Skip to content

Eval bug: Gemma 3 extremly slow prompt processing when using quantized kv cache. #12352

Open
@Bearsaerker

Description

@Bearsaerker

Name and Version

./build/bin/llama-cli --version
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: yes
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce RTX 3060, compute capability 8.6, VMM: yes
version: 0 (unknown)
built with cc (GCC) 13.3.1 20240611 (Red Hat 13.3.1-2) for x86_64-redhat-linux

(newest b4876 version)

Operating systems

Linux

GGML backends

CUDA

Hardware

Ryzen 3900x + rtx 3060 12gb

Models

Gemma-3-12b_Q5_K_M

Problem description & steps to reproduce

Prompt eval time is way slower when using quantized kv cache than standard kv cache. Also I see that the cpu is used when the quantized kv cache is turned on. So I believe that the kv cache is not properly processed by the gpu if the quantized kv cache is provided

First Bad Commit

No response

Relevant log output

# Unquantized kv cache:

./build/bin/llama-server -m '/home/luis/Downloads/llama.cpp-b4876/models/gemma-3-12b-it-Q5_K_M.gguf'  --n-gpu-layers -1 --batch_size 1024 --flash-attn -c 4000 --port 7777 -t 8 -ngl 99
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: yes
ggml_cuda_init: found 1 CUDA devices:
  Device 0: NVIDIA GeForce RTX 3060, compute capability 8.6, VMM: yes
build: 0 (unknown) with cc (GCC) 13.3.1 20240611 (Red Hat 13.3.1-2) for x86_64-redhat-linux
system info: n_threads = 8, n_threads_batch = 8, total_threads = 24

system_info: n_threads = 8 (n_threads_batch = 8) / 24 | CUDA : ARCHS = 860 | FORCE_CUBLAS = 1 | USE_GRAPHS = 1 | PEER_MAX_BATCH_SIZE = 128 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | LLAMAFILE = 1 | OPENMP = 1 | AARCH64_REPACK = 1 | 

