github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 08:59:13 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #10011] OLLAMA_NUM_PARALLEL not working #6563

New Issue
Closed
opened 2026-04-12 18:11:28 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-12 18:11:28 -05:00
Owner
Copy Link

Originally created by @forReason on GitHub (Mar 27, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10011

What is the issue?

OLLAMA_NUM_PARALLEL is not working. It just loads one instance.

Image

Image

Relevant log output

2025/03/27 08:03:56 routes.go:1205: INFO server config env="map[CUDA_VISIBLE_DEVICES:0,1 GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\\Users\\devops01\\.ollama\\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]"
time=2025-03-27T08:03:56.510+01:00 level=INFO source=images.go:432 msg="total blobs: 12"
time=2025-03-27T08:03:56.511+01:00 level=INFO source=images.go:439 msg="total unused blobs removed: 0"
time=2025-03-27T08:03:56.512+01:00 level=INFO source=routes.go:1256 msg="Listening on [::]:11434 (version 0.5.12)"
time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu.go:217 msg="looking for compatible GPUs"
time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:167 msg=packages count=4
time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:214 msg="" package=0 cores=2 efficiency=0 threads=2
time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:214 msg="" package=1 cores=2 efficiency=0 threads=2
time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:214 msg="" package=2 cores=2 efficiency=0 threads=2
time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:214 msg="" package=3 cores=2 efficiency=0 threads=2
time=2025-03-27T08:03:56.910+01:00 level=INFO source=gpu.go:319 msg="detected OS VRAM overhead" id=GPU-cae06f2c-27fb-11b2-9c39-98c46e11c22e library=cuda compute=8.6 driver=11.4 name="NVIDIA A40-48Q" overhead="907.2 MiB"
time=2025-03-27T08:03:57.222+01:00 level=INFO source=gpu.go:319 msg="detected OS VRAM overhead" id=GPU-cdb4fbea-27fb-11b2-847d-42035a3548d7 library=cuda compute=8.6 driver=11.4 name="NVIDIA A40-48Q" overhead="932.5 MiB"
time=2025-03-27T08:03:57.226+01:00 level=INFO source=types.go:130 msg="inference compute" id=GPU-cae06f2c-27fb-11b2-9c39-98c46e11c22e library=cuda variant=v11 compute=8.6 driver=11.4 name="NVIDIA A40-48Q" total="48.0 GiB" available="42.9 GiB"
time=2025-03-27T08:03:57.226+01:00 level=INFO source=types.go:130 msg="inference compute" id=GPU-cdb4fbea-27fb-11b2-847d-42035a3548d7 library=cuda variant=v11 compute=8.6 driver=11.4 name="NVIDIA A40-48Q" total="48.0 GiB" available="42.9 GiB"
[GIN] 2025/03/27 - 08:03:57 | 200 |            0s |       127.0.0.1 | HEAD     "/"
time=2025-03-27T08:03:57.229+01:00 level=WARN source=routes.go:901 msg="bad manifest filepath" name=registry.ollama.ai/library/llama_vision_fp6_custom:latest error="open C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-069235addef1211ea9efd620985e88763b48096fffe53803d9e61a39af9fc866: The system cannot find the file specified."
[GIN] 2025/03/27 - 08:03:57 | 200 |      2.2242ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2025/03/27 - 08:04:01 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 08:05:22 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 08:05:22 | 200 |     33.1633ms |       127.0.0.1 | POST     "/api/generate"
[GIN] 2025/03/27 - 08:05:22 | 200 |      22.717ms |       127.0.0.1 | DELETE   "/api/delete"
[GIN] 2025/03/27 - 10:24:46 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:24:46 | 200 |     19.0364ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2025/03/27 - 10:24:57 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:24:57 | 404 |       2.786ms |       127.0.0.1 | POST     "/api/show"
time=2025-03-27T10:25:09.966+01:00 level=INFO source=images.go:669 msg="request failed: Get \"https://registry.ollama.ai/v2/library/llama_vision_fp6_custom/manifests/latest\": dial tcp: lookup registry.ollama.ai: no such host"
[GIN] 2025/03/27 - 10:25:09 | 200 |   12.0138758s |       127.0.0.1 | POST     "/api/pull"
[GIN] 2025/03/27 - 10:26:41 | 200 |        65.3µs |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:26:41 | 404 |      4.4727ms |       127.0.0.1 | POST     "/api/show"
time=2025-03-27T10:26:53.952+01:00 level=INFO source=images.go:669 msg="request failed: Get \"https://registry.ollama.ai/v2/library/llama_vision_fp6_custom/manifests/latest\": dial tcp: lookup registry.ollama.ai: no such host"
[GIN] 2025/03/27 - 10:26:53 | 200 |   12.0130116s |       127.0.0.1 | POST     "/api/pull"
[GIN] 2025/03/27 - 10:32:12 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:32:12 | 200 |     74.9073ms |       127.0.0.1 | POST     "/api/show"
time=2025-03-27T10:32:12.953+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
time=2025-03-27T10:32:13.040+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128
time=2025-03-27T10:32:13.040+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128
time=2025-03-27T10:32:13.043+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128
