time=2026-10-05T09:30:03.736+02:00 level=INFO source=routes.go:2117 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: LLAMA_ARG_FIT: LLAMA_ARG_FIT_TARGET: NO_PROXY: OLLAMA_CONTEXT_LENGTH:0 OLLAMA_CREATE_REMOTE:false OLLAMA_DEBUG:INFO OLLAMA_DEBUG_LOG_REQUESTS:false OLLAMA_EDITOR: OLLAMA_FLASH_ATTENTION:false OLLAMA_GO_TEMPLATE:true OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_IGPU_ENABLE: OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MAX_TRANSFER_STREAMS:4 OLLAMA_MODELS:C:\\Users\\jerom\\.ollama\\models OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NO_CLOUD:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false OLLAMA_VULKAN:true ROCR_VISIBLE_DEVICES:]" time=2026-10-05T09:30:03.781+02:00 level=INFO source=routes.go:2119 msg="Ollama cloud disabled: false" time=2026-10-05T09:30:03.785+02:00 level=INFO source=images.go:947 msg="total blobs: 8" time=2026-10-05T09:30:03.786+02:00 level=INFO source=images.go:955 msg="total unused blobs removed: 0" time=2026-10-05T09:30:03.790+02:00 level=INFO source=routes.go:2177 msg="Listening on 127.0.0.1:11434 (version 0.35.1)" time=2026-10-05T09:30:03.796+02:00 level=INFO source=runner.go:60 msg="discovering available GPUs..." time=2026-10-05T09:30:03.994+02:00 level=INFO source=model_recommendations.go:177 msg="model recommendations cache sleep scheduled" wait=4h21m45.61569341s consecutive_failures=0 time=2026-10-05T09:30:12.974+02:00 level=INFO source=runner.go:405 msg="dropping integrated GPU; to enable, set OLLAMA_IGPU_ENABLE=1" id=0 library=Vulkan compute=0.0 name=Vulkan0 description="Intel(R) UHD Graphics 770" pci_id="" time=2026-10-05T09:30:14.050+02:00 level=INFO source=types.go:32 msg="inference compute" id=0 filter_id=0 library=CUDA compute=8.6 name=CUDA0 description="NVIDIA GeForce RTX 3080" libdirs=ollama,cuda_v13 driver=13.4 pci_id=0000:01:00.0 type=discrete total="10.0 GiB" available="8.9 GiB" time=2026-10-05T09:30:14.050+02:00 level=INFO source=routes.go:2227 msg="vram-based default context" total_vram="10.0 GiB" default_num_ctx=4096 [GIN] 2026/10/05 - 09:30:14 | 500 | 2.5009ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:30:14 | 500 | 3.0004ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:30:14 | 500 | 3.0004ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:30:14 | 500 | 5.3519ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:30:14 | 500 | 6.5092ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:30:14 | 200 | 9.0071ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:30:14 | 200 | 3.4993ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:30:14 | 200 | 57.7955ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/10/05 - 09:30:17 | 200 | 6.0158ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:30:47 | 200 | 5.5113ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:31:17 | 200 | 5.4419ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:31:47 | 200 | 3.9959ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:32:17 | 200 | 10.536ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:32:47 | 200 | 4.0916ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:32:48 | 200 | 3.6425ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:32:48 | 200 | 5.0084ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/10/05 - 09:32:48 | 200 | 52.5524ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/10/05 - 09:32:48 | 200 | 4.5145ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:32:48 | 200 | 2.5924ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/10/05 - 09:32:48 | 200 | 3.5007ms | 127.0.0.1 | POST "/api/show" time=2026-10-05T09:32:49.671+02:00 level=INFO source=sched.go:1152 msg="disabling mmap for llama-server load by default" model=C:\Users\jerom\.ollama\models\blobs\sha256-1194192cf2a187eb02722edcc3f77b11d21f537048ce04b67ccf8ba78863006a reason=windows_cuda time=2026-10-05T09:32:49.671+02:00 level=INFO source=server.go:100 msg="using llama-server for model" model=C:\Users\jerom\.ollama\models\blobs\sha256-1194192cf2a187eb02722edcc3f77b11d21f537048ce04b67ccf8ba78863006a time=2026-10-05T09:32:49.691+02:00 level=INFO source=llama_server.go:436 msg="starting llama-server" cmd="C:\\Users\\jerom\\AppData\\Local\\Programs\\Ollama\\lib\\ollama\\llama-server.exe --model C:\\Users\\jerom\\.ollama\\models\\blobs\\sha256-1194192cf2a187eb02722edcc3f77b11d21f537048ce04b67ccf8ba78863006a --port 53541 --host 127.0.0.1 --no-webui --offline -c 16384 -np 1 --log-verbosity 4 --no-log-prefix --no-log-timestamps --no-jinja --chat-template chatml --load-mode none --flash-attn auto -b 512 -ub 512 --context-shift --keep 4" time=2026-10-05T09:32:49.699+02:00 level=INFO source=sched.go:618 msg="system memory" total="31.7 GiB" free="4.1 GiB" free_swap="2.9 GiB" time=2026-10-05T09:32:49.699+02:00 level=INFO source=sched.go:625 msg="gpu memory" id=0 library=CUDA available="8.4 GiB" free="8.9 GiB" minimum="457.0 MiB" overhead="0 B" time=2026-10-05T09:32:49.699+02:00 level=INFO source=llama_server.go:1054 msg="loading model via llama-server" model=C:\Users\jerom\.ollama\models\blobs\sha256-1194192cf2a187eb02722edcc3f77b11d21f537048ce04b67ccf8ba78863006a time=2026-10-05T09:32:49.699+02:00 level=INFO source=llama_server.go:1307 msg="waiting for llama-server to start responding" time=2026-10-05T09:32:49.700+02:00 level=INFO source=llama_server.go:1362 msg="waiting for llama-server to become available" status="llm server error" srv llama_server: initializing ... load_backend: loaded CPU backend from C:\Users\jerom\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-alderlake.dll ggml_cuda_init: found 1 CUDA devices (Total VRAM: 10239 MiB): Device 0: NVIDIA GeForce RTX 3080, compute capability 8.6, VMM: yes, VRAM: 10239 MiB load_backend: loaded CUDA backend from C:\Users\jerom\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll cmn common_param: common_params_print_info: build 1 (6f767fe96) with Clang 18.1.8 for Windows AMD64 cmn common_param: common_params_print_info: verbosity = 4 (adjust with the `-lv N` CLI arg) cmn common_param: device_info: cmn common_param: - CPU : 12th Gen Intel(R) Core(TM) i7-12700K (32489 MiB, 3994 MiB free) cmn common_param: - CUDA0 : NVIDIA GeForce RTX 3080 (10239 MiB, 9095 MiB free) cmn common_param: system_info: n_threads = 10 (n_threads_batch = 10) / 20 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX_VNNI = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | LLAMAFILE = 1 | OPENMP = 1 | REPACK = 1 | CUDA : ARCHS = 750,800,860,890,1000,1200 | USE_GRAPHS = 1 | FA_QUANTS = q4_0-q4_0,q8_0-q8_0,f16-f16,bf16-bf16 | srv init_listene: The UI is disabled srv init_listene: Use --ui/--no-ui (or deprecated --webui/--no-webui) to enable/disable srv init: using 19 threads for HTTP server srv llama_server: security: no API key is set and CORS allows all origins (see https://github.com/ggml-org/llama.cpp/pull/25655) srv load_model: loading model 'C:\Users\jerom\.ollama\models\blobs\sha256-1194192cf2a187eb02722edcc3f77b11d21f537048ce04b67ccf8ba78863006a' srv load_model: local path 'C:\Users\jerom\.ollama\models\blobs\sha256-1194192cf2a187eb02722edcc3f77b11d21f537048ce04b67ccf8ba78863006a' cmn common_init_: fitting params to device memory ... cmn common_init_: (for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on) common_params_fit_impl: getting device memory data for initial parameters: time=2026-10-05T09:32:50.208+02:00 level=INFO source=llama_server.go:1362 msg="waiting for llama-server to become available" status="llm server loading model" common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + (19160 = 17524 + 1536 + 100) + -18015 | common_memory_breakdown_print: | - Host | 190 = 166 + 0 + 24 | common_params_fit_impl: projected to use 19160 MiB of device memory vs. 9095 MiB of free device memory common_params_fit_impl: cannot meet free memory target of 1024 MiB, need to reduce device memory by 11089 MiB common_params_fit_impl: context size set by user to 16384 -> no change common_params_fit_impl: getting device memory data with all MoE tensors moved to system memory: common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + ( 2553 = 784 + 1536 + 233) + -1409 | common_memory_breakdown_print: | - Host | 16930 = 16906 + 0 + 24 | common_params_fit_impl: with only dense weights in device memory there is a total surplus of 5517 MiB common_params_fit_impl: id=0, target=8071 MiB common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + ( 266 = 0 + 0 + 266) + 878 | common_memory_breakdown_print: | - Host | 19251 = 17691 + 1536 + 24 | common_params_fit_impl: memory for test allocation by device: common_params_fit_impl: id=0, n_layer= 0, n_part= 0, overflow_type=4, mem= 266 MiB common_params_fit_impl: filling dense-only layers back-to-front: common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + ( 2920 = 1157 + 1536 + 226) + -1775 | common_memory_breakdown_print: | - Host | 16557 = 16533 + 0 + 24 | common_params_fit_impl: memory for test allocation by device: common_params_fit_impl: id=0, n_layer=49, n_part=48, overflow_type=4, mem= 2920 MiB common_params_fit_impl: set ngl_per_device[0].n_layer=49 common_params_fit_impl: - CUDA0 (NVIDIA GeForce RTX 3080): 49 layers, 2920 MiB used, 6174 MiB free common_params_fit_impl: converting dense-only layers to full layers and filling them front-to-back with overflow to next device/system memory: common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + (19160 = 17524 + 1536 + 100) + -18015 | common_memory_breakdown_print: | - Host | 190 = 166 + 0 + 24 | common_params_fit_impl: memory for test allocation by device: common_params_fit_impl: id=0, n_layer=49, n_part= 0, overflow_type=4, mem= 19160 MiB common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + ( 8176 = 6413 + 1536 + 226) + -7031 | common_memory_breakdown_print: | - Host | 11301 = 11277 + 0 + 24 | common_params_fit_impl: memory for test allocation by device: common_params_fit_impl: id=0, n_layer=49, n_part=33, overflow_type=4, mem= 8176 MiB common_params_fit_impl: set ngl_per_device_high[0].(n_layer, n_part)=(49, 33), id_dense_start_high=0 common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + ( 7852 = 6089 + 1536 + 226) + -6707 | common_memory_breakdown_print: | - Host | 11625 = 11601 + 0 + 24 | common_params_fit_impl: memory for test allocation by device: common_params_fit_impl: id=0, n_layer=49, n_part=34, overflow_type=4, mem= 7852 MiB common_params_fit_impl: set ngl_per_device[0].(n_layer, n_part)=(49, 34), id_dense_start=0 common_params_fit_impl: trying to fit one extra layer with overflow_type=LAYER_FRACTION_UP common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + ( 7959 = 6196 + 1536 + 226) + -6814 | common_memory_breakdown_print: | - Host | 11518 = 11494 + 0 + 24 | common_params_fit_impl: memory for test allocation by device: common_params_fit_impl: id=0, n_layer=49, n_part=34, overflow_type=2, mem= 7959 MiB common_params_fit_impl: set ngl_per_device[0].