main: HTTP server is listening, hostname: 127.0.0.1, port: 7777, http threads: 23
main: loading model
srv    load_model: loading model '/home/luis/Downloads/llama.cpp-b4876/models/gemma-3-12b-it-Q5_K_M.gguf'
llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce RTX 3060) - 10456 MiB free
llama_model_loader: loaded meta data with 34 key-value pairs and 626 tensors from /home/luis/Downloads/llama.cpp-b4876/models/gemma-3-12b-it-Q5_K_M.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              = gemma3
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Gemma 3
llama_model_loader: - kv   3:                       general.quantized_by str              = Unsloth
llama_model_loader: - kv   4:                         general.size_label str              = 12B
llama_model_loader: - kv   5:                           general.repo_url str              = https://huggingface.co/unsloth
llama_model_loader: - kv   6:                      gemma3.context_length u32              = 131072
llama_model_loader: - kv   7:                    gemma3.embedding_length u32              = 3840
llama_model_loader: - kv   8:                         gemma3.block_count u32              = 48
llama_model_loader: - kv   9:                 gemma3.feed_forward_length u32              = 15360
llama_model_loader: - kv  10:                gemma3.attention.head_count u32              = 16
llama_model_loader: - kv  11:    gemma3.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  12:                gemma3.attention.key_length u32              = 256
llama_model_loader: - kv  13:              gemma3.attention.value_length u32              = 256
llama_model_loader: - kv  14:                      gemma3.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  15:            gemma3.attention.sliding_window u32              = 1024
llama_model_loader: - kv  16:             gemma3.attention.head_count_kv u32              = 8
llama_model_loader: - kv  17:                   gemma3.rope.scaling.type str              = linear
llama_model_loader: - kv  18:                 gemma3.rope.scaling.factor f32              = 8.000000
llama_model_loader: - kv  19:                       tokenizer.ggml.model str              = llama
llama_model_loader: - kv  20:                         tokenizer.ggml.pre str              = default
llama_model_loader: - kv  21:                      tokenizer.ggml.tokens arr[str,262208]  = ["<pad>", "<eos>", "<bos>", "<unk>", ...
llama_model_loader: - kv  22:                      tokenizer.ggml.scores arr[f32,262208]  = [-1000.000000, -1000.000000, -1000.00...
llama_model_loader: - kv  23:                  tokenizer.ggml.token_type arr[i32,262208]  = [3, 3, 3, 3, 3, 4, 3, 3, 3, 3, 3, 3, ...
llama_model_loader: - kv  24:                tokenizer.ggml.bos_token_id u32              = 2
llama_model_loader: - kv  25:                tokenizer.ggml.eos_token_id u32              = 106
llama_model_loader: - kv  26:            tokenizer.ggml.unknown_token_id u32              = 3
llama_model_loader: - kv  27:            tokenizer.ggml.padding_token_id u32              = 0
llama_model_loader: - kv  28:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  29:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  30:                    tokenizer.chat_template str              = {{ bos_token }}\n{%- if messages[0]['r...
llama_model_loader: - kv  31:            tokenizer.ggml.add_space_prefix bool             = false
llama_model_loader: - kv  32:               general.quantization_version u32              = 2
llama_model_loader: - kv  33:                          general.file_type u32              = 17
llama_model_loader: - type  f32:  289 tensors
llama_model_loader: - type q5_K:  288 tensors
llama_model_loader: - type q6_K:   49 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type   = Q5_K - Medium
print_info: file size   = 7.86 GiB (5.74 BPW) 
load: special tokens cache size = 6415
load: token to piece cache size = 1.9446 MB
print_info: arch             = gemma3
print_info: vocab_only       = 0
print_info: n_ctx_train      = 131072
print_info: n_embd           = 3840
print_info: n_layer          = 48
print_info: n_head           = 16
print_info: n_head_kv        = 8
print_info: n_rot            = 256
print_info: n_swa            = 1024
print_info: n_embd_head_k    = 256
print_info: n_embd_head_v    = 256
print_info: n_gqa            = 2
print_info: n_embd_k_gqa     = 2048
print_info: n_embd_v_gqa     = 2048
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     = 6.2e-02
print_info: n_ff             = 15360
print_info: n_expert         = 0
print_info: n_expert_used    = 0
print_info: causal attn      = 1
print_info: pooling type     = 0
print_info: rope type        = 2
print_info: rope scaling     = linear
print_info: freq_base_train  = 1000000.0
print_info: freq_scale_train = 0.125
print_info: n_ctx_orig_yarn  = 131072
print_info: rope_finetuned   = unknown
print_info: ssm_d_conv       = 0
print_info: ssm_d_inner      = 0
print_info: ssm_d_state      = 0
print_info: ssm_dt_rank      = 0
print_info: ssm_dt_b_c_rms   = 0
print_info: model type       = 12B
print_info: model params     = 11.77 B
print_info: general.name     = Gemma 3
print_info: vocab type       = SPM
print_info: n_vocab          = 262208
print_info: n_merges         = 0
print_info: BOS token        = 2 '<bos>'
print_info: EOS token        = 106 '<end_of_turn>'
print_info: EOT token        = 106 '<end_of_turn>'
print_info: UNK token        = 3 '<unk>'
print_info: PAD token        = 0 '<pad>'
print_info: LF token         = 248 '<0x0A>'
print_info: EOG token        = 106 '<end_of_turn>'
print_info: max token length = 48
load_tensors: loading model tensors, this can take a while... (mmap = true)
load_tensors: offloading 48 repeating layers to GPU
load_tensors: offloading output layer to GPU
load_tensors: offloaded 49/49 layers to GPU
load_tensors:   CPU_Mapped model buffer size =   787.69 MiB
load_tensors:        CUDA0 model buffer size =  8047.63 MiB
.....................................................................................
llama_init_from_model: n_seq_max     = 1
llama_init_from_model: n_ctx         = 4096
llama_init_from_model: n_ctx_per_seq = 4096
llama_init_from_model: n_batch       = 1024
llama_init_from_model: n_ubatch      = 512
llama_init_from_model: flash_attn    = 1
llama_init_from_model: freq_base     = 1000000.0
llama_init_from_model: freq_scale    = 0.125
llama_init_from_model: n_ctx_per_seq (4096) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
llama_kv_cache_init: kv_size = 4096, offload = 1, type_k = 'f16', type_v = 'f16', n_layer = 48, can_shift = 1
llama_kv_cache_init:      CUDA0 KV buffer size =  1536.00 MiB
llama_init_from_model: KV self size  = 1536.00 MiB, K (f16):  768.00 MiB, V (f16):  768.00 MiB
llama_init_from_model:  CUDA_Host  output buffer size =     1.00 MiB
llama_init_from_model:      CUDA0 compute buffer size =   519.62 MiB
llama_init_from_model:  CUDA_Host compute buffer size =    23.51 MiB
llama_init_from_model: graph nodes  = 1737
llama_init_from_model: graph splits = 2
common_init_from_params: setting dry_penalty_last_n to ctx_size = 4096
common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)
srv          init: initializing slots, n_slots = 1
slot         init: id  0 | task -1 | new slot n_ctx_slot = 4096
main: model loaded
main: chat template, chat_template: {{ bos_token }}
{%- if messages[0]['role'] == 'system' -%}
    {%- if messages[0]['content'] is string -%}
        {%- set first_user_prefix = messages[0]['content'] + '