time=2025-03-27T10:32:13.043+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128
time=2025-03-27T10:32:13.051+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128
time=2025-03-27T10:32:13.051+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128
time=2025-03-27T10:32:13.053+01:00 level=INFO source=sched.go:731 msg="new model will fit in available VRAM, loading" model=C:\Users\devops01\.ollama\models\blobs\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 library=cuda parallel=1 required="76.8 GiB"
time=2025-03-27T10:32:13.084+01:00 level=INFO source=server.go:97 msg="system memory" total="256.0 GiB" free="250.0 GiB" free_swap="284.7 GiB"
time=2025-03-27T10:32:13.087+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128
time=2025-03-27T10:32:13.087+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128
time=2025-03-27T10:32:13.087+01:00 level=INFO source=server.go:130 msg=offload library=cuda layers.requested=-1 layers.model=101 layers.offload=101 layers.split=51,50 memory.available="[42.9 GiB 42.8 GiB]" memory.gpu_overhead="0 B" memory.required.full="76.8 GiB" memory.required.partial="76.8 GiB" memory.required.kv="1.6 GiB" memory.required.allocations="[41.1 GiB 35.7 GiB]" memory.weights.total="67.0 GiB" memory.weights.repeating="66.2 GiB" memory.weights.nonrepeating="822.0 MiB" memory.graph.full="1.1 GiB" memory.graph.partial="1.1 GiB" projector.weights="1.9 GiB" projector.graph="2.8 GiB"
time=2025-03-27T10:32:13.095+01:00 level=INFO source=server.go:380 msg="starting llama server" cmd="C:\\Users\\devops01\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --model C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 --ctx-size 2048 --batch-size 512 --n-gpu-layers 101 --mmproj C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-6b6c374d159e097509b33e9fda648c178c903959fc0c7dbfae487cc8d958093e --threads 8 --no-mmap --parallel 1 --tensor-split 51,50 --port 62648"
time=2025-03-27T10:32:13.102+01:00 level=INFO source=sched.go:450 msg="loaded runners" count=1
time=2025-03-27T10:32:13.102+01:00 level=INFO source=server.go:557 msg="waiting for llama runner to start responding"
time=2025-03-27T10:32:13.103+01:00 level=INFO source=server.go:591 msg="waiting for server to become available" status="llm server error"
time=2025-03-27T10:32:13.139+01:00 level=INFO source=runner.go:932 msg="starting go runner"
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 CUDA devices:
  Device 0: NVIDIA A40-48Q, compute capability 8.6, VMM: no
  Device 1: NVIDIA A40-48Q, compute capability 8.6, VMM: no
load_backend: loaded CUDA backend from C:\Users\devops01\AppData\Local\Programs\Ollama\lib\ollama\cuda_v11\ggml-cuda.dll
load_backend: loaded CPU backend from C:\Users\devops01\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-icelake.dll
time=2025-03-27T10:32:14.460+01:00 level=INFO source=runner.go:935 msg=system info="CPU : LLAMAFILE = 1 | CPU : LLAMAFILE = 1 | CUDA : USE_GRAPHS = 1 | PEER_MAX_BATCH_SIZE = 128 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | cgo(clang)" threads=8
time=2025-03-27T10:32:14.462+01:00 level=INFO source=runner.go:993 msg="Server listening on 127.0.0.1:62648"
time=2025-03-27T10:32:14.611+01:00 level=INFO source=server.go:591 msg="waiting for server to become available" status="llm server loading model"
llama_load_model_from_file: using device CUDA0 (NVIDIA A40-48Q) - 43830 MiB free
llama_load_model_from_file: using device CUDA1 (NVIDIA A40-48Q) - 43830 MiB free
llama_model_loader: loaded meta data with 27 key-value pairs and 984 tensors from C:\Users\devops01\.ollama\models\blobs\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 (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              = mllama
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Model
llama_model_loader: - kv   3:                         general.size_label str              = 88B
llama_model_loader: - kv   4:                         mllama.block_count u32              = 100
llama_model_loader: - kv   5:                      mllama.context_length u32              = 131072
llama_model_loader: - kv   6:                    mllama.embedding_length u32              = 8192
llama_model_loader: - kv   7:                 mllama.feed_forward_length u32              = 28672
llama_model_loader: - kv   8:                mllama.attention.head_count u32              = 64
llama_model_loader: - kv   9:             mllama.attention.head_count_kv u32              = 8
llama_model_loader: - kv  10:                      mllama.rope.freq_base f32              = 500000.000000
llama_model_loader: - kv  11:    mllama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  12:                          general.file_type u32              = 18
llama_model_loader: - kv  13:                          mllama.vocab_size u32              = 128256
llama_model_loader: - kv  14:                mllama.rope.dimension_count u32              = 128
llama_model_loader: - kv  15:    mllama.attention.cross_attention_layers arr[i32,20]      = [3, 8, 13, 18, 23, 28, 33, 38, 43, 48...