(n_layer, n_part, overflow_type)=(49, 34, UP), id_dense_start=0 common_params_fit_impl: trying to fit one extra layer with overflow_type=LAYER_FRACTION_GATE common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted | common_memory_breakdown_print: | - CUDA0 (RTX 3080) | 10239 = 9095 + ( 8068 = 6305 + 1536 + 226) + -6923 | common_memory_breakdown_print: | - Host | 11409 = 11385 + 0 + 24 | common_params_fit_impl: memory for test allocation by device: common_params_fit_impl: id=0, n_layer=49, n_part=34, overflow_type=3, mem= 8068 MiB common_params_fit_impl: set ngl_per_device[0].(n_layer, n_part, overflow_type)=(49, 34, GATE), id_dense_start=0 common_params_fit_impl: - CUDA0 (NVIDIA GeForce RTX 3080): 49 layers (34 overflowing), 8068 MiB used, 1026 MiB free common_fit_params: successfully fit params to free device memory common_fit_params: fitting params to free memory took 4.55 seconds llama_model_loader: loaded meta data with 35 key-value pairs and 579 tensors from C:\Users\jerom\.ollama\models\blobs\sha256-1194192cf2a187eb02722edcc3f77b11d21f537048ce04b67ccf8ba78863006a (version GGUF V3 (latest)) llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. llama_model_loader: - kv 0: general.architecture str = qwen3moe llama_model_loader: - kv 1: general.basename str = Qwen3-Coder llama_model_loader: - kv 2: general.file_type u32 = 15 llama_model_loader: - kv 3: general.finetune str = Instruct llama_model_loader: - kv 4: general.license str = apache-2.0 llama_model_loader: - kv 5: general.license.link str = https://huggingface.co/Qwen/Qwen3-Cod... llama_model_loader: - kv 6: general.name str = Qwen3 Coder 30B A3B Instruct llama_model_loader: - kv 7: general.parameter_count u64 = 30532122624 llama_model_loader: - kv 8: general.quantization_version u32 = 2 llama_model_loader: - kv 9: general.size_label str = 30B-A3B llama_model_loader: - kv 10: general.tags arr[str,1] = ["text-generation"] llama_model_loader: - kv 11: general.type str = model llama_model_loader: - kv 12: qwen3moe.attention.head_count u32 = 32 llama_model_loader: - kv 13: qwen3moe.attention.head_count_kv u32 = 4 llama_model_loader: - kv 14: qwen3moe.attention.key_length u32 = 128 llama_model_loader: - kv 15: qwen3moe.attention.layer_norm_rms_epsilon f32 = 0.000001 llama_model_loader: - kv 16: qwen3moe.attention.value_length u32 = 128 llama_model_loader: - kv 17: qwen3moe.block_count u32 = 48 llama_model_loader: - kv 18: qwen3moe.context_length u32 = 262144 llama_model_loader: - kv 19: qwen3moe.embedding_length u32 = 2048 llama_model_loader: - kv 20: qwen3moe.expert_count u32 = 128 llama_model_loader: - kv 21: qwen3moe.expert_feed_forward_length u32 = 768 llama_model_loader: - kv 22: qwen3moe.expert_shared_feed_forward_length u32 = 0 llama_model_loader: - kv 23: qwen3moe.expert_used_count u32 = 8 llama_model_loader: - kv 24: qwen3moe.feed_forward_length u32 = 5472 llama_model_loader: - kv 25: qwen3moe.rope.freq_base f32 = 10000000.000000 llama_model_loader: - kv 26: tokenizer.chat_template str = {% macro render_item_list(item_list, ... llama_model_loader: - kv 27: tokenizer.ggml.add_bos_token bool = false llama_model_loader: - kv 28: tokenizer.ggml.eos_token_id u32 = 151645 llama_model_loader: - kv 29: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... llama_model_loader: - kv 30: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 31: tokenizer.ggml.padding_token_id u32 = 151643 llama_model_loader: - kv 32: tokenizer.ggml.pre str = qwen2 llama_model_loader: - kv 33: tokenizer.ggml.token_type arr[i32,151936] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 34: tokenizer.ggml.tokens arr[str,151936] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - type f32: 241 tensors llama_model_loader: - type q4_K: 289 tensors llama_model_loader: - type q6_K: 49 tensors print_info: file format = GGUF V3 (latest) print_info: file type = Q4_K - Medium print_info: file size = 17.28 GiB (4.86 BPW) llama_prepare_model_devices: using device CUDA0 (NVIDIA GeForce RTX 3080) (0000:01:00.0) - 9095 MiB free load: 0 unused tokens load: control-looking token: 128247 '' was not control-type; this is probably a bug in the model. its type will be overridden load: printing all EOG tokens: load: - 128247 ('') load: - 151643 ('<|endoftext|>') load: - 151645 ('<|im_end|>') load: - 151662 ('<|fim_pad|>') load: - 151663 ('<|repo_name|>') load: - 151664 ('<|file_sep|>') load: special tokens cache size = 27 load: token to piece cache size = 0.9311 MB print_info: arch = qwen3moe print_info: vocab_only = 0 print_info: no_alloc = 0 print_info: n_ctx_train = 262144 print_info: n_embd_inp = 2048 print_info: n_embd = 2048 print_info: n_embd_out = 2048 print_info: n_layer = 48 print_info: n_layer_all = 48 print_info: n_head = 32 print_info: n_head_kv = 4 print_info: n_rot = 128 print_info: n_swa = 0 print_info: is_swa_any = 0 print_info: non_causal_type = 0 print_info: n_embd_head_k = 128 print_info: n_embd_head_v = 128 print_info: n_gqa = 8 print_info: n_embd_k_gqa = 512 print_info: n_embd_v_gqa = 512 print_info: f_norm_eps = 0.0e+00 print_info: f_norm_rms_eps = 1.0e-06 print_info: f_clamp_kqv = 0.0e+00 print_info: f_max_alibi_bias = 0.0e+00 print_info: f_logit_scale = 0.0e+00 print_info: f_attn_scale = 0.0e+00 print_info: f_attn_value_scale = 0.0000 print_info: n_ff = 5472 print_info: n_expert = 128 print_info: n_expert_used = 8 print_info: n_expert_groups = 0 print_info: n_group_used = 0 print_info: causal attn = 1 print_info: pooling type = -1 print_info: rope type = 2 print_info: rope scaling = linear print_info: freq_base_train = 10000000.0 print_info: freq_scale_train = 1 print_info: n_ctx_orig_yarn = 262144 print_info: rope_yarn_log_mul = 0.0000 print_info: rope_finetuned = unknown print_info: model type = 30B.A3B print_info: model params = 30.53 B print_info: general.name = Qwen3 Coder 30B A3B Instruct print_info: n_ff_exp = 768 print_info: vocab type = BPE print_info: n_vocab = 151936 print_info: n_merges = 151387 print_info: BOS token = 11 ',' print_info: EOS token = 151645 '<|im_end|>' print_info: EOT token = 151645 '<|im_end|>' print_info: PAD token = 151643 '<|endoftext|>' print_info: LF token = 198 'Ċ' print_info: FIM PRE token = 151659 '<|fim_prefix|>' print_info: FIM SUF token = 151661 '<|fim_suffix|>' print_info: FIM MID token = 151660 '<|fim_middle|>' print_info: FIM PAD token = 151662 '<|fim_pad|>' print_info: FIM REP token = 151663 '<|repo_name|>' print_info: FIM SEP token = 151664 '<|file_sep|>' print_info: EOG token = 128247 '' print_info: EOG token = 151643 '<|endoftext|>' print_info: EOG token = 151645 '<|im_end|>' print_info: EOG token = 151662 '<|fim_pad|>' print_info: EOG token = 151663 '<|repo_name|>' print_info: EOG token = 151664 '<|file_sep|>' print_info: max token length = 256 load_tensors: loading model tensors, this can take a while... (load_mode = none) load_tensors: offloading output layer to GPU load_tensors: offloading 47 repeating layers to GPU load_tensors: offloaded 49/49 layers to GPU load_tensors: CUDA0 model buffer size = 6305.92 MiB load_tensors: CUDA_Host model buffer size = 11385.42 MiB [GIN] 2026/10/05 - 09:33:17 | 200 | 20.2058ms | 127.0.0.1 | GET "/api/tags" cmn common_init_: added logit bias = -inf cmn common_init_: added <|endoftext|> logit bias = -inf cmn common_init_: added <|im_end|> logit bias = -inf cmn common_init_: added <|fim_pad|> logit bias = -inf cmn common_init_: added <|repo_name|> logit bias = -inf cmn common_init_: added <|file_sep|> logit bias = -inf llama_context: constructing llama_context llama_context: n_seq_max = 1 llama_context: n_ctx = 16384 llama_context: n_ctx_seq = 16384 llama_context: n_batch = 512 llama_context: n_ubatch = 512 llama_context: causal_attn = 1 llama_context: flash_attn = auto llama_context: kv_unified = false llama_context: freq_base = 10000000.0 llama_context: freq_scale = 1 llama_context: n_rs_seq = 0 llama_context: n_outputs_max = 1 llama_context: n_outputs_max_per_seq = 1 llama_context: n_ctx_seq (16384) < n_ctx_train (262144) -- the full capacity of the model will not be utilized llama_context: CUDA_Host output buffer size = 0.58 MiB llama_kv_cache: CUDA0 KV buffer size = 1536.00 MiB llama_kv_cache: size = 1536.00 MiB ( 16384 cells, 48 layers, 1/1 seqs), K (f16): 768.00 MiB, V (f16): 768.00 MiB llama_kv_cache: attn_rot_k = 0, n_embd_head_k_all = 128 llama_kv_cache: attn_rot_v = 0, n_embd_head_k_all = 128 sched_reserve: reserving ... resolve_fused_ops: Flash Attention enabled resolve_fused_ops: resolving fused DeepSeek V4 HC support: resolve_fused_ops: fused DeepSeek V4 HC pre enabled resolve_fused_ops: fused DeepSeek V4 HC comb enabled resolve_fused_ops: fused DeepSeek V4 HC post enabled sched_reserve: CUDA0 compute buffer size = 226.30 MiB sched_reserve: CUDA_Host compute buffer size = 24.01 MiB sched_reserve: graph (pp bs=512, tg bs=1): nodes = 3030 / 3030, splits = 98 / 68, input objects = 4 / 4, input tensors = 6 / 6 sched_reserve: reserve took 34.92 ms, sched copies = 1 cmn init: llama threadpool init, n_threads = 10 cmn common_init_: warming up the model with an empty run - please wait ... (--no-warmup to disable) srv load_model: initializing, n_slots = 1, n_ctx_slot = 16384, kv_unified = 'false' spec common_specu: no implementations specified for speculative decoding slot load_model: id 0 | task -1 | new slot, n_ctx = 16384 srv load_model: prompt cache is enabled, size limit: 8192 MiB srv load_model: use `--cache-ram 0` to disable the prompt cache srv load_model: for more info see https://github.com/ggml-org/llama.cpp/pull/16391 srv load_model: context checkpoints enabled, max = 32, min spacing = 8192 srv init: idle slots will be saved to prompt cache upon starting a new task srv init: init: chat template, example_format: '<|im_start|>system You are a helpful assistant<|im_end|> <|im_start|>user Hello<|im_end|> <|im_start|>assistant Hi there<|im_end|> <|im_start|>user How are you?