ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 1 CUDA devices: Device 0: NVIDIA GeForce RTX 5090, compute capability 12.0, VMM: yes main: setting n_parallel = 4 and kv_unified = true (add -kvu to disable this) build: 7212 (ff90508d6) with cc (Ubuntu 11.4.0-1ubuntu1~22.04.2) 11.4.0 for x86_64-linux-gnu system info: n_threads = 8, n_threads_batch = 8, total_threads = 32 system_info: n_threads = 8 (n_threads_batch = 8) / 32 | CUDA : ARCHS = 890 | USE_GRAPHS = 1 | PEER_MAX_BATCH_SIZE = 128 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX_VNNI = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | LLAMAFILE = 1 | OPENMP = 1 | REPACK = 1 | init: using 31 threads for HTTP server start: binding port with default address family main: loading model srv load_model: loading model '/home/iscas/models/Qwen2-VL-2B-Instruct-GGUF/Qwen2-VL-2B-Instruct-Q8_0.gguf' llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce RTX 5090) (0000:01:00.0) - 29667 MiB free llama_model_loader: loaded meta data with 33 key-value pairs and 338 tensors from /home/iscas/models/Qwen2-VL-2B-Instruct-GGUF/Qwen2-VL-2B-Instruct-Q8_0.gguf (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 = qwen2vl llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = Qwen2 VL 2B Instruct llama_model_loader: - kv 3: general.finetune str = Instruct llama_model_loader: - kv 4: general.basename str = Qwen2-VL llama_model_loader: - kv 5: general.size_label str = 2B llama_model_loader: - kv 6: general.license str = apache-2.0 llama_model_loader: - kv 7: general.base_model.count u32 = 1 llama_model_loader: - kv 8: general.base_model.0.name str = Qwen2 VL 2B Instruct llama_model_loader: - kv 9: general.base_model.0.organization str = Qwen llama_model_loader: - kv 10: general.base_model.0.repo_url str = https://huggingface.co/qwen/Qwen2-VL-... llama_model_loader: - kv 11: general.tags arr[str,2] = ["multimodal", "image-text-to-text"] llama_model_loader: - kv 12: general.languages arr[str,1] = ["en"] llama_model_loader: - kv 13: qwen2vl.block_count u32 = 28 llama_model_loader: - kv 14: qwen2vl.context_length u32 = 32768 llama_model_loader: - kv 15: qwen2vl.embedding_length u32 = 1536 llama_model_loader: - kv 16: qwen2vl.feed_forward_length u32 = 8960 llama_model_loader: - kv 17: qwen2vl.attention.head_count u32 = 12 llama_model_loader: - kv 18: qwen2vl.attention.head_count_kv u32 = 2 llama_model_loader: - kv 19: qwen2vl.rope.freq_base f32 = 1000000.000000 llama_model_loader: - kv 20: qwen2vl.attention.layer_norm_rms_epsilon f32 = 0.000001 llama_model_loader: - kv 21: general.file_type u32 = 7 llama_model_loader: - kv 22: qwen2vl.rope.dimension_sections arr[i32,4] = [16, 24, 24, 0] llama_model_loader: - kv 23: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 24: tokenizer.ggml.pre str = qwen2 llama_model_loader: - kv 25: tokenizer.ggml.tokens arr[str,151936] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 26: tokenizer.ggml.token_type arr[i32,151936] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 27: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... llama_model_loader: - kv 28: tokenizer.ggml.eos_token_id u32 = 151645 llama_model_loader: - kv 29: tokenizer.ggml.padding_token_id u32 = 151643 llama_model_loader: - kv 30: tokenizer.ggml.bos_token_id u32 = 151643 llama_model_loader: - kv 31: tokenizer.chat_template str = {% set image_count = namespace(value=... llama_model_loader: - kv 32: general.quantization_version u32 = 2 llama_model_loader: - type f32: 141 tensors llama_model_loader: - type q8_0: 197 tensors print_info: file format = GGUF V3 (latest) print_info: file type = Q8_0 print_info: file size = 1.53 GiB (8.50 BPW) load: printing all EOG tokens: load: - 151643 ('<|endoftext|>') load: - 151645 ('<|im_end|>') load: special tokens cache size = 14 load: token to piece cache size = 0.9309 MB print_info: arch = qwen2vl print_info: vocab_only = 0 print_info: n_ctx_train = 32768 print_info: n_embd = 1536 print_info: n_embd_inp = 1536 print_info: n_layer = 28 print_info: n_head = 12 print_info: n_head_kv = 2 print_info: n_rot = 128 print_info: n_swa = 0 print_info: is_swa_any = 0 print_info: n_embd_head_k = 128 print_info: n_embd_head_v = 128 print_info: n_gqa = 6 print_info: n_embd_k_gqa = 256 print_info: n_embd_v_gqa = 256 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: n_ff = 8960 print_info: n_expert = 0 print_info: n_expert_used = 0 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 = 8 print_info: rope scaling = linear print_info: freq_base_train = 1000000.0 print_info: freq_scale_train = 1 print_info: n_ctx_orig_yarn = 32768 print_info: rope_finetuned = unknown print_info: mrope sections = [16, 24, 24, 0] print_info: model type = 1.5B print_info: model params = 1.54 B print_info: general.name = Qwen2 VL 2B Instruct print_info: vocab type = BPE print_info: n_vocab = 151936 print_info: n_merges = 151387 print_info: BOS token = 151643 '<|endoftext|>' 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: EOG token = 151643 '<|endoftext|>' print_info: EOG token = 151645 '<|im_end|>' print_info: max token length = 256 load_tensors: loading model tensors, this can take a while... (mmap = true) load_tensors: offloading 28 repeating layers to GPU load_tensors: offloading output layer to GPU load_tensors: offloaded 29/29 layers to GPU load_tensors: CPU_Mapped model buffer size = 236.47 MiB load_tensors: CUDA0 model buffer size = 1564.63 MiB ............................................................................ llama_context: constructing llama_context llama_context: n_seq_max = 4 llama_context: n_ctx = 8192 llama_context: n_ctx_seq = 8192 llama_context: n_batch = 2048 llama_context: n_ubatch = 512 llama_context: causal_attn = 1 llama_context: flash_attn = auto llama_context: kv_unified = true llama_context: freq_base = 1000000.0 llama_context: freq_scale = 1 llama_context: n_ctx_seq (8192) < n_ctx_train (32768) -- the full capacity of the model will not be utilized llama_context: CUDA_Host output buffer size = 2.32 MiB llama_kv_cache: CUDA0 KV buffer size = 224.00 MiB llama_kv_cache: size = 224.00 MiB ( 8192 cells, 28 layers, 4/1 seqs), K (f16): 112.00 MiB, V (f16): 112.00 MiB llama_context: Flash Attention was auto, set to enabled llama_context: CUDA0 compute buffer size = 321.77 MiB llama_context: CUDA_Host compute buffer size = 19.02 MiB llama_context: graph nodes = 959 llama_context: graph splits = 2 common_init_from_params: added <|endoftext|> logit bias = -inf common_init_from_params: added <|im_end|> logit bias = -inf common_init_from_params: setting dry_penalty_last_n to ctx_size = 8192 common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable) srv init: initializing slots, n_slots = 4 slot init: id 0 | task -1 | new slot, n_ctx = 8192 slot init: id 1 | task -1 | new slot, n_ctx = 8192 slot init: id 2 | task -1 | new slot, n_ctx = 8192 slot init: id 3 | task -1 | new slot, n_ctx = 8192 srv init: prompt cache is enabled, size limit: 8192 MiB srv init: use `--cache-ram 0` to disable the prompt cache srv init: for more info see https://github.com/ggml-org/llama.cpp/pull/16391 srv init: thinking = 0 init: chat template, chat_template: {% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system You are a helpful assistant.