llama_model_loader: - kv  16:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  17:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  18:                         tokenizer.ggml.pre str              = llama-bpe
llama_model_loader: - kv  19:                      tokenizer.ggml.tokens arr[str,128257]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  20:                  tokenizer.ggml.token_type arr[i32,128257]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  21:                      tokenizer.ggml.merges arr[str,280147]  = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv  22:                tokenizer.ggml.bos_token_id u32              = 128000
llama_model_loader: - kv  23:                tokenizer.ggml.eos_token_id u32              = 128009
llama_model_loader: - kv  24:            tokenizer.ggml.padding_token_id u32              = 128004
llama_model_loader: - kv  25:                    tokenizer.chat_template str              = {{- bos_token }}\n{%- if custom_tools ...
llama_model_loader: - kv  26:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:  282 tensors
llama_model_loader: - type q6_K:  702 tensors
llm_load_vocab: special tokens cache size = 257
llm_load_vocab: token to piece cache size = 0.7999 MB
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = mllama
llm_load_print_meta: vocab type       = BPE
llm_load_print_meta: n_vocab          = 128256
llm_load_print_meta: n_merges         = 280147
llm_load_print_meta: vocab_only       = 0
llm_load_print_meta: n_ctx_train      = 131072
llm_load_print_meta: n_embd           = 8192
llm_load_print_meta: n_layer          = 100
llm_load_print_meta: n_head           = 64
llm_load_print_meta: n_head_kv        = 8
llm_load_print_meta: n_rot            = 128
llm_load_print_meta: n_swa            = 0
llm_load_print_meta: n_embd_head_k    = 128
llm_load_print_meta: n_embd_head_v    = 128
llm_load_print_meta: n_gqa            = 8
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             = 28672
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  = 500000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn  = 131072
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: ssm_dt_b_c_rms   = 0
llm_load_print_meta: model type       = ?B
llm_load_print_meta: model ftype      = Q6_K
llm_load_print_meta: model params     = 87.67 B
llm_load_print_meta: model size       = 66.98 GiB (6.56 BPW) 
llm_load_print_meta: general.name     = Model
llm_load_print_meta: BOS token        = 128000 '<|begin_of_text|>'
llm_load_print_meta: EOS token        = 128009 '<|eot_id|>'
llm_load_print_meta: EOT token        = 128009 '<|eot_id|>'
llm_load_print_meta: EOM token        = 128008 '<|eom_id|>'
llm_load_print_meta: PAD token        = 128004 '<|finetune_right_pad_id|>'
llm_load_print_meta: LF token         = 128 'Ä'
llm_load_print_meta: EOG token        = 128008 '<|eom_id|>'
llm_load_print_meta: EOG token        = 128009 '<|eot_id|>'
llm_load_print_meta: max token length = 256
llama_model_load: vocab mismatch 128256 !- 128257 ...
llm_load_tensors: offloading 100 repeating layers to GPU
llm_load_tensors: offloading output layer to GPU
llm_load_tensors: offloaded 101/101 layers to GPU
llm_load_tensors:          CPU model buffer size =   822.00 MiB
llm_load_tensors:        CUDA0 model buffer size = 34141.32 MiB
llm_load_tensors:        CUDA1 model buffer size = 33624.43 MiB
llama_new_context_with_model: n_seq_max     = 1
llama_new_context_with_model: n_ctx         = 2048
llama_new_context_with_model: n_ctx_per_seq = 2048
llama_new_context_with_model: n_batch       = 512
llama_new_context_with_model: n_ubatch      = 512
llama_new_context_with_model: flash_attn    = 0
llama_new_context_with_model: freq_base     = 500000.0
llama_new_context_with_model: freq_scale    = 1
llama_new_context_with_model: n_ctx_per_seq (2048) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
llama_kv_cache_init: kv_size = 2048, offload = 1, type_k = 'f16', type_v = 'f16', n_layer = 100, can_shift = 1
llama_kv_cache_init:      CUDA0 KV buffer size =   828.31 MiB
llama_kv_cache_init:      CUDA1 KV buffer size =   812.31 MiB
llama_new_context_with_model: KV self size  = 1640.62 MiB, K (f16):  820.31 MiB, V (f16):  820.31 MiB
llama_new_context_with_model:  CUDA_Host  output buffer size =     0.52 MiB
llama_new_context_with_model: pipeline parallelism enabled (n_copies=4)
llama_new_context_with_model:      CUDA0 compute buffer size =   400.01 MiB
llama_new_context_with_model:      CUDA1 compute buffer size =   400.02 MiB
llama_new_context_with_model:  CUDA_Host compute buffer size =    32.02 MiB
llama_new_context_with_model: graph nodes  = 2566
llama_new_context_with_model: graph splits = 3
mllama_model_load: model name:   Llama-3.2-90B-Vision-Instruct
mllama_model_load: description:  vision encoder for Mllama
mllama_model_load: GGUF version: 3
mllama_model_load: alignment:    32
mllama_model_load: n_tensors:    512
mllama_model_load: n_kv:         17
mllama_model_load: ftype:        f16
mllama_model_load: 
mllama_model_load: mllama_model_load: using CUDA0 backend

