[Performance] Llava-cli offloading image encoding to cuda

[rezacopol]

Running on cuda 12.2.

Looking at running llava 1.5 using llava-cli, image encoding timing is 10x worse than running on Mac m2. I saw this thread that seems it fixed the issue but numbers tell a different story. Any insight?

Running on M2:

encode_image_with_clip: image embedding created: 576 tokens
encode_image_with_clip: image encoded in   129.36 ms by CLIP ( 0.22 ms per image patch)
llama_print_timings:        load time =    1178.11 ms
llama_print_timings:      sample time =       9.86 ms /   141 runs   (    0.07 ms per token, 14294.40 tokens per second)
llama_print_timings: prompt eval time =     938.64 ms /   618 tokens (    1.52 ms per token,   658.40 tokens per second)
llama_print_timings:        eval time =    2115.32 ms /   140 runs   (   15.11 ms per token,    66.18 tokens per second)
llama_print_timings:       total time =    3329.86 ms /   758 tokens

Running on cuda:

encode_image_with_clip: image embedding created: 576 tokens
encode_image_with_clip: image encoded in  1322.46 ms by CLIP (2.30 ms per image patch)
llama_print_timings:        load time =    2987.52 ms
llama_print_timings:      sample time =       7.05 ms /   183 runs   (    0.04 ms per token, 25946.41 tokens per second)
llama_print_timings: prompt eval time =     218.61 ms /   618 tokens (    0.35 ms per token,  2826.94 tokens per second)
llama_print_timings:        eval time =    1603.21 ms /   182 runs   (    8.81 ms per token,   113.52 tokens per second)
llama_print_timings:       total time =    4648.39 ms /   800 tokens

[ggerganov]

It works on my end as expected:

$ ▶ LLAMA_CUDA=1 make -j && ./llava-cli -m ./models/llava-7b-v1.6/ggml-model-q4_0.gguf --mmproj ./models/llava-7b-v1.6/mmproj-model-f16.gguf --image ~/Downloads/cat.png -p "What is the animal doing?" --temp 0.0 -ngl 99
I ccache not found. Consider installing it for faster compilation.
I llama.cpp build info: 
I UNAME_S:   Linux
I UNAME_P:   x86_64
I UNAME_M:   x86_64
I CFLAGS:    -I. -Icommon -D_XOPEN_SOURCE=600 -D_GNU_SOURCE -DNDEBUG -DGGML_USE_LLAMAFILE -DGGML_USE_CUDA -I/usr/local/cuda/include -I/usr/local/cuda/targets/x86_64-linux/include  -std=c11   -fPIC -O3 -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes -Werror=implicit-int -Werror=implicit-function-declaration -pthread -march=native -mtune=native -Wdouble-promotion 
I CXXFLAGS:  -std=c++11 -fPIC -O3 -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wmissing-declarations -Wmissing-noreturn -pthread  -march=native -mtune=native -Wno-array-bounds -Wno-format-truncation -Wextra-semi -I. -Icommon -D_XOPEN_SOURCE=600 -D_GNU_SOURCE -DNDEBUG -DGGML_USE_LLAMAFILE -DGGML_USE_CUDA -I/usr/local/cuda/include -I/usr/local/cuda/targets/x86_64-linux/include 
I NVCCFLAGS: -std=c++11 -O3 -use_fast_math --forward-unknown-to-host-compiler -arch=native -DGGML_CUDA_DMMV_X=32 -DGGML_CUDA_MMV_Y=1 -DK_QUANTS_PER_ITERATION=2 -DGGML_CUDA_PEER_MAX_BATCH_SIZE=128 
I LDFLAGS:   -lcuda -lcublas -lculibos -lcudart -lcublasLt -lpthread -ldl -lrt -L/usr/local/cuda/lib64 -L/usr/lib64 -L/usr/local/cuda/targets/x86_64-linux/lib -L/usr/lib/wsl/lib 
I CC:        cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
I CXX:       g++ (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
I NVCC:      Build cuda_12.2.r12.2/compiler.33191640_0