<|im_end|> {% endif %}<|im_start|>{{ message['role'] }} {% if message['content'] is string %}{{ message['content'] }}<|im_end|> {% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|> {% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant {% endif %}, 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 ' main: model loaded main: server is listening on http://0.0.0.0:8081 main: starting the main loop... srv update_slots: all slots are idle srv log_server_r: request: GET /health 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LRU, t_last = -1 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 0 | processing task slot update_slots: id 3 | task 0 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 362 slot update_slots: id 3 | task 0 | n_tokens = 0, memory_seq_rm [0, end) slot update_slots: id 3 | task 0 | prompt processing progress, n_tokens = 362, batch.n_tokens = 362, progress = 1.000000 slot update_slots: id 3 | task 0 | prompt done, n_tokens = 362, batch.n_tokens = 362 slot print_timing: id 3 | task 0 | prompt eval time = 31.05 ms / 362 tokens ( 0.09 ms per token, 11658.62 tokens per second) eval time = 20.25 ms / 6 tokens ( 3.37 ms per token, 296.34 tokens per second) total time = 51.30 ms / 368 tokens slot release: id 3 | task 0 | stop processing: n_tokens = 367, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 2 | task -1 | selected slot by LRU, t_last = -1 slot launch_slot_: id 2 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 2 | task 7 | processing task slot update_slots: id 2 | task 7 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2341 slot update_slots: id 2 | task 7 | n_tokens = 0, memory_seq_rm [0, end) slot update_slots: id 2 | task 7 | prompt processing progress, n_tokens = 2048, batch.n_tokens = 2048, progress = 0.874840 slot update_slots: id 2 | task 7 | n_tokens = 2048, memory_seq_rm [2048, end) slot update_slots: id 2 | task 7 | prompt processing progress, n_tokens = 2341, batch.n_tokens = 293, progress = 1.000000 slot update_slots: id 2 | task 7 | prompt done, n_tokens = 2341, batch.n_tokens = 293 slot print_timing: id 2 | task 7 | prompt eval time = 67.84 ms / 2341 tokens ( 0.03 ms per token, 34508.17 tokens per second) eval time = 51.78 ms / 21 tokens ( 2.47 ms per token, 405.53 tokens per second) total time = 119.62 ms / 2362 tokens slot release: id 2 | task 7 | stop processing: n_tokens = 2361, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.967 (> 0.100 thold), f_keep = 0.973 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 30 | processing task slot update_slots: id 3 | task 30 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 369 slot update_slots: id 3 | task 30 | n_tokens = 357, memory_seq_rm [357, end) slot update_slots: id 3 | task 30 | prompt processing progress, n_tokens = 369, batch.n_tokens = 12, progress = 1.000000 slot update_slots: id 3 | task 30 | prompt done, n_tokens = 369, batch.n_tokens = 12 slot print_timing: id 3 | task 30 | prompt eval time = 22.50 ms / 12 tokens ( 1.87 ms per token, 533.40 tokens per second) eval time = 14.04 ms / 6 tokens ( 2.34 ms per token, 427.35 tokens per second) total time = 36.54 ms / 18 tokens slot release: id 3 | task 30 | stop processing: n_tokens = 374, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 2 | task -1 | selected slot by LCP similarity, sim_best = 0.925 (> 0.100 thold), f_keep = 0.920 slot launch_slot_: id 2 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 2 | task 37 | processing task slot update_slots: id 2 | task 37 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2348 slot update_slots: id 2 | task 37 | n_tokens = 2172, memory_seq_rm [2172, end) slot update_slots: id 2 | task 37 | prompt processing progress, n_tokens = 2348, batch.n_tokens = 176, progress = 1.000000 slot update_slots: id 2 | task 37 | prompt done, n_tokens = 2348, batch.n_tokens = 176 slot print_timing: id 2 | task 37 | prompt eval time = 11.77 ms / 176 tokens ( 0.07 ms per token, 14955.81 tokens per second) eval time = 47.92 ms / 19 tokens ( 2.52 ms per token, 396.52 tokens per second) total time = 59.69 ms / 195 tokens slot release: id 2 | task 37 | stop processing: n_tokens = 2366, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.976 (> 0.100 thold), f_keep = 0.971 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 57 | processing task slot update_slots: id 3 | task 57 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 372 slot update_slots: id 3 | task 57 | n_tokens = 363, memory_seq_rm [363, end) slot update_slots: id 3 | task 57 | prompt processing progress, n_tokens = 372, batch.n_tokens = 9, progress = 1.000000 slot update_slots: id 3 | task 57 | prompt done, n_tokens = 372, batch.n_tokens = 9 slot print_timing: id 3 | task 57 | prompt eval time = 9.52 ms / 9 tokens ( 1.06 ms per token, 945.48 tokens per second) eval time = 17.07 ms / 6 tokens ( 2.84 ms per token, 351.60 tokens per second) total time = 26.58 ms / 15 tokens slot release: id 3 | task 57 | stop processing: n_tokens = 377, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 2 | task -1 | selected slot by LCP similarity, sim_best = 0.926 (> 0.100 thold), f_keep = 0.921 slot launch_slot_: id 2 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 2 | task 64 | processing task slot update_slots: id 2 | task 64 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2351 slot update_slots: id 2 | task 64 | n_tokens = 2178, memory_seq_rm [2178, end) slot update_slots: id 2 | task 64 | prompt processing progress, n_tokens = 2351, batch.n_tokens = 173, progress = 1.000000 slot update_slots: id 2 | task 64 | prompt done, n_tokens = 2351, batch.n_tokens = 173 slot print_timing: id 2 | task 64 | prompt eval time = 11.31 ms / 173 tokens ( 0.07 ms per token, 15292.14 