' -%}
    {%- else -%}
        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '

' -%}
    {%- endif -%}
    {%- set loop_messages = messages[1:] -%}
{%- else -%}
    {%- set first_user_prefix = "" -%}
    {%- set loop_messages = messages -%}
{%- endif -%}
{%- for message in loop_messages -%}
    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
        {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
    {%- endif -%}
    {%- if (message['role'] == 'assistant') -%}
        {%- set role = "model" -%}
    {%- else -%}
        {%- set role = message['role'] -%}
    {%- endif -%}
    {{ '<start_of_turn>' + role + '
' + (first_user_prefix if loop.first else "") }}
    {%- if message['content'] is string -%}
        {{ message['content'] | trim }}
    {%- elif message['content'] is iterable -%}
        {%- for item in message['content'] -%}
            {%- if item['type'] == 'image' -%}
                {{ '<start_of_image>' }}
            {%- elif item['type'] == 'text' -%}
                {{ item['text'] | trim }}
            {%- endif -%}
        {%- endfor -%}
    {%- else -%}
        {{ raise_exception("Invalid content type") }}
    {%- endif -%}
    {{ '<end_of_turn>
' }}
{%- endfor -%}
{%- if add_generation_prompt -%}
    {{'<start_of_turn>model
'}}
{%- endif -%}
, example_format: '<start_of_turn>user
You are a helpful assistant

Hello<end_of_turn>
<start_of_turn>model
Hi there<end_of_turn>
<start_of_turn>user
How are you?<end_of_turn>
<start_of_turn>model
'
main: server is listening on http://127.0.0.1:7777 - starting the main loop
srv  update_slots: all slots are idle
srv  params_from_: Chat format: Content-only
slot launch_slot_: id  0 | task 0 | processing task
slot update_slots: id  0 | task 0 | new prompt, n_ctx_slot = 4096, n_keep = 0, n_prompt_tokens = 2505
slot update_slots: id  0 | task 0 | kv cache rm [0, end)
slot update_slots: id  0 | task 0 | prompt processing progress, n_past = 1024, n_tokens = 1024, progress = 0.408782
slot update_slots: id  0 | task 0 | kv cache rm [1024, end)
slot update_slots: id  0 | task 0 | prompt processing progress, n_past = 2048, n_tokens = 1024, progress = 0.817565
slot update_slots: id  0 | task 0 | kv cache rm [2048, end)
slot update_slots: id  0 | task 0 | prompt processing progress, n_past = 2505, n_tokens = 457, progress = 1.000000
slot update_slots: id  0 | task 0 | prompt done, n_past = 2505, n_tokens = 457
slot      release: id  0 | task 0 | stop processing: n_past = 3199, truncated = 0
slot print_timing: id  0 | task 0 | 
prompt eval time =    2697.39 ms /  2505 tokens (    1.08 ms per token,   928.67 tokens per second)
       eval time =   24911.73 ms /   695 tokens (   35.84 ms per token,    27.90 tokens per second)
      total time =   27609.12 ms /  3200 tokens
srv  update_slots: all slots are idle
srv  log_server_r: request: POST /v1/chat/completions 127.0.0.1 200