make: Nothing to be done for 'default'.
Log start
clip_model_load: model name:   openai/clip-vit-large-patch14-336
clip_model_load: description:  image encoder for LLaVA
clip_model_load: GGUF version: 3
clip_model_load: alignment:    32
clip_model_load: n_tensors:    378
clip_model_load: n_kv:         19
clip_model_load: ftype:        f16

clip_model_load: loaded meta data with 19 key-value pairs and 378 tensors from ./models/llava-7b-v1.6/mmproj-model-f16.gguf
clip_model_load: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
clip_model_load: - kv   0:                       general.architecture str              = clip
clip_model_load: - kv   1:                      clip.has_text_encoder bool             = false
clip_model_load: - kv   2:                    clip.has_vision_encoder bool             = true
clip_model_load: - kv   3:                   clip.has_llava_projector bool             = true
clip_model_load: - kv   4:                          general.file_type u32              = 1
clip_model_load: - kv   5:                               general.name str              = openai/clip-vit-large-patch14-336
clip_model_load: - kv   6:                        general.description str              = image encoder for LLaVA
clip_model_load: - kv   7:                        clip.projector_type str              = mlp
clip_model_load: - kv   8:                     clip.vision.image_size u32              = 336
clip_model_load: - kv   9:                     clip.vision.patch_size u32              = 14
clip_model_load: - kv  10:               clip.vision.embedding_length u32              = 1024
clip_model_load: - kv  11:            clip.vision.feed_forward_length u32              = 4096
clip_model_load: - kv  12:                 clip.vision.projection_dim u32              = 768
clip_model_load: - kv  13:           clip.vision.attention.head_count u32              = 16
clip_model_load: - kv  14:   clip.vision.attention.layer_norm_epsilon f32              = 0.000010
clip_model_load: - kv  15:                    clip.vision.block_count u32              = 23
clip_model_load: - kv  16:                     clip.vision.image_mean arr[f32,3]       = [0.481455, 0.457828, 0.408211]
clip_model_load: - kv  17:                      clip.vision.image_std arr[f32,3]       = [0.268630, 0.261303, 0.275777]
clip_model_load: - kv  18:                              clip.use_gelu bool             = false
clip_model_load: - type  f32:  236 tensors
clip_model_load: - type  f16:  142 tensors
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:   no
ggml_cuda_init: CUDA_USE_TENSOR_CORES: yes
ggml_cuda_init: found 1 CUDA devices:
  Device 0: NVIDIA GeForce RTX 2060 SUPER, compute capability 7.5, VMM: yes
clip_model_load: CLIP using CUDA backend
clip_model_load: text_encoder:   0
clip_model_load: vision_encoder: 1
clip_model_load: llava_projector:  1
clip_model_load: model size:     595.50 MB
clip_model_load: metadata size:  0.14 MB
clip_model_load: params backend buffer size =  595.50 MB (378 tensors)
key clip.vision.image_grid_pinpoints not found in file
key clip.vision.mm_patch_merge_type not found in file
key clip.vision.image_crop_resolution not found in file
clip_model_load: compute allocated memory: 32.89 MB
llama_model_loader: loaded meta data with 24 key-value pairs and 291 tensors from ./models/llava-7b-v1.6/ggml-model-q4_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              = llama
llama_model_loader: - kv   1:                               general.name str              = huggingface
llama_model_loader: - kv   2:                       llama.context_length u32              = 32768
llama_model_loader: - kv   3:                     llama.embedding_length u32              = 4096
llama_model_loader: - kv   4:                          llama.block_count u32              = 32
llama_model_loader: - kv   5:                  llama.feed_forward_length u32              = 14336
llama_model_loader: - kv   6:                 llama.rope.dimension_count u32              = 128
llama_model_loader: - kv   7:                 llama.attention.head_count u32              = 32
llama_model_loader: - kv   8:              llama.attention.head_count_kv u32              = 8
llama_model_loader: - kv   9:     llama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  10:                       llama.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  11:                          general.file_type u32              = 2
llama_model_loader: - kv  12:                       tokenizer.ggml.model str              = llama