tokens per second) eval time = 138.40 ms / 55 tokens ( 2.52 ms per token, 397.40 tokens per second) total time = 149.71 ms / 228 tokens slot release: id 2 | task 64 | stop processing: n_tokens = 2405, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.976 (> 0.100 thold), f_keep = 0.963 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 120 | processing task slot update_slots: id 3 | task 120 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 372 slot update_slots: id 3 | task 120 | n_tokens = 363, memory_seq_rm [363, end) slot update_slots: id 3 | task 120 | prompt processing progress, n_tokens = 372, batch.n_tokens = 9, progress = 1.000000 slot update_slots: id 3 | task 120 | prompt done, n_tokens = 372, batch.n_tokens = 9 slot print_timing: id 3 | task 120 | prompt eval time = 8.70 ms / 9 tokens ( 0.97 ms per token, 1034.60 tokens per second) eval time = 14.44 ms / 6 tokens ( 2.41 ms per token, 415.45 tokens per second) total time = 23.14 ms / 15 tokens slot release: id 3 | task 120 | stop processing: n_tokens = 377, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 2 | task -1 | selected slot by LCP similarity, sim_best = 0.926 (> 0.100 thold), f_keep = 0.906 slot launch_slot_: id 2 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 2 | task 127 | processing task slot update_slots: id 2 | task 127 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2351 slot update_slots: id 2 | task 127 | n_tokens = 2178, memory_seq_rm [2178, end) slot update_slots: id 2 | task 127 | prompt processing progress, n_tokens = 2351, batch.n_tokens = 173, progress = 1.000000 slot update_slots: id 2 | task 127 | prompt done, n_tokens = 2351, batch.n_tokens = 173 slot print_timing: id 2 | task 127 | prompt eval time = 11.21 ms / 173 tokens ( 0.06 ms per token, 15427.14 tokens per second) eval time = 126.01 ms / 44 tokens ( 2.86 ms per token, 349.18 tokens per second) total time = 137.22 ms / 217 tokens slot release: id 2 | task 127 | stop processing: n_tokens = 2394, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.976 (> 0.100 thold), f_keep = 0.963 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 172 | processing task slot update_slots: id 3 | task 172 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 372 slot update_slots: id 3 | task 172 | n_tokens = 363, memory_seq_rm [363, end) slot update_slots: id 3 | task 172 | prompt processing progress, n_tokens = 372, batch.n_tokens = 9, progress = 1.000000 slot update_slots: id 3 | task 172 | prompt done, n_tokens = 372, batch.n_tokens = 9 slot print_timing: id 3 | task 172 | prompt eval time = 10.32 ms / 9 tokens ( 1.15 ms per token, 872.09 tokens per second) eval time = 14.42 ms / 6 tokens ( 2.40 ms per token, 416.20 tokens per second) total time = 24.74 ms / 15 tokens slot release: id 3 | task 172 | stop processing: n_tokens = 377, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LRU, t_last = -1 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 179 | processing task slot update_slots: id 1 | task 179 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2773 slot update_slots: id 1 | task 179 | n_tokens = 0, memory_seq_rm [0, end) slot update_slots: id 1 | task 179 | prompt processing progress, n_tokens = 2048, batch.n_tokens = 2048, progress = 0.738550 slot update_slots: id 1 | task 179 | n_tokens = 2048, memory_seq_rm [2048, end) slot update_slots: id 1 | task 179 | prompt processing progress, n_tokens = 2773, batch.n_tokens = 725, progress = 1.000000 slot update_slots: id 1 | task 179 | prompt done, n_tokens = 2773, batch.n_tokens = 725 slot print_timing: id 1 | task 179 | prompt eval time = 95.75 ms / 2773 tokens ( 0.03 ms per token, 28961.44 tokens per second) eval time = 973.64 ms / 339 tokens ( 2.87 ms per token, 348.18 tokens per second) total time = 1069.39 ms / 3112 tokens slot release: id 1 | task 179 | stop processing: n_tokens = 3111, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.891 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 520 | processing task slot update_slots: id 1 | task 520 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2773 slot update_slots: id 1 | task 520 | need to evaluate at least 1 token for each active slot (n_past = 2773, task.n_tokens() = 2773) slot update_slots: id 1 | task 520 | n_past was set to 2772 slot update_slots: id 1 | task 520 | n_tokens = 2772, memory_seq_rm [2772, end) slot update_slots: id 1 | task 520 | prompt processing progress, n_tokens = 2773, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 520 | prompt done, n_tokens = 2773, batch.n_tokens = 1 slot print_timing: id 1 | task 520 | prompt eval time = 11.48 ms / 1 tokens ( 11.48 ms per token, 87.12 tokens per second) eval time = 1276.92 ms / 429 tokens ( 2.98 ms per token, 335.96 tokens per second) total time = 1288.40 ms / 430 tokens slot release: id 1 | task 520 | stop processing: n_tokens = 3201, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.866 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 950 | processing task slot update_slots: id 1 | task 950 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2773 slot update_slots: id 1 | task 950 | need to evaluate at least 1 token for each active slot (n_past = 2773, task.n_tokens() = 2773) slot update_slots: id 1 | task 950 | n_past was set to 2772 slot update_slots: id 1 | task 950 | n_tokens = 2772, memory_seq_rm [2772, end) slot update_slots: id 1 | task 950 | prompt processing progress, n_tokens = 2773, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 950 | prompt done, n_tokens = 2773, batch.n_tokens = 1 slot print_timing: id 1 | task 950 | prompt eval time = 11.41 ms / 1 tokens ( 11.41 ms per token, 87.67 tokens per second) eval time = 1014.58 ms / 365 tokens ( 2.78 ms per token, 359.75 tokens per second) total time = 1025.99 ms / 366 tokens slot release: id 1 | task 950 | stop processing: n_tokens = 3137, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.962 (> 0.100 thold), f_keep = 0.944 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 1316 | processing task slot update_slots: id 3 | task 1316 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 370 slot update_slots: id 3 | task 1316 | n_tokens = 356, memory_seq_rm [356, end) slot update_slots: id 3 | task 1316 | prompt processing progress, n_tokens = 370, batch.n_tokens = 14, progress = 1.000000 slot update_slots: id 3 | task 1316 | prompt done, n_tokens = 370, batch.n_tokens = 14 slot print_timing: id 3 | task 1316 | prompt eval time = 7.90 ms / 14 tokens ( 0.56 ms per token, 1772.15 tokens per second) eval time = 14.43 ms / 6 tokens ( 2.40 ms per token, 415.94 tokens per second) total time = 22.33 ms / 20 tokens slot release: id 3 | task 1316 | stop processing: n_tokens = 375, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 0.936 (> 0.100 thold), f_keep = 0.827 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 1323 | processing task slot update_slots: id 1 | task 1323 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2771 slot update_slots: id 1 | task 1323 | n_tokens = 2593, memory_seq_rm [2593, end) slot update_slots: id 1 | task 1323 | prompt processing progress, n_tokens = 2771, batch.n_tokens = 178, progress = 1.000000 slot update_slots: id 1 | task 1323 | prompt done, n_tokens = 2771, batch.n_tokens = 178 slot print_timing: id 1 | task 1323 | prompt eval time = 20.14 ms / 178 tokens ( 0.11 ms per token, 8838.57 tokens per second) eval time = 939.10 ms / 326 tokens ( 2.88 ms per token, 347.14 tokens per second) total time = 959.24 ms / 504 tokens slot release: id 1 | task 1323 | stop processing: n_tokens = 3096, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.895 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 1650 | processing task slot update_slots: id 1 | task 1650 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2771 slot update_slots: id 1 | task 1650 | need to evaluate at least 1 token for each active slot (n_past = 2771, task.n_tokens() = 2771) slot update_slots: id 1 | task 1650 | n_past was set to 2770 slot update_slots: id 1 | task 1650 | n_tokens = 2770, memory_seq_rm [2770, end) slot update_slots: id 1 | task 1650 | prompt processing progress, n_tokens = 2771, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 1650 | prompt done, n_tokens = 2771, batch.n_tokens = 1 slot print_timing: id 1 | task 1650 | prompt eval time = 11.59 ms / 1 tokens ( 11.59 ms per token, 86.30 tokens per second) eval time = 1012.17 ms / 352 tokens ( 2.88 ms per token, 347.77 tokens per second) total time = 1023.76 ms / 353 tokens slot release: id 1 | task 1650 | stop processing: n_tokens = 3122, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.888 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 2003 | processing task slot update_slots: id 1 | task 2003 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2771 slot update_slots: id 1 | task 2003 | need to evaluate at least 1 token for each active slot (n_past = 2771, task.n_tokens() = 2771) slot update_slots: id 1 | task 2003 | n_past was set to 2770 slot update_slots: id 1 | task 2003 | n_tokens = 2770, memory_seq_rm [2770, end) slot update_slots: id 1 | task 2003 | prompt processing progress, n_tokens = 2771, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 2003 | prompt done, n_tokens = 2771, batch.n_tokens = 1 slot print_timing: id 1 | task 2003 | prompt eval time = 11.50 ms / 1 tokens ( 11.50 ms per token, 86.96 tokens per second) eval time = 1323.31 ms / 455 tokens ( 2.91 ms per token, 343.83 tokens per second) total time = 1334.81 ms / 456 tokens slot release: id 1 | task 2003 | stop processing: n_tokens = 3225, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.976 (> 0.100 thold), f_keep = 0.968 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 2459 | processing task slot update_slots: id 3 | task 2459 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 372 slot update_slots: id 3 | task 2459 | n_tokens = 363, memory_seq_rm [363, end) slot update_slots: id 3 | task 2459 | prompt processing progress, n_tokens = 372, batch.n_tokens = 9, progress = 1.000000 slot update_slots: id 3 | task 2459 | prompt done, n_tokens = 372, batch.n_tokens = 9 slot print_timing: id 3 | task 2459 | prompt eval time = 7.76 ms / 9 tokens ( 0.86 ms per token, 1159.20 tokens per second) eval time = 18.54 ms / 6 tokens ( 3.09 ms per token, 323.55 tokens per second) total time = 26.31 ms / 15 tokens slot release: id 3 | task 2459 | stop processing: n_tokens = 377, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 0.938 (> 0.100 thold), f_keep = 0.806 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 2466 | processing task slot update_slots: id 1 | task 2466 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2773 slot update_slots: id 1 | task 2466 | n_tokens = 2600, memory_seq_rm [2600, end) slot update_slots: id 1 | task 2466 | prompt processing progress, n_tokens = 2773, batch.n_tokens = 173, progress = 1.000000 slot update_slots: id 1 | task 2466 | prompt done, n_tokens = 2773, batch.n_tokens = 173 slot print_timing: id 1 | task 2466 | prompt eval time = 20.09 ms / 173 tokens ( 0.12 ms per token, 8611.25 tokens per second) eval time = 1158.57 ms / 408 tokens ( 2.84 ms per token, 352.16 tokens per second) total time = 1178.66 ms / 581 tokens slot release: id 1 | task 2466 | stop processing: n_tokens = 3180, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.872 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 2875 | processing task slot update_slots: id 1 | task 2875 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2773 slot update_slots: id 1 | task 2875 | need to evaluate at least 1 token for each active slot (n_past = 2773, task.n_tokens() = 2773) slot update_slots: id 1 | task 2875 | n_past was set to 2772 slot update_slots: id 1 | task 2875 | n_tokens = 2772, memory_seq_rm [2772, end) slot update_slots: id 1 | task 2875 | prompt processing progress, n_tokens = 2773, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 2875 | prompt done, n_tokens = 2773, batch.n_tokens = 1 slot print_timing: id 1 | task 2875 | prompt eval time = 11.71 ms / 1 tokens ( 11.71 ms per token, 85.36 tokens per second) eval time = 1132.82 ms / 400 tokens ( 2.83 ms per token, 353.10 tokens per second) total time = 1144.53 ms / 401 tokens slot release: id 1 | task 2875 | stop processing: n_tokens = 3172, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.874 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 3276 | processing task slot update_slots: id 1 | task 3276 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2773 slot update_slots: id 1 | task 3276 | need to evaluate at least 1 token for each active slot (n_past = 2773, task.n_tokens() = 2773) slot update_slots: id 1 | task 3276 | n_past was set to 2772 slot update_slots: id 1 | task 3276 | n_tokens = 2772, memory_seq_rm [2772, end) slot update_slots: id 1 | task 3276 | prompt processing progress, n_tokens = 2773, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 3276 | prompt done, n_tokens = 