<|im_end|> <|im_start|>assistant ' srv init: init: chat template, thinking = 0 srv init: preserve_reasoning kwarg: not supported by template srv llama_server: model loaded srv llama_server: listening on http://127.0.0.1:53541 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:33:48 | 200 | 69.9818ms | 127.0.0.1 | GET "/api/tags" time=2026-10-05T09:33:48.197+02:00 level=INFO source=llama_server.go:1374 msg="llama-server started in 58.45 seconds" time=2026-10-05T09:33:48.197+02:00 level=INFO source=images.go:395 msg="template selection" model=registry.ollama.ai/library/qwen3-coder:30b selected=renderer_parser renderer=qwen3-coder parser=qwen3-coder go_template=null chat_template="[tools completion]" harmony=null renderer_parser="[completion tools]" time=2026-10-05T09:33:48.210+02:00 level=INFO source=sched.go:733 msg="loaded runners" count=1 time=2026-10-05T09:33:48.212+02:00 level=INFO source=llama_server.go:1307 msg="waiting for llama-server to start responding" time=2026-10-05T09:33:48.215+02:00 level=INFO source=llama_server.go:1374 msg="llama-server started in 58.52 seconds" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - skipping, slot is empty slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = -1 srv get_availabl: updating prompt cache srv load: - looking for better prompt, base f_keep = -1.000, f_sim = 0.000 srv update: - cache state: 0 prompts, 0.000 MiB (limits: 8192.000 MiB, 16384 tokens, 8589934592 est) srv get_availabl: prompt cache update took 0.02 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 0 | processing task, is_child = 0 slot operator(): id 0 | task 0 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1517 slot operator(): id 0 | task 0 | cached n_tokens = 0, memory_seq_rm [0, end) slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 512, progress = 0.34, t = 13.30 s / 38.50 tokens per second slot operator(): id 0 | task 0 | cached n_tokens = 512, memory_seq_rm [512, end) slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 1024, progress = 0.68, t = 24.57 s / 41.67 tokens per second slot operator(): id 0 | task 0 | cached n_tokens = 1024, memory_seq_rm [1024, end) slot init_sampler: id 0 | task 0 | init sampler, took 0.75 ms, tokens: text = 1517, total = 1517 [GIN] 2026/10/05 - 09:34:18 | 200 | 222.6001ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:34:49 | 200 | 39.562ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 0 | n_gen = 100, tg = 4.09 t/s, tg_3s = 4.13 t/s slot print_timing: id 0 | task 0 | n_gen = 129, tg = 4.70 t/s, tg_3s = 9.51 t/s slot print_timing: id 0 | task 0 | n_gen = 151, tg = 4.82 t/s, tg_3s = 5.66 t/s slot print_timing: id 0 | task 0 | n_gen = 169, tg = 4.63 t/s, tg_3s = 3.49 t/s slot print_timing: id 0 | task 0 | n_gen = 184, tg = 4.48 t/s, tg_3s = 3.31 t/s [GIN] 2026/10/05 - 09:35:19 | 200 | 31.5214ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 0 | n_gen = 187, tg = 4.15 t/s, tg_3s = 0.75 t/s slot print_timing: id 0 | task 0 | n_gen = 208, tg = 4.32 t/s, tg_3s = 6.90 t/s slot print_timing: id 0 | task 0 | n_gen = 259, tg = 5.06 t/s, tg_3s = 16.82 t/s slot print_timing: id 0 | task 0 | prompt eval time = 50170.70 ms / 1517 tokens ( 33.07 ms per token, 30.24 tokens per second) slot print_timing: id 0 | task 0 | eval time = 51182.87 ms / 263 tokens ( 195.35 ms per token, 5.12 tokens per second) slot print_timing: id 0 | task 0 | total time = 101353.57 ms / 1780 tokens slot print_timing: id 0 | task 0 | graphs reused = 261 slot release: id 0 | task 0 | stop processing: n_tokens = 1779, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:35:29 | 200 | 2m41s | 127.0.0.1 | POST "/api/chat" [GIN] 2026/10/05 - 09:35:49 | 200 | 19.9915ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:36:19 | 200 | 33.0771ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:36:49 | 200 | 28.0722ms | 127.0.0.1 | GET "/api/tags" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1437) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 160002594 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1779, total state size = 166.803 MiB (draft: 0.000 MiB) [GIN] 2026/10/05 - 09:37:21 | 200 | 50.1016ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:37:21 | 200 | 11.0155ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/10/05 - 09:37:21 | 200 | 16.5395ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/10/05 - 09:37:25 | 200 | 56.6571ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:37:25 | 200 | 27.5415ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/10/05 - 09:37:25 | 200 | 121.6917ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/10/05 - 09:37:55 | 200 | 28.5504ms | 127.0.0.1 | GET "/api/tags" srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1779, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv update: - cache state: 1 prompts, 166.803 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53A28070: 1779 tokens, checkpoints: 0, 166.803 MiB srv get_availabl: prompt cache update took 53860.66 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 266 | processing task, is_child = 0 slot operator(): id 0 | task 266 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1437 slot operator(): id 0 | task 266 | cached n_tokens = 4, memory_seq_rm [4, end) slot print_timing: id 0 | task 266 | prompt processing, n_tokens = 512, progress = 0.36, t = 6.03 s / 84.89 tokens per second slot operator(): id 0 | task 266 | cached n_tokens = 516, memory_seq_rm [516, end) slot print_timing: id 0 | task 266 | prompt processing, n_tokens = 1024, progress = 0.72, t = 12.35 s / 82.91 tokens per second slot operator(): id 0 | task 266 | cached n_tokens = 1028, memory_seq_rm [1028, end) slot init_sampler: id 0 | task 266 | init sampler, took 1.17 ms, tokens: text = 1437, total = 1437 [GIN] 2026/10/05 - 09:38:25 | 200 | 11.0999ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 266 | prompt eval time = 18132.71 ms / 1433 tokens ( 12.65 ms per token, 79.03 tokens per second) slot print_timing: id 0 | task 266 | eval time = 2314.78 ms / 44 tokens ( 53.83 ms per token, 18.58 tokens per second) slot print_timing: id 0 | task 266 | total time = 20447.49 ms / 1477 tokens slot print_timing: id 0 | task 266 | graphs reused = 303 slot release: id 0 | task 266 | stop processing: n_tokens = 1480, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:38:26 | 200 | 1m14s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.794 (1436/1808) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, f_sim_best = 0.794 (> 0.100 thold), f_keep = 0.970 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 313 | processing task, is_child = 0 slot operator(): id 0 | task 313 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1808 slot operator(): id 0 | task 313 | cached n_tokens = 1436, memory_seq_rm [1436, end) slot init_sampler: id 0 | task 313 | init sampler, took 0.47 ms, tokens: text = 1808, total = 1808 slot print_timing: id 0 | task 313 | n_gen = 100, tg = 17.30 t/s, tg_3s = 17.47 t/s slot print_timing: id 0 | task 313 | n_gen = 157, tg = 17.80 t/s, tg_3s = 18.74 t/s slot print_timing: id 0 | task 313 | n_gen = 208, tg = 17.57 t/s, tg_3s = 16.91 t/s slot print_timing: id 0 | task 313 | n_gen = 261, tg = 17.55 t/s, tg_3s = 17.45 t/s slot print_timing: id 0 | task 313 | n_gen = 313, tg = 17.50 t/s, tg_3s = 17.26 t/s slot print_timing: id 0 | task 313 | n_gen = 364, tg = 17.40 t/s, tg_3s = 16.83 t/s [GIN] 2026/10/05 - 09:38:55 | 200 | 9.1963ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 313 | n_gen = 412, tg = 17.22 t/s, tg_3s = 15.99 t/s slot print_timing: id 0 | task 313 | n_gen = 460, tg = 17.09 t/s, tg_3s = 16.00 t/s slot print_timing: id 0 | task 313 | n_gen = 479, tg = 15.98 t/s, tg_3s = 6.22 t/s slot print_timing: id 0 | task 313 | n_gen = 495, tg = 14.55 t/s, tg_3s = 3.96 t/s slot print_timing: id 0 | task 313 | n_gen = 507, tg = 13.49 t/s, tg_3s = 3.38 t/s slot print_timing: id 0 | task 313 | n_gen = 523, tg = 12.88 t/s, tg_3s = 5.32 t/s slot print_timing: id 0 | task 313 | prompt eval time = 5905.45 ms / 372 tokens ( 15.87 ms per token, 62.99 tokens per second) slot print_timing: id 0 | task 313 | eval time = 40885.55 ms / 528 tokens ( 77.58 ms per token, 12.89 tokens per second) slot print_timing: id 0 | task 313 | total time = 46791.00 ms / 900 tokens slot print_timing: id 0 | task 313 | graphs reused = 827 slot release: id 0 | task 313 | stop processing: n_tokens = 2335, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:39:13 | 200 | 46.8643037s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.004 (5/1269) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 383840099 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 2335, total state size = 218.934 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.004 srv load: - prompt with length 1779, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 2335, lcp = 5, f_keep = 0.002, f_sim = 0.004 srv update: - cache state: 2 prompts, 385.737 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53A28070: 1779 tokens, checkpoints: 0, 166.803 MiB srv update: - prompt 000001EC4EA3E9D0: 2335 tokens, checkpoints: 0, 218.934 MiB srv get_availabl: prompt cache update took 130.51 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 842 | processing task, is_child = 0 slot operator(): id 0 | task 842 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1269 slot operator(): id 0 | task 842 | cached n_tokens = 5, memory_seq_rm [5, end) slot print_timing: id 0 | task 842 | prompt processing, n_tokens = 512, progress = 0.41, t = 5.49 s / 93.19 tokens per second slot operator(): id 0 | task 842 | cached n_tokens = 517, memory_seq_rm [517, end) slot print_timing: id 0 | task 842 | prompt processing, n_tokens = 1024, progress = 0.81, t = 10.99 s / 93.14 tokens per second slot operator(): id 0 | task 842 | cached n_tokens = 1029, memory_seq_rm [1029, end) slot init_sampler: id 0 | task 842 | init sampler, took 0.36 ms, tokens: text = 1269, total = 1269 [GIN] 2026/10/05 - 09:39:25 | 200 | 29.5958ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 842 | n_gen = 100, tg = 16.33 t/s, tg_3s = 16.50 t/s slot print_timing: id 0 | task 842 | prompt eval time = 14760.58 ms / 1264 tokens ( 11.68 ms per token, 85.63 tokens per second) slot print_timing: id 0 | task 842 | eval time = 8619.16 ms / 151 tokens ( 57.46 ms per token, 17.40 tokens per second) slot print_timing: id 0 | task 842 | total time = 23379.75 ms / 1415 tokens slot print_timing: id 0 | task 842 | graphs reused = 975 slot release: id 0 | task 842 | stop processing: n_tokens = 1419, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:39:37 | 200 | 23.5736524s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1517) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 407545944 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1419, total state size = 133.049 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 1779, lcp = 1517, f_keep = 0.853, f_sim = 1.000 srv load: - prompt with length 2335, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1419, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - found better prompt with f_keep = 0.853, f_sim = 1.000 srv update: - cache state: 2 prompts, 351.983 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC4EA3E9D0: 2335 tokens, checkpoints: 0, 218.934 MiB srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv get_availabl: prompt cache update took 405.40 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 996 | processing task, is_child = 0 slot operator(): id 0 | task 996 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1517 slot operator(): id 0 | task 996 | need to evaluate at least 1 token for each active slot (n_past = 1517, task.n_tokens() = 1517) slot operator(): id 0 | task 996 | n_past was set to 1516 slot operator(): id 0 | task 996 | cached n_tokens = 1516, memory_seq_rm [1516, end) slot init_sampler: id 0 | task 996 | init sampler, took 0.37 ms, tokens: text = 1517, total = 1517 slot print_timing: id 0 | task 996 | n_gen = 100, tg = 