llama_model_loader: - kv  13:                      tokenizer.ggml.tokens arr[str,32000]   = ["<unk>", "<s>", "</s>", "<0x00>", "<...
llama_model_loader: - kv  14:                      tokenizer.ggml.scores arr[f32,32000]   = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv  15:                  tokenizer.ggml.token_type arr[i32,32000]   = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...
llama_model_loader: - kv  16:                tokenizer.ggml.bos_token_id u32              = 1
llama_model_loader: - kv  17:                tokenizer.ggml.eos_token_id u32              = 2
llama_model_loader: - kv  18:            tokenizer.ggml.unknown_token_id u32              = 0
llama_model_loader: - kv  19:            tokenizer.ggml.padding_token_id u32              = 0
llama_model_loader: - kv  20:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  21:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  22:                    tokenizer.chat_template str              = {{ bos_token }}{% for message in mess...
llama_model_loader: - kv  23:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:   65 tensors
llama_model_loader: - type q4_0:  225 tensors
llama_model_loader: - type q6_K:    1 tensors
llm_load_vocab: special tokens definition check successful ( 259/32000 ).
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = llama
llm_load_print_meta: vocab type       = SPM
llm_load_print_meta: n_vocab          = 32000
llm_load_print_meta: n_merges         = 0
llm_load_print_meta: n_ctx_train      = 32768
llm_load_print_meta: n_embd           = 4096
llm_load_print_meta: n_head           = 32
llm_load_print_meta: n_head_kv        = 8
llm_load_print_meta: n_layer          = 32
llm_load_print_meta: n_rot            = 128
llm_load_print_meta: n_embd_head_k    = 128
llm_load_print_meta: n_embd_head_v    = 128
llm_load_print_meta: n_gqa            = 4
llm_load_print_meta: n_embd_k_gqa     = 1024
llm_load_print_meta: n_embd_v_gqa     = 1024
llm_load_print_meta: f_norm_eps       = 0.0e+00
llm_load_print_meta: f_norm_rms_eps   = 1.0e-05
llm_load_print_meta: f_clamp_kqv      = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale    = 0.0e+00
llm_load_print_meta: n_ff             = 14336
llm_load_print_meta: n_expert         = 0
llm_load_print_meta: n_expert_used    = 0
llm_load_print_meta: causal attn      = 1
llm_load_print_meta: pooling type     = 0
llm_load_print_meta: rope type        = 0
llm_load_print_meta: rope scaling     = linear
llm_load_print_meta: freq_base_train  = 1000000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_yarn_orig_ctx  = 32768
llm_load_print_meta: rope_finetuned   = unknown
llm_load_print_meta: ssm_d_conv       = 0
llm_load_print_meta: ssm_d_inner      = 0
llm_load_print_meta: ssm_d_state      = 0
llm_load_print_meta: ssm_dt_rank      = 0
llm_load_print_meta: model type       = 8B
llm_load_print_meta: model ftype      = Q4_0
llm_load_print_meta: model params     = 7.24 B
llm_load_print_meta: model size       = 3.83 GiB (4.54 BPW) 
llm_load_print_meta: general.name     = huggingface
llm_load_print_meta: BOS token        = 1 '<s>'
llm_load_print_meta: EOS token        = 2 '</s>'
llm_load_print_meta: UNK token        = 0 '<unk>'
llm_load_print_meta: PAD token        = 0 '<unk>'
llm_load_print_meta: LF token         = 13 '<0x0A>'
llm_load_tensors: ggml ctx size =    0.30 MiB
llm_load_tensors: offloading 32 repeating layers to GPU
llm_load_tensors: offloading non-repeating layers to GPU
llm_load_tensors: offloaded 33/33 layers to GPU
llm_load_tensors:        CPU buffer size =    70.31 MiB
llm_load_tensors:      CUDA0 buffer size =  3847.55 MiB
..................................................................................................
llama_new_context_with_model: n_ctx      = 2048
llama_new_context_with_model: n_batch    = 2048
llama_new_context_with_model: n_ubatch   = 512
llama_new_context_with_model: freq_base  = 1000000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init:      CUDA0 KV buffer size =   256.00 MiB
llama_new_context_with_model: KV self size  =  256.00 MiB, K (f16):  128.00 MiB, V (f16):  128.00 MiB
llama_new_context_with_model:  CUDA_Host  output buffer size =     0.12 MiB
llama_new_context_with_model:      CUDA0 compute buffer size =   164.00 MiB
llama_new_context_with_model:  CUDA_Host compute buffer size =    12.01 MiB
llama_new_context_with_model: graph nodes  = 1030
llama_new_context_with_model: graph splits = 2
encode_image_with_clip: image embedding created: 576 tokens