2773, batch.n_tokens = 1 slot print_timing: id 1 | task 3276 | prompt eval time = 12.11 ms / 1 tokens ( 12.11 ms per token, 82.56 tokens per second) eval time = 888.74 ms / 298 tokens ( 2.98 ms per token, 335.31 tokens per second) total time = 900.85 ms / 299 tokens slot release: id 1 | task 3276 | stop processing: n_tokens = 3070, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.952 (> 0.100 thold), f_keep = 0.944 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 3575 | processing task slot update_slots: id 3 | task 3575 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 374 slot update_slots: id 3 | task 3575 | n_tokens = 356, memory_seq_rm [356, end) slot update_slots: id 3 | task 3575 | prompt processing progress, n_tokens = 374, batch.n_tokens = 18, progress = 1.000000 slot update_slots: id 3 | task 3575 | prompt done, n_tokens = 374, batch.n_tokens = 18 slot print_timing: id 3 | task 3575 | prompt eval time = 12.58 ms / 18 tokens ( 0.70 ms per token, 1431.30 tokens per second) eval time = 20.22 ms / 6 tokens ( 3.37 ms per token, 296.78 tokens per second) total time = 32.79 ms / 24 tokens slot release: id 3 | task 3575 | stop processing: n_tokens = 379, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 0.934 (> 0.100 thold), f_keep = 0.845 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 3582 | processing task slot update_slots: id 1 | task 3582 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2775 slot update_slots: id 1 | task 3582 | n_tokens = 2593, memory_seq_rm [2593, end) slot update_slots: id 1 | task 3582 | prompt processing progress, n_tokens = 2775, batch.n_tokens = 182, progress = 1.000000 slot update_slots: id 1 | task 3582 | prompt done, n_tokens = 2775, batch.n_tokens = 182 slot print_timing: id 1 | task 3582 | prompt eval time = 21.34 ms / 182 tokens ( 0.12 ms per token, 8527.79 tokens per second) eval time = 748.01 ms / 253 tokens ( 2.96 ms per token, 338.23 tokens per second) total time = 769.35 ms / 435 tokens slot release: id 1 | task 3582 | stop processing: n_tokens = 3027, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.917 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 3836 | processing task slot update_slots: id 1 | task 3836 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2775 slot update_slots: id 1 | task 3836 | need to evaluate at least 1 token for each active slot (n_past = 2775, task.n_tokens() = 2775) slot update_slots: id 1 | task 3836 | n_past was set to 2774 slot update_slots: id 1 | task 3836 | n_tokens = 2774, memory_seq_rm [2774, end) slot update_slots: id 1 | task 3836 | prompt processing progress, n_tokens = 2775, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 3836 | prompt done, n_tokens = 2775, batch.n_tokens = 1 decode: failed to find a memory slot for batch of size 1 srv try_clear_id: purging slot 2 with 2394 tokens slot clear_slot: id 2 | task -1 | clearing slot with 2394 tokens srv update_slots: failed to find free space in the KV cache, retrying with smaller batch size, i = 0, n_batch = 2048, ret = 1 decode: failed to find a memory slot for batch of size 1 srv try_clear_id: purging slot 3 with 379 tokens slot clear_slot: id 3 | task -1 | clearing slot with 379 tokens srv update_slots: failed to find free space in the KV cache, retrying with smaller batch size, i = 0, n_batch = 2048, ret = 1 slot print_timing: id 1 | task 3836 | prompt eval time = 11.81 ms / 1 tokens ( 11.81 ms per token, 84.68 tokens per second) eval time = 16454.85 ms / 5417 tokens ( 3.04 ms per token, 329.20 tokens per second) total time = 16466.66 ms / 5418 tokens slot release: id 1 | task 3836 | stop processing: n_tokens = 8191, truncated = 1 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.339 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 8191, total state size = 224.067 MiB srv load: - looking for better prompt, base f_keep = 0.339, sim = 1.000 srv update: - cache state: 1 prompts, 224.067 MiB (limits: 8192.000 MiB, 8192 tokens, 299466 est) srv update: - prompt 0x5f76d2dd0400: 8191 tokens, checkpoints: 0, 224.067 MiB srv get_availabl: prompt cache update took 77.26 ms slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 9254 | processing task slot update_slots: id 1 | task 9254 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2775 slot update_slots: id 1 | task 9254 | need to evaluate at least 1 token for each active slot (n_past = 2775, task.n_tokens() = 2775) slot update_slots: id 1 | task 9254 | n_past was set to 2774 slot update_slots: id 1 | task 9254 | n_tokens = 2774, memory_seq_rm [2774, end) slot update_slots: id 1 | task 9254 | prompt processing progress, n_tokens = 2775, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 9254 | prompt done, n_tokens = 2775, batch.n_tokens = 1 slot print_timing: id 1 | task 9254 | prompt eval time = 12.92 ms / 1 tokens ( 12.92 ms per token, 77.40 tokens per second) eval time = 1111.32 ms / 405 tokens ( 2.74 ms per token, 364.43 tokens per second) total time = 1124.24 ms / 406 tokens slot release: id 1 | task 9254 | stop processing: n_tokens = 3179, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = -1 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 0 | task 9660 | processing task slot update_slots: id 0 | task 9660 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 377 slot update_slots: id 0 | task 9660 | n_tokens = 0, memory_seq_rm [0, end) slot update_slots: id 0 | task 9660 | prompt processing progress, n_tokens = 377, batch.n_tokens = 377, progress = 1.000000 slot update_slots: id 0 | task 9660 | prompt done, n_tokens = 377, batch.n_tokens = 377 slot print_timing: id 0 | task 9660 | prompt eval time = 22.20 ms / 377 tokens ( 0.06 ms per token, 16979.69 tokens per second) eval time = 15.28 ms / 6 tokens ( 2.55 ms per token, 392.64 tokens per second) total time = 37.48 ms / 383 tokens slot release: id 0 | task 9660 | stop processing: n_tokens = 382, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 0.934 (> 0.100 thold), f_keep = 0.816 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 9667 | processing task slot update_slots: id 1 | task 9667 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2778 slot update_slots: id 1 | task 9667 | n_tokens = 2594, memory_seq_rm [2594, end) slot update_slots: id 1 | task 9667 | prompt processing progress, n_tokens = 2778, batch.n_tokens = 184, progress = 1.000000 slot update_slots: id 1 | task 9667 | prompt done, n_tokens = 2778, batch.n_tokens = 184 slot print_timing: id 1 | task 9667 | prompt