# Quantized kv cache:
 ./build/bin/llama-server -m '/home/luis/Downloads/llama.cpp-b4876/models/gemma-3-12b-it-Q5_K_M.gguf'  --n-gpu-layers -1 --cache-type-k q8_0 --cache-type-v q8_0 --batch_size 1024 --flash-attn -c 4000 --port 7777 -t 8 -ngl 99
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: yes
ggml_cuda_init: found 1 CUDA devices:
  Device 0: NVIDIA GeForce RTX 3060, compute capability 8.6, VMM: yes
build: 0 (unknown) with cc (GCC) 13.3.1 20240611 (Red Hat 13.3.1-2) for x86_64-redhat-linux
system info: n_threads = 8, n_threads_batch = 8, total_threads = 24

system_info: n_threads = 8 (n_threads_batch = 8) / 24 | CUDA : ARCHS = 860 | FORCE_CUBLAS = 1 | USE_GRAPHS = 1 | PEER_MAX_BATCH_SIZE = 128 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | LLAMAFILE = 1 | OPENMP = 1 | AARCH64_REPACK = 1 | 

main: HTTP server is listening, hostname: 127.0.0.1, port: 7777, http threads: 23
main: loading model
srv    load_model: loading model '/home/luis/Downloads/llama.cpp-b4876/models/gemma-3-12b-it-Q5_K_M.gguf'
llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce RTX 3060) - 10500 MiB free
llama_model_loader: loaded meta data with 34 key-value pairs and 626 tensors from /home/luis/Downloads/llama.cpp-b4876/models/gemma-3-12b-it-Q5_K_M.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              = gemma3
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Gemma 3
llama_model_loader: - kv   3:                       general.quantized_by str              = Unsloth
llama_model_loader: - kv   4:                         general.size_label str              = 12B
llama_model_loader: - kv   5:                           general.repo_url str              = https://huggingface.co/unsloth
llama_model_loader: - kv   6:                      gemma3.context_length u32              = 131072
llama_model_loader: - kv   7:                    gemma3.embedding_length u32              = 3840
llama_model_loader: - kv   8:                         gemma3.block_count u32              = 48
llama_model_loader: - kv   9:                 gemma3.feed_forward_length u32              = 15360
llama_model_loader: - kv  10:                gemma3.attention.head_count u32              = 16
llama_model_loader: - kv  11:    gemma3.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  12:                gemma3.attention.key_length u32              = 256
llama_model_loader: - kv  13:              gemma3.attention.value_length u32              = 256
llama_model_loader: - kv  14:                      gemma3.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  15:            gemma3.attention.sliding_window u32              = 1024
llama_model_loader: - kv  16:             gemma3.attention.head_count_kv u32              = 8
llama_model_loader: - kv  17:                   gemma3.rope.scaling.type str              = linear
llama_model_loader: - kv  18:                 gemma3.rope.scaling.factor f32              = 8.000000
llama_model_loader: - kv  19:                       tokenizer.ggml.model str              = llama
llama_model_loader: - kv  20:                         tokenizer.ggml.pre str              = default
llama_model_loader: - kv  21:                      tokenizer.ggml.tokens arr[str,262208]  = ["<pad>", "<eos>", "<bos>", "<unk>", ...
llama_model_loader: - kv  22:                      tokenizer.ggml.scores arr[f32,262208]  = [-1000.000000, -1000.000000, -1000.00...
llama_model_loader: - kv  23:                  tokenizer.ggml.token_type arr[i32,262208]  = [3, 3, 3, 3, 3, 4, 3, 3, 3, 3, 3, 3, ...
llama_model_loader: - kv  24:                tokenizer.ggml.bos_token_id u32              = 2
llama_model_loader: - kv  25:                tokenizer.ggml.eos_token_id u32              = 106
llama_model_loader: - kv  26:            tokenizer.ggml.unknown_token_id u32              = 3
llama_model_loader: - kv  27:            tokenizer.ggml.padding_token_id u32              = 0
llama_model_loader: - kv  28:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  29:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  30:                    tokenizer.chat_template str              = {{ bos_token }}\n{%- if messages[0]['r...