encode_image_with_clip: image encoded in    95.39 ms by CLIP (    0.17 ms per image patch)

 The animal, a cat, is sitting on a couch and holding a knife. 

llama_print_timings:        load time =    1715.01 ms
llama_print_timings:      sample time =       0.25 ms /    18 runs   (    0.01 ms per token, 71146.25 tokens per second)
llama_print_timings: prompt eval time =     742.14 ms /   620 tokens (    1.20 ms per token,   835.43 tokens per second)
llama_print_timings:        eval time =     322.39 ms /    17 runs   (   18.96 ms per token,    52.73 tokens per second)
llama_print_timings:       total time =    2043.04 ms /   637 tokens

[rezacopol]

Any idea/pointer where I should look? not sure why image encoder tower seems to be run on CPU :/

[Jeximo]

Any idea/pointer where I should look?

@rezacopol It's unclear what parameters you're using. My guess is you didn't enable GPU, ie. -ngl 99:

llama.cpp/examples/main/README.md

Line 318 in 4e9a7f7

- `-ngl N, --n-gpu-layers N`: When compiled with appropriate support (currently CLBlast or cuBLAS), this option allows offloading some layers to the GPU for computation. Generally results in increased performance.

[rezacopol]

I did that and make sure things are offloaded to gpu using nvidia-smi command, here is the full command and log I get. I also tried with a machine with multiple GPU and found running it with -sm none helps.

./llava-cli -m ../ggml_llava-v1.5-7b/ggml-model-q5_k.gguf --mmproj ../ggml_llava-v1.5-7b/mmproj-model-f16.gguf --image ../photos/kitchen_4p.jpg -p "describe the photo in detail" -ngl 100 -sm none

Log start
llama_model_loader: loaded meta data with 19 key-value pairs and 291 tensors from ../ggml_llava-v1.5-7b/ggml-model-q5_k.gguf (version GGUF V2)
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv   0:                       general.architecture str              = llama
llama_model_loader: - kv   1:                               general.name str              = LLaMA v2
llama_model_loader: - kv   2:                       llama.context_length u32              = 4096
llama_model_loader: - kv   3:                     llama.embedding_length u32              = 4096
llama_model_loader: - kv   4:                          llama.block_count u32              = 32
llama_model_loader: - kv   5:                  llama.feed_forward_length u32              = 11008
llama_model_loader: - kv   6:                 llama.rope.dimension_count u32              = 128
llama_model_loader: - kv   7:                 llama.attention.head_count u32              = 32
llama_model_loader: - kv   8:              llama.attention.head_count_kv u32              = 32
llama_model_loader: - kv   9:     llama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  10:                          general.file_type u32              = 17
llama_model_loader: - kv  11:                       tokenizer.ggml.model str              = llama
llama_model_loader: - kv  12:                      tokenizer.ggml.tokens arr[str,32000]   = ["<unk>", "<s>", "</s>", "<0x00>", "<...
llama_model_loader: - kv  13:                      tokenizer.ggml.scores arr[f32,32000]   = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv  14:                  tokenizer.ggml.token_type arr[i32,32000]   = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...
llama_model_loader: - kv  15:                tokenizer.ggml.bos_token_id u32              = 1
llama_model_loader: - kv  16:                tokenizer.ggml.eos_token_id u32              = 2
llama_model_loader: - kv  17:            tokenizer.ggml.padding_token_id u32              = 0
llama_model_loader: - kv  18:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:   65 tensors
llama_model_loader: - type q5_K:  193 tensors
llama_model_loader: - type q6_K:   33 tensors
llm_load_vocab: special tokens definition check successful ( 259/32000 ).
llm_load_print_meta: format           = GGUF V2
llm_load_print_meta: arch             = llama
llm_load_print_meta: vocab type       = SPM
llm_load_print_meta: n_vocab          = 32000
llm_load_print_meta: n_merges         = 0
llm_load_print_meta: n_ctx_train      = 4096
llm_load_print_meta: n_embd           = 4096
llm_load_print_meta: n_head           = 32
llm_load_print_meta: n_head_kv        = 32
llm_load_print_meta: n_layer          = 32
llm_load_print_meta: n_rot            = 128
llm_load_print_meta: n_embd_head_k    = 128
llm_load_print_meta: n_embd_head_v    = 128
llm_load_print_meta: n_gqa            = 1
llm_load_print_meta: n_embd_k_gqa     = 4096
llm_load_print_meta: n_embd_v_gqa     = 4096
llm_load_print_meta: f_norm_eps       = 0.0e+00
llm_load_print_meta: f_norm_rms_eps   = 1.0e-05
llm_load_print_meta: f_clamp_kqv      = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale    = 0.0e+00
llm_load_print_meta: n_ff             = 11008
llm_load_print_meta: n_expert         = 0
llm_load_print_meta: n_expert_used    = 0
llm_load_print_meta: causal attn      = 1
llm_load_print_meta: pooling type     = 0
llm_load_print_meta: rope type        = 0
llm_load_print_meta: rope scaling     = linear
llm_load_print_meta: freq_base_train  = 10000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_yarn_orig_ctx  = 4096
llm_load_print_meta: rope_finetuned   = unknown
llm_load_print_meta: ssm_d_conv       = 0
llm_load_print_meta: ssm_d_inner      = 0
llm_load_print_meta: ssm_d_state      = 0
llm_load_print_meta: ssm_dt_rank      = 0
llm_load_print_meta: model type       = 7B
llm_load_print_meta: model ftype      = Q5_K - Medium
llm_load_print_meta: model params     = 6.74 B
llm_load_print_meta: model size       = 4.45 GiB (5.68 BPW) 
llm_load_print_meta: general.name     = LLaMA v2
llm_load_print_meta: BOS token        = 1 '<s>'
llm_load_print_meta: EOS token        = 2 '</s>'
llm_load_print_meta: UNK token        = 0 '<unk>'
llm_load_print_meta: PAD token        = 0 '<unk>'
llm_load_print_meta: LF token         = 13 '<0x0A>'
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:   no
ggml_cuda_init: CUDA_USE_TENSOR_CORES: yes
ggml_cuda_init: found 4 CUDA devices:
  Device 0: NVIDIA A100-SXM4-80GB, compute capability 8.0, VMM: yes
  Device 1: NVIDIA A100-SXM4-80GB, compute capability 8.0, VMM: yes
  Device 2: NVIDIA A100-SXM4-80GB, compute capability 8.0, VMM: yes
  Device 3: NVIDIA A100-SXM4-80GB, compute capability 8.0, VMM: yes
llm_load_tensors: ggml ctx size =    0.30 MiB
llm_load_tensors: offloading 32 repeating layers to GPU
llm_load_tensors: offloading non-repeating layers to GPU
llm_load_tensors: offloaded 33/33 layers to GPU
llm_load_tensors:        CPU buffer size =    85.94 MiB
llm_load_tensors:      CUDA0 buffer size =  4474.94 MiB
..................................................................................................
clip_model_load: model name:   openai/clip-vit-large-patch14-336
clip_model_load: description:  image encoder for LLaVA
clip_model_load: GGUF version: 2
clip_model_load: alignment:    32
clip_model_load: n_tensors:    377
clip_model_load: n_kv:         18
clip_model_load: ftype:        f16