eval time = 21.12 ms / 184 tokens ( 0.11 ms per token, 8712.53 tokens per second) eval time = 1386.72 ms / 511 tokens ( 2.71 ms per token, 368.50 tokens per second) total time = 1407.84 ms / 695 tokens slot release: id 1 | task 9667 | stop processing: n_tokens = 3288, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.949 (> 0.100 thold), f_keep = 0.932 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 0 | task 10179 | processing task slot update_slots: id 0 | task 10179 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 375 slot update_slots: id 0 | task 10179 | n_tokens = 356, memory_seq_rm [356, end) slot update_slots: id 0 | task 10179 | prompt processing progress, n_tokens = 375, batch.n_tokens = 19, progress = 1.000000 slot update_slots: id 0 | task 10179 | prompt done, n_tokens = 375, batch.n_tokens = 19 slot print_timing: id 0 | task 10179 | prompt eval time = 10.11 ms / 19 tokens ( 0.53 ms per token, 1879.89 tokens per second) eval time = 17.70 ms / 6 tokens ( 2.95 ms per token, 339.08 tokens per second) total time = 27.80 ms / 25 tokens slot release: id 0 | task 10179 | stop processing: n_tokens = 380, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 0.934 (> 0.100 thold), f_keep = 0.789 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 10186 | processing task slot update_slots: id 1 | task 10186 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2776 slot update_slots: id 1 | task 10186 | n_tokens = 2593, memory_seq_rm [2593, end) slot update_slots: id 1 | task 10186 | prompt processing progress, n_tokens = 2776, batch.n_tokens = 183, progress = 1.000000 slot update_slots: id 1 | task 10186 | prompt done, n_tokens = 2776, batch.n_tokens = 183 slot print_timing: id 1 | task 10186 | prompt eval time = 21.46 ms / 183 tokens ( 0.12 ms per token, 8525.90 tokens per second) eval time = 971.93 ms / 348 tokens ( 2.79 ms per token, 358.05 tokens per second) total time = 993.39 ms / 531 tokens slot release: id 1 | task 10186 | stop processing: n_tokens = 3123, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.889 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 10535 | processing task slot update_slots: id 1 | task 10535 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2776 slot update_slots: id 1 | task 10535 | need to evaluate at least 1 token for each active slot (n_past = 2776, task.n_tokens() = 2776) slot update_slots: id 1 | task 10535 | n_past was set to 2775 slot update_slots: id 1 | task 10535 | n_tokens = 2775, memory_seq_rm [2775, end) slot update_slots: id 1 | task 10535 | prompt processing progress, n_tokens = 2776, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 10535 | prompt done, n_tokens = 2776, batch.n_tokens = 1 slot print_timing: id 1 | task 10535 | prompt eval time = 11.86 ms / 1 tokens ( 11.86 ms per token, 84.30 tokens per second) eval time = 986.41 ms / 348 tokens ( 2.83 ms per token, 352.79 tokens per second) total time = 998.28 ms / 349 tokens slot release: id 1 | task 10535 | stop processing: n_tokens = 3123, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 1 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.889 slot launch_slot_: id 1 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 1 | task 10884 | processing task slot update_slots: id 1 | task 10884 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2776 slot update_slots: id 1 | task 10884 | need to evaluate at least 1 token for each active slot (n_past = 2776, task.n_tokens() = 2776) slot update_slots: id 1 | task 10884 | n_past was set to 2775 slot update_slots: id 1 | task 10884 | n_tokens = 2775, memory_seq_rm [2775, end) slot update_slots: id 1 | task 10884 | prompt processing progress, n_tokens = 2776, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 1 | task 10884 | prompt done, n_tokens = 2776, batch.n_tokens = 1 decode: failed to find a memory slot for batch of size 1 srv try_clear_id: purging slot 0 with 380 tokens slot clear_slot: id 0 | task -1 | clearing slot with 380 tokens srv update_slots: failed to find free space in the KV cache, retrying with smaller batch size, i = 0, n_batch = 2048, ret = 1 slot print_timing: id 1 | task 10884 | prompt eval time = 11.54 ms / 1 tokens ( 11.54 ms per token, 86.62 tokens per second) eval time = 15291.22 ms / 5416 tokens ( 2.82 ms per token, 354.19 tokens per second) total time = 15302.77 ms / 5417 tokens slot release: id 1 | task 10884 | stop processing: n_tokens = 8191, truncated = 1 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 2 | task -1 | selected slot by LRU, t_last = 284097512062 slot launch_slot_: id 2 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 2 | task 16301 | processing task slot update_slots: id 2 | task 16301 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 378 slot update_slots: id 2 | task 16301 | n_tokens = 0, memory_seq_rm [0, end) slot update_slots: id 2 | task 16301 | prompt processing progress, n_tokens = 378, batch.n_tokens = 378, progress = 1.000000 slot update_slots: id 2 | task 16301 | prompt done, n_tokens = 378, batch.n_tokens = 378 decode: failed to find a memory slot for batch of size 378 srv try_clear_id: purging slot 1 with 8191 tokens slot clear_slot: id 1 | task -1 | clearing slot with 8191 tokens srv update_slots: failed to find free space in the KV cache, retrying with smaller batch size, i = 0, n_batch = 2048, ret = 1 slot print_timing: id 2 | task 16301 | prompt eval time = 15.66 ms / 378 tokens ( 0.04 ms per token, 24142.56 tokens per second) eval time = 14.88 ms / 6 tokens ( 2.48 ms per token, 403.23 tokens per second) total time = 30.54 ms / 384 tokens slot release: id 2 | task 16301 | stop processing: n_tokens = 383, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LRU, t_last = 284111914218 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 16308 | processing task slot update_slots: id 3 | task 16308 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2779 slot update_slots: id 3 | task 16308 | n_tokens = 0, memory_seq_rm [0, end) slot update_slots: id 3 | task 16308 | prompt processing progress, n_tokens = 2048, batch.n_tokens = 2048, progress = 0.736956 slot update_slots: id 3 | task 16308 | n_tokens = 2048, memory_seq_rm [2048, end) slot update_slots: id 3 | task 16308 | prompt processing progress, n_tokens = 2779, batch.n_tokens = 731, progress = 1.000000 slot update_slots: id 3 | task 16308 | prompt done, n_tokens = 2779, batch.n_tokens = 731 slot print_timing: id 3 | task 16308 | prompt eval time = 89.56 ms / 2779 tokens ( 0.03 ms per token, 31030.86 tokens per second) eval time = 1024.78 ms / 399 tokens ( 2.57 ms per token, 389.35 tokens per second) total time = 1114.33 ms / 3178 tokens slot release: id 3 | task 16308 | stop processing: n_tokens = 3177, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 2 | task -1 | selected slot by LCP similarity, sim_best = 0.949 (> 0.100 thold), f_keep = 0.930 slot launch_slot_: id 2 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 2 | task 16709 | processing task slot update_slots: id 2 | task 16709 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 375 slot update_slots: id 2 | task 16709 | n_tokens = 356, memory_seq_rm [356, end) slot update_slots: id 2 | task 16709 | prompt processing progress, n_tokens = 375, batch.n_tokens = 19, progress = 1.000000 slot update_slots: id 2 | task 16709 | prompt done, n_tokens = 375, batch.n_tokens = 19 slot print_timing: id 2 | task 16709 | prompt eval time = 7.88 ms / 19 tokens ( 0.41 ms per token, 2409.64 tokens per second) eval time = 16.28 ms / 6 tokens ( 2.71 ms per token, 368.64 tokens per second) total time = 24.16 ms / 25 tokens slot release: id 2 | task 16709 | stop processing: n_tokens = 380, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.934 (> 0.100 thold), f_keep = 0.816 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 16716 | processing task slot update_slots: id 3 | task 16716 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2776 slot update_slots: id 3 | task 16716 | n_tokens = 2593, memory_seq_rm [2593, end) slot update_slots: id 3 | task 16716 | prompt processing progress, n_tokens = 2776, batch.n_tokens = 183, progress = 1.000000 slot update_slots: id 3 | task 16716 | prompt done, n_tokens = 2776, batch.n_tokens = 183 slot print_timing: id 3 | task 16716 | prompt eval time = 20.99 ms / 183 tokens ( 0.11 ms per token, 8719.68 tokens per second) eval time = 1130.38 ms / 435 tokens ( 2.60 ms per token, 384.83 tokens per second) total time = 1151.37 ms / 618 tokens slot release: id 3 | task 16716 | stop processing: n_tokens = 3210, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.865 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 17152 | processing task slot update_slots: id 3 | task 17152 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2776 slot update_slots: id 3 | task 17152 | need to evaluate at least 1 token for each active slot (n_past = 2776, task.n_tokens() = 2776) slot update_slots: id 3 | task 17152 | n_past was set to 2775 slot update_slots: id 3 | task 17152 | n_tokens = 2775, memory_seq_rm [2775, end) slot update_slots: id 3 | task 17152 | prompt processing progress, n_tokens = 2776, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 3 | task 17152 | prompt done, n_tokens = 2776, batch.n_tokens = 1 slot print_timing: id 3 | task 17152 | prompt eval time = 11.40 ms / 1 tokens ( 11.40 ms per token, 87.75 tokens per second) eval time = 1205.67 ms / 446 tokens ( 2.70 ms per token, 369.92 tokens per second) total time = 1217.07 ms / 447 tokens slot release: id 3 | task 17152 | stop processing: n_tokens = 3221, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.862 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 17599 | processing task slot update_slots: id 3 | task 17599 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2776 slot update_slots: id 3 | task 17599 | need to evaluate at least 1 token for each active slot (n_past = 2776, task.n_tokens() = 2776) slot update_slots: id 3 | task 17599 | n_past was set to 2775 slot update_slots: id 3 | task 17599 | n_tokens = 2775, memory_seq_rm [2775, end) slot update_slots: id 3 | task 17599 | prompt processing progress, n_tokens = 2776, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 3 | task 17599 | prompt done, n_tokens = 2776, batch.n_tokens = 1 slot print_timing: id 3 | task 17599 | prompt eval time = 11.40 ms / 1 tokens ( 11.40 ms per token, 87.73 tokens per second) eval time = 1322.92 ms / 511 tokens ( 2.59 ms per token, 386.27 tokens per second) total time = 1334.32 ms / 512 tokens slot release: id 3 | task 17599 | stop processing: n_tokens = 3286, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 2 | task -1 | selected slot by LCP similarity, sim_best = 0.904 (> 0.100 thold), f_keep = 0.937 slot launch_slot_: id 2 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 2 | task 18111 | processing task slot update_slots: id 2 | task 18111 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 394 slot update_slots: id 2 | task 18111 | n_tokens = 356, memory_seq_rm [356, end) slot update_slots: id 2 | task 18111 | prompt processing progress, n_tokens = 394, batch.n_tokens = 38, progress = 1.000000 slot update_slots: id 2 | task 18111 | prompt done, n_tokens = 394, batch.n_tokens = 38 slot print_timing: id 2 | task 18111 | prompt eval time = 10.93 ms / 38 tokens ( 0.29 ms per token, 3476.03 tokens per second) eval time = 15.30 ms / 6 tokens ( 2.55 ms per token, 392.16 tokens per second) total time = 26.23 ms / 44 tokens slot release: id 2 | task 18111 | stop processing: n_tokens = 399, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.928 (> 0.100 thold), f_keep = 0.789 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 18118 | processing task slot update_slots: id 3 | task 18118 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2795 slot update_slots: id 3 | task 18118 | n_tokens = 2593, memory_seq_rm [2593, end) slot update_slots: id 3 | task 18118 | prompt processing progress, n_tokens = 2795, batch.n_tokens = 202, progress = 1.000000 slot update_slots: id 3 | task 18118 | prompt done, n_tokens = 2795, batch.n_tokens = 202 decode: failed to find a memory slot for batch of size 1 srv try_clear_id: purging slot 2 with 399 tokens slot clear_slot: id 2 | task -1 | clearing slot with 399 tokens srv update_slots: failed to find free space in the KV cache, retrying with smaller batch size, i = 0, n_batch = 2048, ret = 1 slot print_timing: id 3 | task 18118 | prompt eval time = 20.43 ms / 202 tokens ( 0.10 ms per token, 9886.94 tokens per second) eval time = 15154.56 ms / 5397 tokens ( 2.81 ms per token, 356.13 tokens per second) total time = 15174.99 ms / 5599 tokens slot release: id 3 | task 18118 | stop processing: n_tokens = 8191, truncated = 1 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.341 srv get_availabl: updating prompt cache srv prompt_save: - saving prompt with length 8191, total state size = 224.067 MiB srv load: - looking for better prompt, base f_keep = 0.341, sim = 1.000 srv update: - cache state: 2 prompts, 448.134 MiB (limits: 8192.000 MiB, 8192 tokens, 299466 est) srv update: - prompt 0x5f76d2dd0400: 8191 tokens, checkpoints: 0, 224.067 MiB srv update: - prompt 0x5f76d2536550: 8191 tokens, checkpoints: 0, 224.067 MiB srv get_availabl: prompt cache update took 77.20 ms slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 23516 | processing task slot update_slots: id 3 | task 23516 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2795 slot update_slots: id 3 | task 23516 | need to evaluate at least 1 token for each active slot (n_past = 2795, task.n_tokens() = 2795) slot update_slots: id 3 | task 23516 | n_past was set to 2794 slot update_slots: id 3 | task 23516 | n_tokens = 2794, memory_seq_rm [2794, end) slot update_slots: id 3 | task 23516 | prompt processing progress, n_tokens = 2795, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 3 | task 23516 | prompt done, n_tokens = 2795, batch.n_tokens = 1 slot print_timing: id 3 | task 23516 | prompt eval time = 11.88 ms / 1 tokens ( 11.88 ms per token, 84.16 tokens per second) eval time = 747.43 ms / 292 tokens ( 2.56 ms per token, 390.67 tokens per second) total time = 759.31 ms / 293 tokens slot release: id 3 | task 23516 | stop processing: n_tokens = 3086, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.906 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 23809 | processing task slot update_slots: id 3 | task 23809 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2795 slot update_slots: id 3 | task 23809 | need to evaluate at least 1 token for each active slot (n_past = 2795, task.n_tokens() = 2795) slot update_slots: id 3 | task 23809 | n_past was set to 2794 slot update_slots: id 3 | task 23809 | n_tokens = 2794, memory_seq_rm [2794, end) slot update_slots: id 3 | task 23809 | prompt processing progress, n_tokens = 2795, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 3 | task 23809 | prompt done, n_tokens = 2795, batch.n_tokens = 1 slot print_timing: id 3 | task 23809 | prompt eval time = 10.76 ms / 1 tokens ( 10.76 ms per token, 92.92 tokens per second) eval time = 750.29 ms / 290 tokens ( 2.59 ms per token, 386.52 tokens per second) total time = 761.06 ms / 291 tokens slot release: id 3 | task 23809 | stop processing: n_tokens = 3084, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 284134057035 slot launch_slot_: id 0 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 0 | task 24100 | processing task slot update_slots: id 0 | task 24100 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 397 slot update_slots: id 0 | task 24100 | n_tokens = 0, memory_seq_rm [0, end) slot update_slots: id 0 | task 24100 | prompt processing progress, n_tokens = 397, batch.n_tokens = 397, progress = 1.000000 slot update_slots: id 0 | task 24100 | prompt done, n_tokens = 397, batch.n_tokens = 397 slot print_timing: id 0 | task 24100 | prompt eval time = 21.96 ms / 397 tokens ( 0.06 ms per token, 18076.68 tokens per second) eval time = 14.86 ms / 6 tokens ( 2.48 ms per token, 403.82 tokens per second) total time = 36.82 ms / 403 tokens slot release: id 0 | task 24100 | stop processing: n_tokens = 402, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 0.927 (> 0.100 thold), f_keep = 0.841 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 24107 | processing task slot update_slots: id 3 | task 24107 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2798 slot update_slots: id 3 | task 24107 | n_tokens = 2593, memory_seq_rm [2593, end) slot update_slots: id 3 | task 24107 | prompt processing progress, n_tokens = 2798, batch.n_tokens = 205, progress = 1.000000 slot update_slots: id 3 | task 24107 | prompt done, n_tokens = 2798, batch.n_tokens = 205 slot print_timing: id 3 | task 24107 | prompt eval time = 21.18 ms / 205 tokens ( 0.10 ms per token, 9679.86 tokens per second) eval time = 1191.36 ms / 437 tokens ( 2.73 ms per token, 366.81 tokens per second) total time = 1212.54 ms / 642 tokens slot release: id 3 | task 24107 | stop processing: n_tokens = 3234, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.865 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 24545 | processing task slot update_slots: id 3 | task 24545 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2798 slot update_slots: id 3 | task 24545 | need to evaluate at least 1 token for each active slot (n_past = 2798, task.n_tokens() = 2798) slot update_slots: id 3 | task 24545 | n_past was set to 2797 slot update_slots: id 3 | task 24545 | n_tokens = 2797, memory_seq_rm [2797, end) slot update_slots: id 3 | task 24545 | prompt processing progress, n_tokens = 2798, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 3 | task 24545 | prompt done, n_tokens = 2798, batch.n_tokens = 1 slot print_timing: id 3 | task 24545 | prompt eval time = 12.41 ms / 1 tokens ( 12.41 ms per token, 80.59 tokens per second) eval time = 1437.48 ms / 537 tokens ( 2.68 ms per token, 373.57 tokens per second) total time = 1449.89 ms / 538 tokens slot release: id 3 | task 24545 | stop processing: n_tokens = 3334, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200 srv params_from_: Chat format: Content-only slot get_availabl: id 3 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 0.839 slot launch_slot_: id 3 | task -1 | sampler chain: logits -> logit-bias -> penalties -> dry -> top-n-sigma -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist slot launch_slot_: id 3 | task 25083 | processing task slot update_slots: id 3 | task 25083 | new prompt, n_ctx_slot = 8192, n_keep = 0, task.n_tokens = 2798 slot update_slots: id 3 | task 25083 | need to evaluate at least 1 token for each active slot (n_past = 2798, task.n_tokens() = 2798) slot update_slots: id 3 | task 25083 | n_past was set to 2797 slot update_slots: id 3 | task 25083 | n_tokens = 2797, memory_seq_rm [2797, end) slot update_slots: id 3 | task 25083 | prompt processing progress, n_tokens = 2798, batch.n_tokens = 1, progress = 1.000000 slot update_slots: id 3 | task 25083 | prompt done, n_tokens = 2798, batch.n_tokens = 1 slot print_timing: id 3 | task 25083 | prompt eval time = 11.78 ms / 1 tokens ( 11.78 ms per token, 84.90 tokens per second) eval time = 781.65 ms / 278 tokens ( 2.81 ms per token, 355.66 tokens per second) total time = 793.43 ms / 279 tokens slot release: id 3 | task 25083 | stop processing: n_tokens = 3075, truncated = 0 srv update_slots: all slots are idle srv log_server_r: request: POST /v1/chat/completions 127.0.0.1 200