llama_model_loader: - kv  31:            tokenizer.ggml.add_space_prefix bool             = false
llama_model_loader: - kv  32:               general.quantization_version u32              = 2
llama_model_loader: - kv  33:                          general.file_type u32              = 17
llama_model_loader: - type  f32:  289 tensors
llama_model_loader: - type q5_K:  288 tensors
llama_model_loader: - type q6_K:   49 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type   = Q5_K - Medium
print_info: file size   = 7.86 GiB (5.74 BPW) 
load: special tokens cache size = 6415
load: token to piece cache size = 1.9446 MB
print_info: arch             = gemma3
print_info: vocab_only       = 0
print_info: n_ctx_train      = 131072
print_info: n_embd           = 3840
print_info: n_layer          = 48
print_info: n_head           = 16
print_info: n_head_kv        = 8
print_info: n_rot            = 256
print_info: n_swa            = 1024
print_info: n_embd_head_k    = 256
print_info: n_embd_head_v    = 256
print_info: n_gqa            = 2
print_info: n_embd_k_gqa     = 2048
print_info: n_embd_v_gqa     = 2048
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     = 6.2e-02
print_info: n_ff             = 15360
print_info: n_expert         = 0
print_info: n_expert_used    = 0
print_info: causal attn      = 1
print_info: pooling type     = 0
print_info: rope type        = 2
print_info: rope scaling     = linear
print_info: freq_base_train  = 1000000.0
print_info: freq_scale_train = 0.125
print_info: n_ctx_orig_yarn  = 131072
print_info: rope_finetuned   = unknown
print_info: ssm_d_conv       = 0
print_info: ssm_d_inner      = 0
print_info: ssm_d_state      = 0
print_info: ssm_dt_rank      = 0
print_info: ssm_dt_b_c_rms   = 0
print_info: model type       = 12B
print_info: model params     = 11.77 B
print_info: general.name     = Gemma 3
print_info: vocab type       = SPM
print_info: n_vocab          = 262208
print_info: n_merges         = 0
print_info: BOS token        = 2 '<bos>'
print_info: EOS token        = 106 '<end_of_turn>'
print_info: EOT token        = 106 '<end_of_turn>'
print_info: UNK token        = 3 '<unk>'
print_info: PAD token        = 0 '<pad>'
print_info: LF token         = 248 '<0x0A>'
print_info: EOG token        = 106 '<end_of_turn>'
print_info: max token length = 48
load_tensors: loading model tensors, this can take a while... (mmap = true)
load_tensors: offloading 48 repeating layers to GPU
load_tensors: offloading output layer to GPU
load_tensors: offloaded 49/49 layers to GPU
load_tensors:   CPU_Mapped model buffer size =   787.69 MiB
load_tensors:        CUDA0 model buffer size =  8047.63 MiB
.....................................................................................
llama_init_from_model: n_seq_max     = 1
llama_init_from_model: n_ctx         = 4096
llama_init_from_model: n_ctx_per_seq = 4096
llama_init_from_model: n_batch       = 1024
llama_init_from_model: n_ubatch      = 512
llama_init_from_model: flash_attn    = 1
llama_init_from_model: freq_base     = 1000000.0
llama_init_from_model: freq_scale    = 0.125
llama_init_from_model: n_ctx_per_seq (4096) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
llama_kv_cache_init: kv_size = 4096, offload = 1, type_k = 'q8_0', type_v = 'q8_0', n_layer = 48, can_shift = 1
llama_kv_cache_init:      CUDA0 KV buffer size =   816.00 MiB
llama_init_from_model: KV self size  =  816.00 MiB, K (q8_0):  408.00 MiB, V (q8_0):  408.00 MiB
llama_init_from_model:  CUDA_Host  output buffer size =     1.00 MiB
llama_init_from_model:      CUDA0 compute buffer size =   519.62 MiB
llama_init_from_model:  CUDA_Host compute buffer size =    45.01 MiB
llama_init_from_model: graph nodes  = 1737
llama_init_from_model: graph splits = 98
common_init_from_params: setting dry_penalty_last_n to ctx_size = 4096
common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)
srv          init: initializing slots, n_slots = 1
slot         init: id  0 | task -1 | new slot n_ctx_slot = 4096
main: model loaded
main: chat template, chat_template: {{ bos_token }}
{%- if messages[0]['role'] == 'system' -%}
    {%- if messages[0]['content'] is string -%}
        {%- set first_user_prefix = messages[0]['content'] + '