clip_model_load: loaded meta data with 18 key-value pairs and 377 tensors from ../ggml_llava-v1.5-7b/mmproj-model-f16.gguf
clip_model_load: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
clip_model_load: - kv   0:                       general.architecture str              = clip
clip_model_load: - kv   1:                      clip.has_text_encoder bool             = false
clip_model_load: - kv   2:                    clip.has_vision_encoder bool             = true
clip_model_load: - kv   3:                   clip.has_llava_projector bool             = true
clip_model_load: - kv   4:                          general.file_type u32              = 1
clip_model_load: - kv   5:                               general.name str              = openai/clip-vit-large-patch14-336
clip_model_load: - kv   6:                        general.description str              = image encoder for LLaVA
clip_model_load: - kv   7:                     clip.vision.image_size u32              = 336
clip_model_load: - kv   8:                     clip.vision.patch_size u32              = 14
clip_model_load: - kv   9:               clip.vision.embedding_length u32              = 1024
clip_model_load: - kv  10:            clip.vision.feed_forward_length u32              = 4096
clip_model_load: - kv  11:                 clip.vision.projection_dim u32              = 768
clip_model_load: - kv  12:           clip.vision.attention.head_count u32              = 16
clip_model_load: - kv  13:   clip.vision.attention.layer_norm_epsilon f32              = 0.000010
clip_model_load: - kv  14:                    clip.vision.block_count u32              = 23
clip_model_load: - kv  15:                     clip.vision.image_mean arr[f32,3]       = [0.481455, 0.457828, 0.408211]
clip_model_load: - kv  16:                      clip.vision.image_std arr[f32,3]       = [0.268630, 0.261303, 0.275777]
clip_model_load: - kv  17:                              clip.use_gelu bool             = false
clip_model_load: - type  f32:  235 tensors
clip_model_load: - type  f16:  142 tensors
clip_model_load: CLIP using CUDA backend
clip_model_load: text_encoder:   0
clip_model_load: vision_encoder: 1
clip_model_load: llava_projector:  1
clip_model_load: model size:     595.49 MB
clip_model_load: metadata size:  0.14 MB
clip_model_load: params backend buffer size =  595.49 MB (377 tensors)
key clip.vision.image_grid_pinpoints not found in file
key clip.vision.mm_patch_merge_type not found in file
key clip.vision.image_crop_resolution not found in file
clip_model_load: compute allocated memory: 32.89 MB
llama_new_context_with_model: n_ctx      = 2048
llama_new_context_with_model: n_batch    = 2048
llama_new_context_with_model: n_ubatch   = 512
llama_new_context_with_model: flash_attn = 0
llama_new_context_with_model: freq_base  = 10000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init:      CUDA0 KV buffer size =  1024.00 MiB
llama_new_context_with_model: KV self size  = 1024.00 MiB, K (f16):  512.00 MiB, V (f16):  512.00 MiB
llama_new_context_with_model:  CUDA_Host  output buffer size =     0.12 MiB
llama_new_context_with_model:      CUDA0 compute buffer size =   164.00 MiB
llama_new_context_with_model:  CUDA_Host compute buffer size =    12.01 MiB
llama_new_context_with_model: graph nodes  = 1030
llama_new_context_with_model: graph splits = 2
encode_image_with_clip: image embedding created: 576 tokens

encode_image_with_clip: image encoded in  3527.74 ms by CLIP (    6.12 ms per image patch)

 The image depicts a group of five people gathered in a kitchen. They are standing around the kitchen counter, engaged in conversation and enjoying each other's company. Various items are visible on the counter, such as a few wine glasses, a bowl, a knife, and a spoon. 

There are also a couple of bottles, a vase, a sink, and an oven in the kitchen. A potted plant is placed near the oven, adding a touch of greenery to the space. The atmosphere appears to be casual and friendly, with the group enjoying the time spent together in the kitchen.

llama_print_timings:        load time =    8921.92 ms
llama_print_timings:      sample time =       5.24 ms /   137 runs   (    0.04 ms per token, 26130.08 tokens per second)
llama_print_timings: prompt eval time =     205.79 ms /   620 tokens (    0.33 ms per token,  3012.74 tokens per second)
llama_print_timings:        eval time =    1180.02 ms /   136 runs   (    8.68 ms per token,   115.25 tokens per second)
llama_print_timings:       total time =   10146.64 ms /   756 tokens

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.