14.03 t/s, tg_3s = 14.17 t/s slot print_timing: id 0 | task 996 | n_gen = 143, tg = 14.06 t/s, tg_3s = 14.12 t/s slot print_timing: id 0 | task 996 | n_gen = 185, tg = 14.02 t/s, tg_3s = 13.90 t/s slot print_timing: id 0 | task 996 | n_gen = 233, tg = 14.38 t/s, tg_3s = 15.96 t/s [GIN] 2026/10/05 - 09:39:55 | 200 | 6.5147ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 996 | prompt eval time = 61.01 ms / 1 tokens ( 61.01 ms per token, 16.39 tokens per second) slot print_timing: id 0 | task 996 | eval time = 18035.34 ms / 268 tokens ( 67.55 ms per token, 14.80 tokens per second) slot print_timing: id 0 | task 996 | total time = 18096.35 ms / 269 tokens slot print_timing: id 0 | task 996 | graphs reused = 1242 slot release: id 0 | task 996 | stop processing: n_tokens = 1784, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:39:56 | 200 | 18.5590018s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 1.000 (1517/1517) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, f_sim_best = 1.000 (> 0.100 thold), f_keep = 0.850 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 1265 | processing task, is_child = 0 slot operator(): id 0 | task 1265 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1517 slot operator(): id 0 | task 1265 | need to evaluate at least 1 token for each active slot (n_past = 1517, task.n_tokens() = 1517) slot operator(): id 0 | task 1265 | n_past was set to 1516 slot operator(): id 0 | task 1265 | cached n_tokens = 1516, memory_seq_rm [1516, end) slot init_sampler: id 0 | task 1265 | init sampler, took 0.34 ms, tokens: text = 1517, total = 1517 slot print_timing: id 0 | task 1265 | n_gen = 100, tg = 7.46 t/s, tg_3s = 7.54 t/s slot print_timing: id 0 | task 1265 | n_gen = 123, tg = 6.88 t/s, tg_3s = 5.15 t/s slot print_timing: id 0 | task 1265 | n_gen = 142, tg = 6.78 t/s, tg_3s = 6.17 t/s slot print_timing: id 0 | task 1265 | n_gen = 163, tg = 6.80 t/s, tg_3s = 6.97 t/s slot print_timing: id 0 | task 1265 | n_gen = 187, tg = 6.63 t/s, tg_3s = 5.67 t/s [GIN] 2026/10/05 - 09:40:25 | 200 | 12.0154ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 1265 | n_gen = 210, tg = 6.72 t/s, tg_3s = 7.62 t/s slot print_timing: id 0 | task 1265 | n_gen = 236, tg = 6.63 t/s, tg_3s = 5.99 t/s slot print_timing: id 0 | task 1265 | n_gen = 254, tg = 6.58 t/s, tg_3s = 5.94 t/s slot print_timing: id 0 | task 1265 | prompt eval time = 57.91 ms / 1 tokens ( 57.91 ms per token, 17.27 tokens per second) slot print_timing: id 0 | task 1265 | eval time = 40852.28 ms / 290 tokens ( 141.36 ms per token, 7.07 tokens per second) slot print_timing: id 0 | task 1265 | total time = 40910.18 ms / 291 tokens slot print_timing: id 0 | task 1265 | graphs reused = 1529 slot release: id 0 | task 1265 | stop processing: n_tokens = 1806, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:40:37 | 200 | 40.9457201s | 127.0.0.1 | POST "/api/chat" [GIN] 2026/10/05 - 09:40:55 | 200 | 6.0012ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:41:25 | 200 | 8.0111ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:41:56 | 200 | 267.7074ms | 127.0.0.1 | GET "/api/tags" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1547) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 468178308 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1806, total state size = 169.334 MiB (draft: 0.000 MiB) [GIN] 2026/10/05 - 09:42:26 | 200 | 140.0457ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:42:56 | 200 | 52.2605ms | 127.0.0.1 | GET "/api/tags" srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 2335, lcp = 628, f_keep = 0.269, f_sim = 0.406 srv load: - prompt with length 1419, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 1806, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - found better prompt with f_keep = 0.269, f_sim = 0.406 srv update: - cache state: 2 prompts, 302.383 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EC53D699B0: 1806 tokens, checkpoints: 0, 169.334 MiB srv get_availabl: prompt cache update took 59636.18 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 1556 | processing task, is_child = 0 slot operator(): id 0 | task 1556 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1547 slot operator(): id 0 | task 1556 | cached n_tokens = 628, memory_seq_rm [628, end) [GIN] 2026/10/05 - 09:43:26 | 200 | 39.687ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 1556 | prompt processing, n_tokens = 512, progress = 0.74, t = 8.49 s / 60.28 tokens per second slot operator(): id 0 | task 1556 | cached n_tokens = 1140, memory_seq_rm [1140, end) slot init_sampler: id 0 | task 1556 | init sampler, took 1.18 ms, tokens: text = 1547, total = 1547 [GIN] 2026/10/05 - 09:43:56 | 200 | 37.1222ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 1556 | prompt eval time = 34002.53 ms / 919 tokens ( 37.00 ms per token, 27.03 tokens per second) slot print_timing: id 0 | task 1556 | eval time = 2809.15 ms / 44 tokens ( 65.33 ms per token, 15.31 tokens per second) slot print_timing: id 0 | task 1556 | total time = 36811.68 ms / 963 tokens slot print_timing: id 0 | task 1556 | graphs reused = 1571 slot release: id 0 | task 1556 | stop processing: n_tokens = 1590, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:43:58 | 200 | 1m37s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.762 (1546/2028) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, f_sim_best = 0.762 (> 0.100 thold), f_keep = 0.972 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 1602 | processing task, is_child = 0 slot operator(): id 0 | task 1602 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 2028 slot operator(): id 0 | task 1602 | cached n_tokens = 1546, memory_seq_rm [1546, end) slot init_sampler: id 0 | task 1602 | init sampler, took 0.62 ms, tokens: text = 2028, total = 2028 slot print_timing: id 0 | task 1602 | n_gen = 100, tg = 16.36 t/s, tg_3s = 16.52 t/s slot print_timing: id 0 | task 1602 | n_gen = 149, tg = 16.32 t/s, tg_3s = 16.26 t/s slot print_timing: id 0 | task 1602 | n_gen = 196, tg = 16.10 t/s, tg_3s = 15.43 t/s slot print_timing: id 0 | task 1602 | n_gen = 244, tg = 16.06 t/s, tg_3s = 15.91 t/s slot print_timing: id 0 | task 1602 | n_gen = 296, tg = 16.27 t/s, tg_3s = 17.33 t/s slot print_timing: id 0 | task 1602 | n_gen = 344, tg = 16.21 t/s, tg_3s = 15.83 t/s slot print_timing: id 0 | task 1602 | n_gen = 381, tg = 15.69 t/s, tg_3s = 12.13 t/s [GIN] 2026/10/05 - 09:44:27 | 200 | 299.011ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 1602 | n_gen = 417, tg = 15.25 t/s, tg_3s = 11.76 t/s slot print_timing: id 0 | task 1602 | n_gen = 454, tg = 14.96 t/s, tg_3s = 12.30 t/s slot print_timing: id 0 | task 1602 | n_gen = 503, tg = 15.08 t/s, tg_3s = 16.28 t/s slot print_timing: id 0 | task 1602 | n_gen = 542, tg = 14.90 t/s, tg_3s = 12.98 t/s slot print_timing: id 0 | task 1602 | prompt eval time = 4274.11 ms / 482 tokens ( 8.87 ms per token, 112.77 tokens per second) slot print_timing: id 0 | task 1602 | eval time = 37341.68 ms / 554 tokens ( 67.53 ms per token, 14.81 tokens per second) slot print_timing: id 0 | task 1602 | total time = 41615.79 ms / 1036 tokens slot print_timing: id 0 | task 1602 | graphs reused = 2120 slot release: id 0 | task 1602 | stop processing: n_tokens = 2581, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:44:39 | 200 | 41.7163771s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.004 (5/1278) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 710158692 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 2581, total state size = 241.999 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.004 srv load: - prompt with length 1419, lcp = 101, f_keep = 0.071, f_sim = 0.079 srv load: - prompt with length 1806, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 2581, lcp = 5, f_keep = 0.002, f_sim = 0.004 srv update: - cache state: 3 prompts, 544.382 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EC53D699B0: 1806 tokens, checkpoints: 0, 169.334 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv get_availabl: prompt cache update took 204.24 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 2157 | processing task, is_child = 0 slot operator(): id 0 | task 2157 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1278 slot operator(): id 0 | task 2157 | cached n_tokens = 5, memory_seq_rm [5, end) slot print_timing: id 0 | task 2157 | prompt processing, n_tokens = 512, progress = 0.40, t = 3.72 s / 137.54 tokens per second slot operator(): id 0 | task 2157 | cached n_tokens = 517, memory_seq_rm [517, end) slot print_timing: id 0 | task 2157 | prompt processing, n_tokens = 1024, progress = 0.81, t = 7.46 s / 137.26 tokens per second slot operator(): id 0 | task 2157 | cached n_tokens = 1029, memory_seq_rm [1029, end) slot init_sampler: id 0 | task 2157 | init sampler, took 0.50 ms, tokens: text = 1278, total = 1278 slot print_timing: id 0 | task 2157 | n_gen = 100, tg = 15.13 t/s, tg_3s = 15.28 t/s [GIN] 2026/10/05 - 09:44:57 | 200 | 13.5336ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 2157 | prompt eval time = 10461.61 ms / 1273 tokens ( 8.22 ms per token, 121.68 tokens per second) slot print_timing: id 0 | task 2157 | eval time = 8532.70 ms / 136 tokens ( 63.21 ms per token, 15.82 tokens per second) slot print_timing: id 0 | task 2157 | total time = 18994.31 ms / 1409 tokens slot print_timing: id 0 | task 2157 | graphs reused = 2253 slot release: id 0 | task 2157 | stop processing: n_tokens = 1413, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:44:59 | 200 | 19.2595037s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1517) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 729529888 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1413, total state size = 132.486 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 1419, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 1806, lcp = 1517, f_keep = 0.840, f_sim = 1.000 srv load: - prompt with length 2581, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1413, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - found better prompt with f_keep = 0.840, f_sim = 1.000 srv update: - cache state: 3 prompts, 507.534 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv get_availabl: prompt cache update took 853.42 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 2296 | processing task, is_child = 0 slot operator(): id 0 | task 2296 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1517 slot operator(): id 0 | task 2296 | need to evaluate at least 1 token for each active slot (n_past = 1517, task.n_tokens() = 1517) slot operator(): id 0 | task 2296 | n_past was set to 1516 slot operator(): id 0 | task 2296 | cached n_tokens = 1516, memory_seq_rm [1516, end) slot init_sampler: id 0 | task 2296 | init sampler, took 0.37 ms, tokens: text = 1517, total = 1517 slot print_timing: id 0 | task 2296 | n_gen = 100, tg = 17.38 t/s, tg_3s = 17.55 t/s slot print_timing: id 0 | task 2296 | n_gen = 142, tg = 14.11 t/s, tg_3s = 9.78 t/s slot print_timing: id 0 | task 2296 | n_gen = 178, tg = 13.63 t/s, tg_3s = 12.00 t/s slot print_timing: id 0 | task 2296 | n_gen = 224, tg = 13.89 t/s, tg_3s = 14.99 t/s slot print_timing: id 0 | task 2296 | n_gen = 268, tg = 13.99 t/s, tg_3s = 14.56 t/s slot print_timing: id 0 | task 2296 | prompt eval time = 102.09 ms / 1 tokens ( 102.09 ms per token, 9.80 tokens per second) slot print_timing: id 0 | task 2296 | eval time = 19714.97 ms / 277 tokens ( 71.43 ms per token, 14.00 tokens per second) slot print_timing: id 0 | task 2296 | total time = 19817.06 ms / 278 tokens slot print_timing: id 0 | task 2296 | graphs reused = 2528 slot release: id 0 | task 2296 | stop processing: n_tokens = 1793, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:45:21 | 200 | 20.699706s | 127.0.0.1 | POST "/api/chat" [GIN] 2026/10/05 - 09:45:27 | 200 | 8.0084ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:45:57 | 200 | 8.6081ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:46:27 | 200 | 18.8066ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:46:57 | 200 | 11.063ms | 127.0.0.1 | GET "/api/tags" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1537) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 751312767 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1793, total state size = 168.115 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1419, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 2581, lcp = 630, f_keep = 0.244, f_sim = 0.410 srv load: - prompt with length 1413, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 1793, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv update: - cache state: 4 prompts, 675.649 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE3017E160: 1793 tokens, checkpoints: 0, 168.115 MiB srv get_availabl: prompt cache update took 20145.57 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 2574 | processing task, is_child = 0 slot operator(): id 0 | task 2574 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1537 slot operator(): id 0 | task 2574 | cached n_tokens = 4, memory_seq_rm [4, end) [GIN] 2026/10/05 - 09:47:27 | 200 | 9.0083ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 2574 | prompt processing, n_tokens = 512, progress = 0.34, t = 14.74 s / 34.74 tokens per second slot operator(): id 0 | task 2574 | cached n_tokens = 516, memory_seq_rm [516, end) [GIN] 2026/10/05 - 09:47:57 | 200 | 9.0135ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:48:28 | 200 | 131.3441ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 2574 | prompt processing, n_tokens = 1024, progress = 0.67, t = 72.94 s / 14.04 tokens per second slot operator(): id 0 | task 2574 | cached n_tokens = 1028, memory_seq_rm [1028, end) slot init_sampler: id 0 | task 2574 | init sampler, took 1.67 ms, tokens: text = 1537, total = 1537 [GIN] 2026/10/05 - 09:48:58 | 200 | 47.4019ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 2574 | prompt eval time = 85030.20 ms / 1533 tokens ( 55.47 ms per token, 18.03 tokens per second) slot print_timing: id 0 | task 2574 | eval time = 35854.27 ms / 44 tokens ( 833.82 ms per token, 1.20 tokens per second) slot print_timing: id 0 | task 2574 | total time = 120884.48 ms / 1577 tokens slot print_timing: id 0 | task 2574 | graphs reused = 2570 slot release: id 0 | task 2574 | stop processing: n_tokens = 1580, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:49:24 | 200 | 2m21s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.761 (1536/2018) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, f_sim_best = 0.761 (> 0.100 thold), f_keep = 0.972 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 2621 | processing task, is_child = 0 slot operator(): id 0 | task 2621 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 2018 slot operator(): id 0 | task 2621 | cached n_tokens = 1536, memory_seq_rm [1536, end) slot init_sampler: id 0 | task 2621 | init sampler, took 0.56 ms, tokens: text = 2018, total = 2018 [GIN] 2026/10/05 - 09:49:28 | 200 | 8.0354ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:49:58 | 200 | 10.501ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 2621 | n_gen = 100, tg = 17.24 t/s, tg_3s = 17.42 t/s slot print_timing: id 0 | task 2621 | n_gen = 151, tg = 17.11 t/s, tg_3s = 16.85 t/s slot print_timing: id 0 | task 2621 | n_gen = 196, tg = 16.53 t/s, tg_3s = 14.87 t/s slot print_timing: id 0 | task 2621 | n_gen = 210, tg = 14.04 t/s, tg_3s = 4.52 t/s slot print_timing: id 0 | task 2621 | n_gen = 226, tg = 12.54 t/s, tg_3s = 5.23 t/s slot print_timing: id 0 | task 2621 | n_gen = 233, tg = 8.24 t/s, tg_3s = 0.68 t/s slot print_timing: id 0 | task 2621 | n_gen = 244, tg = 7.64 t/s, tg_3s = 3.04 t/s [GIN] 2026/10/05 - 09:50:28 | 200 | 14ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 2621 | n_gen = 259, tg = 7.41 t/s, tg_3s = 4.97 t/s slot print_timing: id 0 | task 2621 | n_gen = 294, tg = 7.73 t/s, tg_3s = 11.37 t/s slot print_timing: id 0 | task 2621 | n_gen = 331, tg = 8.06 t/s, tg_3s = 12.17 t/s slot print_timing: id 0 | task 2621 | n_gen = 360, tg = 8.16 t/s, tg_3s = 9.50 t/s slot print_timing: id 0 | task 2621 | n_gen = 399, tg = 8.47 t/s, tg_3s = 12.91 t/s slot print_timing: id 0 | task 2621 | n_gen = 444, tg = 8.85 t/s, tg_3s = 14.76 t/s slot print_timing: id 0 | task 2621 | n_gen = 484, tg = 9.10 t/s, tg_3s = 13.13 t/s slot print_timing: id 0 | task 2621 | prompt eval time = 31685.88 ms / 482 tokens ( 65.74 ms per token, 15.21 tokens per second) slot print_timing: id 0 | task 2621 | eval time = 53892.46 ms / 494 tokens ( 109.32 ms per token, 9.15 tokens per second) slot print_timing: id 0 | task 2621 | total time = 85578.34 ms / 976 tokens slot print_timing: id 0 | task 2621 | graphs reused = 3060 slot release: id 0 | task 2621 | stop processing: n_tokens = 2511, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:50:50 | 200 | 1m25s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.004 (5/1158) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1080600907 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 2511, total state size = 235.436 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.004 srv load: - prompt with length 1419, lcp = 101, f_keep = 0.071, f_sim = 0.087 srv load: - prompt with length 2581, lcp = 5, f_keep = 0.002, f_sim = 0.004 srv load: - prompt with length 1413, lcp = 102, f_keep = 0.072, f_sim = 0.088 srv load: - prompt with length 1793, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 2511, lcp = 5, f_keep = 0.002, f_sim = 0.004 srv update: - cache state: 5 prompts, 911.086 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE3017E160: 1793 tokens, checkpoints: 0, 168.115 MiB srv update: - prompt 000001EC53D69B90: 2511 tokens, checkpoints: 0, 235.436 MiB srv get_availabl: prompt cache update took 773.64 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 3116 | processing task, is_child = 0 slot operator(): id 0 | task 3116 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1158 slot operator(): id 0 | task 3116 | cached n_tokens = 5, memory_seq_rm [5, end) slot print_timing: id 0 | task 3116 | prompt processing, n_tokens = 512, progress = 0.45, t = 3.90 s / 131.13 tokens per second slot operator(): id 0 | task 3116 | cached n_tokens = 517, memory_seq_rm [517, end) [GIN] 2026/10/05 - 09:50:58 | 200 | 9.0734ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 3116 | prompt processing, n_tokens = 1024, progress = 0.89, t = 8.57 s / 119.47 tokens per second slot operator(): id 0 | task 3116 | cached n_tokens = 1029, memory_seq_rm [1029, end) slot init_sampler: id 0 | task 3116 | init sampler, took 0.76 ms, tokens: text = 1158, total = 1158 slot print_timing: id 0 | task 3116 | n_gen = 100, tg = 16.25 t/s, tg_3s = 16.42 t/s slot print_timing: id 0 | task 3116 | n_gen = 137, tg = 14.88 t/s, tg_3s = 12.14 t/s slot print_timing: id 0 | task 3116 | prompt eval time = 11264.72 ms / 1153 tokens ( 9.77 ms per token, 102.35 tokens per second) slot print_timing: id 0 | task 3116 | eval time = 9972.38 ms / 147 tokens ( 68.30 ms per token, 14.64 tokens per second) slot print_timing: id 0 | task 3116 | total time = 21237.10 ms / 1300 tokens slot print_timing: id 0 | task 3116 | graphs reused = 3204 slot release: id 0 | task 3116 | stop processing: n_tokens = 1304, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:51:14 | 200 | 23.514128s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1517) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1104371779 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1304, total state size = 122.266 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 1419, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 2581, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1413, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 1793, lcp = 1517, f_keep = 0.846, f_sim = 1.000 srv load: - prompt with length 2511, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1304, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - found better prompt with f_keep = 0.846, f_sim = 1.000 srv update: - cache state: 5 prompts, 865.236 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EC53D69B90: 2511 tokens, checkpoints: 0, 235.436 MiB srv update: - prompt 000001EE40A30850: 1304 tokens, checkpoints: 0, 122.266 MiB srv get_availabl: prompt cache update took 1069.14 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 3266 | processing task, is_child = 0 slot operator(): id 0 | task 3266 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1517 slot operator(): id 0 | task 3266 | need to evaluate at least 1 token for each active slot (n_past = 1517, task.n_tokens() = 1517) slot operator(): id 0 | task 3266 | n_past was set to 1516 slot operator(): id 0 | task 3266 | cached n_tokens = 1516, memory_seq_rm [1516, end) slot init_sampler: id 0 | task 3266 | init sampler, took 0.34 ms, tokens: text = 1517, total = 1517 slot print_timing: id 0 | task 3266 | n_gen = 100, tg = 9.52 t/s, tg_3s = 9.61 t/s [GIN] 2026/10/05 - 09:51:28 | 200 | 15.4995ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 3266 | n_gen = 142, tg = 10.50 t/s, tg_3s = 13.86 t/s slot print_timing: id 0 | task 3266 | n_gen = 186, tg = 11.20 t/s, tg_3s = 14.28 t/s slot print_timing: id 0 | task 3266 | n_gen = 194, tg = 7.39 t/s, tg_3s = 0.83 t/s slot print_timing: id 0 | task 3266 | n_gen = 222, tg = 7.58 t/s, tg_3s = 9.28 t/s slot print_timing: id 0 | task 3266 | n_gen = 243, tg = 7.12 t/s, tg_3s = 4.31 t/s slot print_timing: id 0 | task 3266 | n_gen = 251, tg = 6.62 t/s, tg_3s = 2.14 t/s slot print_timing: id 0 | task 3266 | n_gen = 259, tg = 6.30 t/s, tg_3s = 2.49 t/s [GIN] 2026/10/05 - 09:52:00 | 200 | 45.5228062s | 127.0.0.1 | POST "/api/chat" slot print_timing: id 0 | task 3266 | prompt eval time = 210.85 ms / 1 tokens ( 210.85 ms per token, 4.74 tokens per second) slot print_timing: id 0 | task 3266 | eval time = 44033.25 ms / 262 tokens ( 168.71 ms