' -%}
    {%- else -%}
        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '

' -%}
    {%- endif -%}
    {%- set loop_messages = messages[1:] -%}
{%- else -%}
    {%- set first_user_prefix = "" -%}
    {%- set loop_messages = messages -%}
{%- endif -%}
{%- for message in loop_messages -%}
    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
        {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
    {%- endif -%}
    {%- if (message['role'] == 'assistant') -%}
        {%- set role = "model" -%}
    {%- else -%}
        {%- set role = message['role'] -%}
    {%- endif -%}
    {{ '<start_of_turn>' + role + '
' + (first_user_prefix if loop.first else "") }}
    {%- if message['content'] is string -%}
        {{ message['content'] | trim }}
    {%- elif message['content'] is iterable -%}
        {%- for item in message['content'] -%}
            {%- if item['type'] == 'image' -%}
                {{ '<start_of_image>' }}
            {%- elif item['type'] == 'text' -%}
                {{ item['text'] | trim }}
            {%- endif -%}
        {%- endfor -%}
    {%- else -%}
        {{ raise_exception("Invalid content type") }}
    {%- endif -%}
    {{ '<end_of_turn>
' }}
{%- endfor -%}
{%- if add_generation_prompt -%}
    {{'<start_of_turn>model
'}}
{%- endif -%}
, example_format: '<start_of_turn>user
You are a helpful assistant

Hello<end_of_turn>
<start_of_turn>model
Hi there<end_of_turn>
<start_of_turn>user
How are you?<end_of_turn>
<start_of_turn>model
'
main: server is listening on http://127.0.0.1:7777 - starting the main loop
srv  update_slots: all slots are idle
srv  params_from_: Chat format: Content-only
slot launch_slot_: id  0 | task 0 | processing task
slot update_slots: id  0 | task 0 | new prompt, n_ctx_slot = 4096, n_keep = 0, n_prompt_tokens = 2505
slot update_slots: id  0 | task 0 | kv cache rm [0, end)
slot update_slots: id  0 | task 0 | prompt processing progress, n_past = 1024, n_tokens = 1024, progress = 0.408782
slot update_slots: id  0 | task 0 | kv cache rm [1024, end)
slot update_slots: id  0 | task 0 | prompt processing progress, n_past = 2048, n_tokens = 1024, progress = 0.817565
slot update_slots: id  0 | task 0 | kv cache rm [2048, end)
slot update_slots: id  0 | task 0 | prompt processing progress, n_past = 2505, n_tokens = 457, progress = 1.000000
slot update_slots: id  0 | task 0 | prompt done, n_past = 2505, n_tokens = 457
slot      release: id  0 | task 0 | stop processing: n_past = 3209, truncated = 0
slot print_timing: id  0 | task 0 | 
prompt eval time =   23150.35 ms /  2505 tokens (    9.24 ms per token,   108.21 tokens per second)
       eval time =   78076.05 ms /   705 tokens (  110.75 ms per token,     9.03 tokens per second)
      total time =  101226.41 ms /  3210 tokens
srv  update_slots: all slots are idle
srv  log_server_r: request: POST /v1/chat/completions 127.0.0.1 200

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions