb8 pc=0x5ed082800448 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc0001d5fe8 sp=0xc0001d5fe0 pc=0x5ed0826172c1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by net/http.(*Server).Serve in goroutine 1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:3290 +0x4b4 2月 21 08:55:27 GBITDeepSeek ollama[698750]: goroutine 22 gp=0xc000252700 m=nil [IO wait]: 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.gopark(0x10?, 0x10?, 0xf0?, 0x5d?, 0xb?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/proc.go:402 +0xce fp=0xc000135da8 sp=0xc000135d88 pc=0x5ed0825e54ee 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.netpollblock(0x5ed08264ba38?, 0x825ae006?, 0xd0?) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/netpoll.go:573 +0xf7 fp=0xc000135de0 sp=0xc000135da8 pc=0x5ed0825dd737 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.runtime_pollWait(0x744117b87f28, 0x72) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/netpoll.go:345 +0x85 fp=0xc000135e00 sp=0xc000135de0 pc=0x5ed082611f85 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*pollDesc).wait(0xc0002b4000?, 0xc0002921f1?, 0x0) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_poll_runtime.go:84 +0x27 fp=0xc000135e28 sp=0xc000135e00 pc=0x5ed082661ea7 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*pollDesc).waitRead(...) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_poll_runtime.go:89 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll.(*FD).Read(0xc0002b4000, {0xc0002921f1, 0x1, 0x1}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: internal/poll/fd_unix.go:164 +0x27a fp=0xc000135ec0 sp=0xc000135e28 pc=0x5ed0826629fa 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*netFD).Read(0xc0002b4000, {0xc0002921f1?, 0xc000135f48?, 0x5ed082613bb0?}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/fd_posix.go:55 +0x25 fp=0xc000135f08 sp=0xc000135ec0 pc=0x5ed0826d0ea5 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*conn).Read(0xc0002b6000, {0xc0002921f1?, 0x0?, 0x5ed082e32100?}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/net.go:185 +0x45 fp=0xc000135f50 sp=0xc000135f08 pc=0x5ed0826db165 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net.(*TCPConn).Read(0x5ed082d0a890?, {0xc0002921f1?, 0x0?, 0x0?}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: <autogenerated>:1 +0x25 fp=0xc000135f80 sp=0xc000135f50 pc=0x5ed0826e6b45 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*connReader).backgroundRead(0xc0002921e0) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:681 +0x37 fp=0xc000135fc8 sp=0xc000135f80 pc=0x5ed0827f59d7 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http.(*connReader).startBackgroundRead.gowrap2() 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:677 +0x25 fp=0xc000135fe0 sp=0xc000135fc8 pc=0x5ed0827f5905 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime.goexit({}) 2月 21 08:55:27 GBITDeepSeek ollama[698750]: runtime/asm_amd64.s:1695 +0x1 fp=0xc000135fe8 sp=0xc000135fe0 pc=0x5ed0826172c1 2月 21 08:55:27 GBITDeepSeek ollama[698750]: created by net/http.(*connReader).startBackgroundRead in goroutine 35 2月 21 08:55:27 GBITDeepSeek ollama[698750]: net/http/server.go:677 +0xba 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rax 0x0 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rbx 0x7440aa200000 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rcx 0x7440f20969fc 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rdx 0x6 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rdi 0xaf0c4 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rsi 0xaf0c8 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rbp 0xaf0c8 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rsp 0x7440aa1de3e0 2月 21 08:55:27 GBITDeepSeek ollama[698750]: r8 0x7440aa1de4b0 2月 21 08:55:27 GBITDeepSeek ollama[698750]: r9 0x7440aa1de480 2月 21 08:55:27 GBITDeepSeek ollama[698750]: r10 0x8 2月 21 08:55:27 GBITDeepSeek ollama[698750]: r11 0x246 2月 21 08:55:27 GBITDeepSeek ollama[698750]: r12 0x6 2月 21 08:55:27 GBITDeepSeek ollama[698750]: r13 0x16 2月 21 08:55:27 GBITDeepSeek ollama[698750]: r14 0x74411788cd30 2月 21 08:55:27 GBITDeepSeek ollama[698750]: r15 0x0 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rip 0x7440f20969fc 2月 21 08:55:27 GBITDeepSeek ollama[698750]: rflags 0x246 2月 21 08:55:27 GBITDeepSeek ollama[698750]: cs 0x33 2月 21 08:55:27 GBITDeepSeek ollama[698750]: fs 0x0 2月 21 08:55:27 GBITDeepSeek ollama[698750]: gs 0x0 2月 21 08:55:27 GBITDeepSeek ollama[698750]: [GIN] 2025/02/21 - 08:55:27 | 200 | 10.421581803s | 172.17.0.2 | POST "/api/chat"
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
client2
cloud
compatibility
context-length
create
docker
documentation
embeddings
engine
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
install
integration
intel
js
linux
macos
memory
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama-ollama#6030
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?