per token, 5.93 tokens per second) slot print_timing: id 0 | task 3266 | total time = 44244.10 ms / 263 tokens slot print_timing: id 0 | task 3266 | graphs reused = 3465 slot release: id 0 | task 3266 | stop processing: n_tokens = 1778, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:52:01 | 200 | 2.3444969s | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:52:31 | 200 | 7.5011ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:53:01 | 200 | 17.5714ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:53:31 | 200 | 234.5008ms | 127.0.0.1 | GET "/api/tags" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1525) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1150904984 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1778, total state size = 166.709 MiB (draft: 0.000 MiB) [GIN] 2026/10/05 - 09:54:01 | 200 | 23.0299ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:54:31 | 200 | 7.2632ms | 127.0.0.1 | GET "/api/tags" srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1419, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 2581, lcp = 628, f_keep = 0.243, f_sim = 0.412 srv load: - prompt with length 1413, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 2511, lcp = 628, f_keep = 0.250, f_sim = 0.412 srv load: - prompt with length 1304, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 1778, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - found better prompt with f_keep = 0.250, f_sim = 0.412 srv update: - cache state: 5 prompts, 796.509 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE40A30850: 1304 tokens, checkpoints: 0, 122.266 MiB srv update: - prompt 000001EE40A30CB0: 1778 tokens, checkpoints: 0, 166.709 MiB srv get_availabl: prompt cache update took 51743.00 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 3529 | processing task, is_child = 0 slot operator(): id 0 | task 3529 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1525 slot operator(): id 0 | task 3529 | cached n_tokens = 628, memory_seq_rm [628, end) slot print_timing: id 0 | task 3529 | prompt processing, n_tokens = 512, progress = 0.75, t = 13.58 s / 37.71 tokens per second slot operator(): id 0 | task 3529 | cached n_tokens = 1140, memory_seq_rm [1140, end) slot init_sampler: id 0 | task 3529 | init sampler, took 0.99 ms, tokens: text = 1525, total = 1525 slot print_timing: id 0 | task 3529 | prompt eval time = 17054.73 ms / 897 tokens ( 19.01 ms per token, 52.60 tokens per second) slot print_timing: id 0 | task 3529 | eval time = 2051.74 ms / 44 tokens ( 47.71 ms per token, 20.96 tokens per second) slot print_timing: id 0 | task 3529 | total time = 19106.47 ms / 941 tokens slot print_timing: id 0 | task 3529 | graphs reused = 3506 slot release: id 0 | task 3529 | stop processing: n_tokens = 1568, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:54:54 | 200 | 1m11s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.760 (1524/2006) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, f_sim_best = 0.760 (> 0.100 thold), f_keep = 0.972 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 3575 | processing task, is_child = 0 slot operator(): id 0 | task 3575 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 2006 slot operator(): id 0 | task 3575 | cached n_tokens = 1524, memory_seq_rm [1524, end) slot init_sampler: id 0 | task 3575 | init sampler, took 0.63 ms, tokens: text = 2006, total = 2006 [GIN] 2026/10/05 - 09:55:01 | 200 | 42.7036ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 3575 | n_gen = 100, tg = 22.11 t/s, tg_3s = 22.33 t/s slot print_timing: id 0 | task 3575 | n_gen = 161, tg = 21.35 t/s, tg_3s = 20.22 t/s slot print_timing: id 0 | task 3575 | n_gen = 217, tg = 20.50 t/s, tg_3s = 18.41 t/s slot print_timing: id 0 | task 3575 | n_gen = 271, tg = 19.90 t/s, tg_3s = 17.84 t/s slot print_timing: id 0 | task 3575 | n_gen = 312, tg = 18.18 t/s, tg_3s = 11.58 t/s slot print_timing: id 0 | task 3575 | n_gen = 323, tg = 15.99 t/s, tg_3s = 3.63 t/s slot print_timing: id 0 | task 3575 | n_gen = 338, tg = 14.52 t/s, tg_3s = 4.87 t/s slot print_timing: id 0 | task 3575 | n_gen = 352, tg = 13.03 t/s, tg_3s = 3.76 t/s slot print_timing: id 0 | task 3575 | n_gen = 368, tg = 11.59 t/s, tg_3s = 3.39 t/s [GIN] 2026/10/05 - 09:55:31 | 200 | 6.0077ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 3575 | n_gen = 386, tg = 11.09 t/s, tg_3s = 5.92 t/s slot print_timing: id 0 | task 3575 | n_gen = 402, tg = 10.32 t/s, tg_3s = 3.86 t/s slot print_timing: id 0 | task 3575 | n_gen = 418, tg = 9.76 t/s, tg_3s = 4.14 t/s slot print_timing: id 0 | task 3575 | n_gen = 433, tg = 9.45 t/s, tg_3s = 4.96 t/s slot print_timing: id 0 | task 3575 | n_gen = 456, tg = 9.33 t/s, tg_3s = 7.61 t/s slot print_timing: id 0 | task 3575 | prompt eval time = 3821.68 ms / 482 tokens ( 7.93 ms per token, 126.12 tokens per second) slot print_timing: id 0 | task 3575 | eval time = 50357.05 ms / 482 tokens ( 104.69 ms per token, 9.55 tokens per second) slot print_timing: id 0 | task 3575 | total time = 54178.73 ms / 964 tokens slot print_timing: id 0 | task 3575 | graphs reused = 3984 slot release: id 0 | task 3575 | stop processing: n_tokens = 2487, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:55:48 | 200 | 54.2285418s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.004 (5/1191) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1379033858 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 2487, total state size = 233.186 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.004 srv load: - prompt with length 1419, lcp = 101, f_keep = 0.071, f_sim = 0.085 srv load: - prompt with length 2581, lcp = 5, f_keep = 0.002, f_sim = 0.004 srv load: - prompt with length 1413, lcp = 102, f_keep = 0.072, f_sim = 0.086 srv load: - prompt with length 1304, lcp = 102, f_keep = 0.078, f_sim = 0.086 srv load: - prompt with length 1778, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 2487, lcp = 5, f_keep = 0.002, f_sim = 0.004 srv update: - cache state: 6 prompts, 1029.695 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE40A30850: 1304 tokens, checkpoints: 0, 122.266 MiB srv update: - prompt 000001EE40A30CB0: 1778 tokens, checkpoints: 0, 166.709 MiB srv update: - prompt 000001EE34181330: 2487 tokens, checkpoints: 0, 233.186 MiB srv get_availabl: prompt cache update took 181.40 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 4058 | processing task, is_child = 0 slot operator(): id 0 | task 4058 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1191 slot operator(): id 0 | task 4058 | cached n_tokens = 5, memory_seq_rm [5, end) [GIN] 2026/10/05 - 09:56:01 | 200 | 65.7136ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:56:31 | 200 | 64.9076ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 4058 | prompt processing, n_tokens = 512, progress = 0.43, t = 49.79 s / 10.28 tokens per second slot operator(): id 0 | task 4058 | cached n_tokens = 517, memory_seq_rm [517, end) slot print_timing: id 0 | task 4058 | prompt processing, n_tokens = 1024, progress = 0.86, t = 57.72 s / 17.74 tokens per second slot operator(): id 0 | task 4058 | cached n_tokens = 1029, memory_seq_rm [1029, end) slot init_sampler: id 0 | task 4058 | init sampler, took 0.92 ms, tokens: text = 1191, total = 1191 slot print_timing: id 0 | task 4058 | n_gen = 100, tg = 21.33 t/s, tg_3s = 21.55 t/s slot print_timing: id 0 | task 4058 | prompt eval time = 60668.86 ms / 1186 tokens ( 51.15 ms per token, 19.55 tokens per second) slot print_timing: id 0 | task 4058 | eval time = 5124.01 ms / 110 tokens ( 47.01 ms per token, 21.27 tokens per second) slot print_timing: id 0 | task 4058 | total time = 65792.87 ms / 1296 tokens slot print_timing: id 0 | task 4058 | graphs reused = 4091 slot release: id 0 | task 4058 | stop processing: n_tokens = 1300, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:56:54 | 200 | 1m6s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1517) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1445164918 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1300, total state size = 121.891 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 1419, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 2581, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1413, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 1304, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - prompt with length 1778, lcp = 1517, f_keep = 0.853, f_sim = 1.000 srv load: - prompt with length 2487, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1300, lcp = 4, f_keep = 0.003, f_sim = 0.003 srv load: - found better prompt with f_keep = 0.853, f_sim = 1.000 srv update: - cache state: 6 prompts, 984.877 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE40A30850: 1304 tokens, checkpoints: 0, 122.266 MiB srv update: - prompt 000001EE34181330: 2487 tokens, checkpoints: 0, 233.186 MiB srv update: - prompt 000001EE34180250: 1300 tokens, checkpoints: 0, 121.891 MiB srv get_availabl: prompt cache update took 1020.63 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 4171 | processing task, is_child = 0 slot operator(): id 0 | task 4171 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1517 slot operator(): id 0 | task 4171 | need to evaluate at least 1 token for each active slot (n_past = 1517, task.n_tokens() = 1517) slot operator(): id 0 | task 4171 | n_past was set to 1516 slot operator(): id 0 | task 4171 | cached n_tokens = 1516, memory_seq_rm [1516, end) slot init_sampler: id 0 | task 4171 | init sampler, took 0.36 ms, tokens: text = 1517, total = 1517 [GIN] 2026/10/05 - 09:57:01 | 200 | 20.5224ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 4171 | n_gen = 100, tg = 17.11 t/s, tg_3s = 17.28 t/s slot print_timing: id 0 | task 4171 | n_gen = 143, tg = 16.11 t/s, tg_3s = 14.21 t/s slot print_timing: id 0 | task 4171 | n_gen = 187, tg = 15.67 t/s, tg_3s = 14.41 t/s slot print_timing: id 0 | task 4171 | n_gen = 232, tg = 15.49 t/s, tg_3s = 14.78 t/s slot print_timing: id 0 | task 4171 | prompt eval time = 60.35 ms / 1 tokens ( 60.35 ms per token, 16.57 tokens per second) slot print_timing: id 0 | task 4171 | eval time = 17847.85 ms / 285 tokens ( 62.84 ms per token, 15.91 tokens per second) slot print_timing: id 0 | task 4171 | total time = 17908.20 ms / 286 tokens slot print_timing: id 0 | task 4171 | graphs reused = 4374 slot release: id 0 | task 4171 | stop processing: n_tokens = 1801, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:57:15 | 200 | 19.2206668s | 127.0.0.1 | POST "/api/chat" [GIN] 2026/10/05 - 09:57:31 | 200 | 5.4997ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:58:01 | 200 | 15.5028ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 09:58:31 | 200 | 13.936ms | 127.0.0.1 | GET "/api/tags" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.003 (4/1554) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1465488063 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1801, total state size = 168.865 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 1419, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 2581, lcp = 630, f_keep = 0.244, f_sim = 0.405 srv load: - prompt with length 1413, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 1304, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 2487, lcp = 628, f_keep = 0.253, f_sim = 0.404 srv load: - prompt with length 1300, lcp = 5, f_keep = 0.004, f_sim = 0.003 srv load: - prompt with length 1801, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - found better prompt with f_keep = 0.253, f_sim = 0.404 srv update: - cache state: 6 prompts, 920.557 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE40A30850: 1304 tokens, checkpoints: 0, 122.266 MiB srv update: - prompt 000001EE34180250: 1300 tokens, checkpoints: 0, 121.891 MiB srv update: - prompt 000001EC4E9D5DA0: 1801 tokens, checkpoints: 0, 168.865 MiB srv get_availabl: prompt cache update took 1105.38 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 4457 | processing task, is_child = 0 slot operator(): id 0 | task 4457 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1554 slot operator(): id 0 | task 4457 | cached n_tokens = 628, memory_seq_rm [628, end) [GIN] 2026/10/05 - 09:59:01 | 200 | 7.979ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 4457 | prompt processing, n_tokens = 512, progress = 0.73, t = 6.22 s / 82.33 tokens per second slot operator(): id 0 | task 4457 | cached n_tokens = 1140, memory_seq_rm [1140, end) slot init_sampler: id 0 | task 4457 | init sampler, took 0.45 ms, tokens: text = 1554, total = 1554 slot print_timing: id 0 | task 4457 | prompt eval time = 11898.03 ms / 926 tokens ( 12.85 ms per token, 77.83 tokens per second) slot print_timing: id 0 | task 4457 | eval time = 2474.32 ms / 44 tokens ( 57.54 ms per token, 17.38 tokens per second) slot print_timing: id 0 | task 4457 | total time = 14372.35 ms / 970 tokens slot print_timing: id 0 | task 4457 | graphs reused = 4416 slot release: id 0 | task 4457 | stop processing: n_tokens = 1597, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:59:13 | 200 | 15.5538588s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.763 (1553/2035) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, f_sim_best = 0.763 (> 0.100 thold), f_keep = 0.972 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 4503 | processing task, is_child = 0 slot operator(): id 0 | task 4503 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 2035 slot operator(): id 0 | task 4503 | cached n_tokens = 1553, memory_seq_rm [1553, end) slot init_sampler: id 0 | task 4503 | init sampler, took 0.64 ms, tokens: text = 2035, total = 2035 slot print_timing: id 0 | task 4503 | n_gen = 100, tg = 15.94 t/s, tg_3s = 16.10 t/s slot print_timing: id 0 | task 4503 | n_gen = 148, tg = 15.88 t/s, tg_3s = 15.75 t/s [GIN] 2026/10/05 - 09:59:31 | 200 | 30.0614ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 4503 | n_gen = 194, tg = 15.68 t/s, tg_3s = 15.06 t/s slot print_timing: id 0 | task 4503 | n_gen = 240, tg = 15.56 t/s, tg_3s = 15.10 t/s slot print_timing: id 0 | task 4503 | n_gen = 274, tg = 14.85 t/s, tg_3s = 11.22 t/s slot print_timing: id 0 | task 4503 | n_gen = 307, tg = 14.23 t/s, tg_3s = 10.58 t/s slot print_timing: id 0 | task 4503 | n_gen = 339, tg = 13.78 t/s, tg_3s = 10.59 t/s slot print_timing: id 0 | task 4503 | n_gen = 383, tg = 13.39 t/s, tg_3s = 11.01 t/s slot print_timing: id 0 | task 4503 | n_gen = 423, tg = 13.36 t/s, tg_3s = 13.08 t/s slot print_timing: id 0 | task 4503 | n_gen = 464, tg = 13.38 t/s, tg_3s = 13.58 t/s slot print_timing: id 0 | task 4503 | n_gen = 506, tg = 13.42 t/s, tg_3s = 13.91 t/s slot print_timing: id 0 | task 4503 | prompt eval time = 7315.34 ms / 482 tokens ( 15.18 ms per token, 65.89 tokens per second) slot print_timing: id 0 | task 4503 | eval time = 38984.93 ms / 525 tokens ( 74.40 ms per token, 13.44 tokens per second) slot print_timing: id 0 | task 4503 | total time = 46300.27 ms / 1007 tokens slot print_timing: id 0 | task 4503 | graphs reused = 4937 slot release: id 0 | task 4503 | stop processing: n_tokens = 2559, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 09:59:59 | 200 | 46.3280122s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.004 (5/1177) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1629665517 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 2559, total state size = 239.937 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.002, f_sim = 0.004 srv load: - prompt with length 1419, lcp = 101, f_keep = 0.071, f_sim = 0.086 srv load: - prompt with length 2581, lcp = 5, f_keep = 0.002, f_sim = 0.004 srv load: - prompt with length 1413, lcp = 102, f_keep = 0.072, f_sim = 0.087 srv load: - prompt with length 1304, lcp = 102, f_keep = 0.078, f_sim = 0.087 srv load: - prompt with length 1300, lcp = 102, f_keep = 0.078, f_sim = 0.087 srv load: - prompt with length 1801, lcp = 4, f_keep = 0.002, f_sim = 0.003 srv load: - prompt with length 2559, lcp = 5, f_keep = 0.002, f_sim = 0.004 srv update: - cache state: 7 prompts, 1160.493 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE40A30850: 1304 tokens, checkpoints: 0, 122.266 MiB srv update: - prompt 000001EE34180250: 1300 tokens, checkpoints: 0, 121.891 MiB srv update: - prompt 000001EC4E9D5DA0: 1801 tokens, checkpoints: 0, 168.865 MiB srv update: - prompt 000001EE34180C50: 2559 tokens, checkpoints: 0, 239.937 MiB srv get_availabl: prompt cache update took 289.37 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 5029 | processing task, is_child = 0 slot operator(): id 0 | task 5029 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 1177 slot operator(): id 0 | task 5029 | cached n_tokens = 5, memory_seq_rm [5, end) [GIN] 2026/10/05 - 10:00:01 | 200 | 6.1657ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 5029 | prompt processing, n_tokens = 512, progress = 0.44, t = 8.95 s / 57.22 tokens per second slot operator(): id 0 | task 5029 | cached n_tokens = 517, memory_seq_rm [517, end) slot print_timing: id 0 | task 5029 | prompt processing, n_tokens = 1024, progress = 0.87, t = 28.44 s / 36.00 tokens per second slot operator(): id 0 | task 5029 | cached n_tokens = 1029, memory_seq_rm [1029, end) slot init_sampler: id 0 | task 5029 | init sampler, took 0.35 ms, tokens: text = 1177, total = 1177 [GIN] 2026/10/05 - 10:00:31 | 200 | 6.5075ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 5029 | n_gen = 100, tg = 8.25 t/s, tg_3s = 8.33 t/s slot print_timing: id 0 | task 5029 | n_gen = 129, tg = 8.51 t/s, tg_3s = 9.51 t/s slot print_timing: id 0 | task 5029 | n_gen = 140, tg = 7.52 t/s, tg_3s = 3.20 t/s slot print_timing: id 0 | task 5029 | prompt eval time = 34607.16 ms / 1172 tokens ( 29.53 ms per token, 33.87 tokens per second) slot print_timing: id 0 | task 5029 | eval time = 19215.31 ms / 152 tokens ( 127.25 ms per token, 7.86 tokens per second) slot print_timing: id 0 | task 5029 | total time = 53822.47 ms / 1324 tokens slot print_timing: id 0 | task 5029 | graphs reused = 5086 slot release: id 0 | task 5029 | stop processing: n_tokens = 1328, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 10:00:53 | 200 | 54.1695953s | 127.0.0.1 | POST "/api/chat" srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.002 (5/2147) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1683931349 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 1328, total state size = 124.516 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 1419, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 2581, lcp = 43, f_keep = 0.017, f_sim = 0.020 srv load: - prompt with length 1413, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 1304, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 1300, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 1801, lcp = 4, f_keep = 0.002, f_sim = 0.002 srv load: - prompt with length 2559, lcp = 43, f_keep = 0.017, f_sim = 0.020 srv load: - prompt with length 1328, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv update: - cache state: 8 prompts, 1285.010 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE40A30850: 1304 tokens, checkpoints: 0, 122.266 MiB srv update: - prompt 000001EE34180250: 1300 tokens, checkpoints: 0, 121.891 MiB srv update: - prompt 000001EC4E9D5DA0: 1801 tokens, checkpoints: 0, 168.865 MiB srv update: - prompt 000001EE34180C50: 2559 tokens, checkpoints: 0, 239.937 MiB srv update: - prompt 000001EE34181330: 1328 tokens, checkpoints: 0, 124.516 MiB srv get_availabl: prompt cache update took 1461.98 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 5184 | processing task, is_child = 0 slot operator(): id 0 | task 5184 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 2147 slot operator(): id 0 | task 5184 | cached n_tokens = 5, memory_seq_rm [5, end) [GIN] 2026/10/05 - 10:01:01 | 200 | 5.0172ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 5184 | prompt processing, n_tokens = 512, progress = 0.24, t = 13.37 s / 38.29 tokens per second slot operator(): id 0 | task 5184 | cached n_tokens = 517, memory_seq_rm [517, end) slot print_timing: id 0 | task 5184 | prompt processing, n_tokens = 1024, progress = 0.48, t = 18.50 s / 55.36 tokens per second slot operator(): id 0 | task 5184 | cached n_tokens = 1029, memory_seq_rm [1029, end) slot print_timing: id 0 | task 5184 | prompt processing, n_tokens = 1536, progress = 0.72, t = 30.53 s / 50.30 tokens per second slot operator(): id 0 | task 5184 | cached n_tokens = 1541, memory_seq_rm [1541, end) [GIN] 2026/10/05 - 10:01:31 | 200 | 8.5734ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 5184 | prompt processing, n_tokens = 2048, progress = 0.96, t = 51.55 s / 39.73 tokens per second slot operator(): id 0 | task 5184 | cached n_tokens = 2053, memory_seq_rm [2053, end) slot init_sampler: id 0 | task 5184 | init sampler, took 0.57 ms, tokens: text = 2147, total = 2147 slot print_timing: id 0 | task 5184 | prompt eval time = 59349.18 ms / 2142 tokens ( 27.71 ms per token, 36.09 tokens per second) slot print_timing: id 0 | task 5184 | eval time = 3330.24 ms / 64 tokens ( 52.86 ms per token, 18.92 tokens per second) slot print_timing: id 0 | task 5184 | total time = 62679.42 ms / 2206 tokens slot print_timing: id 0 | task 5184 | graphs reused = 5148 [GIN] 2026/10/05 - 10:01:58 | 200 | 1m4s | 127.0.0.1 | POST "/api/chat" slot release: id 0 | task 5184 | stop processing: n_tokens = 2210, truncated = 0 srv update_slots: all slots are idle srv server_strea: conv_id= (empty=1) slot get_availabl: id 0 | task -1 | - checking sim = 0.025 (61/2422) > 0.100 slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 1748275744 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 2210, total state size = 207.214 MiB (draft: 0.000 MiB) srv load: - looking for better prompt, base f_keep = 0.028, f_sim = 0.025 srv load: - prompt with length 1419, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 2581, lcp = 43, f_keep = 0.017, f_sim = 0.018 srv load: - prompt with length 1413, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 1304, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 1300, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 1801, lcp = 4, f_keep = 0.002, f_sim = 0.002 srv load: - prompt with length 2559, lcp = 43, f_keep = 0.017, f_sim = 0.018 srv load: - prompt with length 1328, lcp = 5, f_keep = 0.004, f_sim = 0.002 srv load: - prompt with length 2210, lcp = 61, f_keep = 0.028, f_sim = 0.025 srv update: - cache state: 9 prompts, 1492.223 MiB (limits: 8192.000 MiB, 16384 tokens, 87370 est) srv update: - prompt 000001EC53D688D0: 1419 tokens, checkpoints: 0, 133.049 MiB srv update: - prompt 000001EE34252960: 2581 tokens, checkpoints: 0, 241.999 MiB srv update: - prompt 000001EE342530E0: 1413 tokens, checkpoints: 0, 132.486 MiB srv update: - prompt 000001EE40A30850: 1304 tokens, checkpoints: 0, 122.266 MiB srv update: - prompt 000001EE34180250: 1300 tokens, checkpoints: 0, 121.891 MiB srv update: - prompt 000001EC4E9D5DA0: 1801 tokens, checkpoints: 0, 168.865 MiB srv update: - prompt 000001EE34180C50: 2559 tokens, checkpoints: 0, 239.937 MiB srv update: - prompt 000001EE34181330: 1328 tokens, checkpoints: 0, 124.516 MiB srv update: - prompt 000001EC4EA3EBB0: 2210 tokens, checkpoints: 0, 207.214 MiB srv get_availabl: prompt cache update took 225.18 ms slot launch_slot_: id 0 | task -1 | sampler chain: logits -> penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist slot launch_slot_: id 0 | task -1 | sampler params: repeat_last_n = 64, repeat_penalty = 1.050, frequency_penalty = 0.000, presence_penalty = 0.000 dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64 top_k = 20, top_p = 0.800, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900 slot launch_slot_: id 0 | task 5253 | processing task, is_child = 0 slot operator(): id 0 | task 5253 | new prompt, n_ctx_slot = 16384, n_keep = 4, task.n_tokens = 2422 slot operator(): id 0 | task 5253 | cached n_tokens = 61, memory_seq_rm [61, end) [GIN] 2026/10/05 - 10:02:01 | 200 | 8.4997ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 5253 | prompt processing, n_tokens = 512, progress = 0.24, t = 5.58 s / 91.79 tokens per second slot operator(): id 0 | task 5253 | cached n_tokens = 573, memory_seq_rm [573, end) slot print_timing: id 0 | task 5253 | prompt processing, n_tokens = 1024, progress = 0.45, t = 14.42 s / 71.00 tokens per second slot operator(): id 0 | task 5253 | cached n_tokens = 1085, memory_seq_rm [1085, end) slot print_timing: id 0 | task 5253 | prompt processing, n_tokens = 1536, progress = 0.66, t = 26.12 s / 58.81 tokens per second slot operator(): id 0 | task 5253 | cached n_tokens = 1597, memory_seq_rm [1597, end) [GIN] 2026/10/05 - 10:02:32 | 200 | 83.6372ms | 127.0.0.1 | GET "/api/tags" slot print_timing: id 0 | task 5253 | prompt processing, n_tokens = 2048, progress = 0.87, t = 34.86 s / 58.75 tokens per second slot operator(): id 0 | task 5253 | cached n_tokens = 2109, memory_seq_rm [2109, end) slot init_sampler: id 0 | task 5253 | init sampler, took 0.64 ms, tokens: text = 2422, total = 2422 slot print_timing: id 0 | task 5253 | n_gen = 100, tg = 10.40 t/s, tg_3s = 10.50 t/s [GIN] 2026/10/05 - 10:02:51 | 200 | 53.6479506s | 127.0.0.1 | POST "/api/chat" srv stop: cancel task, id_task = 5253 slot release: id 0 | task 5253 | stop processing: n_tokens = 2546, truncated = 0 srv update_slots: all slots are idle [GIN] 2026/10/05 - 10:03:03 | 200 | 889.3398ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:03:33 | 200 | 11.0292ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:04:03 | 200 | 5.4677ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:04:33 | 200 | 7.5866ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:05:03 | 200 | 4.9995ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:05:34 | 200 | 713.5056ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:06:04 | 200 | 114.6115ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:06:34 | 200 | 3.5356ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:07:04 | 200 | 12.0147ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:07:34 | 200 | 148.4639ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:08:04 | 200 | 7.0025ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:08:34 | 200 | 8.5083ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:09:04 | 200 | 3.0002ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:09:34 | 200 | 5.9939ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:10:04 | 200 | 3.4991ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:10:34 | 200 | 5.0073ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:11:04 | 200 | 3.5005ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:11:34 | 200 | 12.5355ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:12:04 | 200 | 3.5026ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:12:34 | 200 | 3.0064ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:13:04 | 200 | 4.0115ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:13:34 | 200 | 4.415ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:14:04 | 200 | 2.9989ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:14:34 | 200 | 6.0119ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:15:04 | 200 | 4.9941ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:15:35 | 200 | 178.0968ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:16:05 | 200 | 10.4766ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:16:35 | 200 | 3.5313ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:17:05 | 200 | 3.0129ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:17:35 | 200 | 4.8663ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:18:05 | 200 | 7.5074ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:18:35 | 200 | 4.0181ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:19:05 | 200 | 3.5ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:19:35 | 200 | 4.5196ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:20:05 | 200 | 45.5421ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:20:35 | 200 | 42.045ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:21:05 | 200 | 3.5934ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:21:35 | 200 | 10.0207ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:22:05 | 200 | 9.9995ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:22:35 | 200 | 210.6627ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:23:05 | 200 | 17.508ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:23:35 | 200 | 3.9999ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:24:05 | 200 | 12.0216ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:24:36 | 200 | 38.0291ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:25:06 | 200 | 4.0077ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:25:36 | 200 | 3.612ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:26:06 | 200 | 2.9999ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:26:36 | 200 | 4.5031ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:27:06 | 200 | 5.5005ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:27:36 | 200 | 4.0083ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:28:06 | 200 | 3.9985ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:28:36 | 200 | 4.4996ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:29:06 | 200 | 3.5083ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:29:36 | 200 | 6.041ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:30:06 | 200 | 4.0112ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:30:36 | 200 | 3.393ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:31:06 | 200 | 5.0937ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:31:36 | 200 | 31.5873ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:32:06 | 200 | 4.5045ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:32:36 | 200 | 3.4998ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:33:06 | 200 | 3.8543ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:33:36 | 200 | 4.0361ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:34:06 | 200 | 3.3663ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:34:36 | 200 | 3.0398ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:35:06 | 200 | 3.8636ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:35:36 | 200 | 3.0583ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:36:06 | 200 | 6.6848ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:36:36 | 200 | 4.0084ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:37:06 | 200 | 4.7544ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:37:36 | 200 | 4.9303ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:38:06 | 200 | 3.4996ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:38:37 | 200 | 4.4515ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:39:07 | 200 | 5.0044ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:39:37 | 200 | 12.3629ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:40:07 | 200 | 4.999ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:40:37 | 200 | 3.5077ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:41:07 | 200 | 4.5037ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:41:37 | 200 | 3.5844ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:42:07 | 200 | 3.0001ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:42:37 | 200 | 5.2904ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:43:07 | 200 | 3.9992ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:43:37 | 200 | 9.0152ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:44:07 | 200 | 3.4039ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:44:37 | 200 | 2.9976ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:45:08 | 200 | 4.4997ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:45:38 | 200 | 5.0055ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:46:08 | 200 | 5.5032ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:46:38 | 200 | 3.5ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:47:08 | 200 | 3.5082ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:47:38 | 200 | 3.5625ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:48:08 | 200 | 3.5046ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:48:38 | 200 | 3.4996ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:49:08 | 200 | 4.0032ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:49:38 | 200 | 3.999ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:50:08 | 200 | 5.5005ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:50:38 | 200 | 5.3171ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:51:08 | 200 | 4.5039ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:51:38 | 200 | 4.0038ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:52:08 | 200 | 3ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:52:38 | 200 | 2.9999ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/10/05 - 10:53:08 | 200 | 331.1531ms | 127.0.0.1 | GET "/api/tags"