mllama_model_load: compute allocated memory: 2853.34 MB
time=2025-03-27T10:32:49.460+01:00 level=INFO source=server.go:596 msg="llama runner started in 36.36 seconds"
[GIN] 2025/03/27 - 10:32:49 | 200 |   36.5347842s |       127.0.0.1 | POST     "/api/generate"
[GIN] 2025/03/27 - 10:33:07 | 200 |        85.6µs |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:33:07 | 200 |       635.5µs |       127.0.0.1 | GET      "/api/ps"
[GIN] 2025/03/27 - 10:33:34 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:33:34 | 200 |      33.761ms |       127.0.0.1 | POST     "/api/show"
time=2025-03-27T10:33:34.857+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
[GIN] 2025/03/27 - 10:33:34 | 200 |     37.4451ms |       127.0.0.1 | POST     "/api/generate"
[GIN] 2025/03/27 - 10:34:00 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:34:00 | 200 |     32.5477ms |       127.0.0.1 | POST     "/api/show"
time=2025-03-27T10:34:00.626+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
[GIN] 2025/03/27 - 10:34:00 | 200 |     30.9911ms |       127.0.0.1 | POST     "/api/generate"
[GIN] 2025/03/27 - 10:34:16 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:34:16 | 200 |     33.7433ms |       127.0.0.1 | POST     "/api/show"
time=2025-03-27T10:34:16.250+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
[GIN] 2025/03/27 - 10:34:16 | 200 |     33.6178ms |       127.0.0.1 | POST     "/api/generate"
time=2025-03-27T10:34:38.089+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
time=2025-03-27T10:34:39.562+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
time=2025-03-27T10:34:40.185+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
time=2025-03-27T10:34:40.590+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
[GIN] 2025/03/27 - 10:35:07 | 200 |    29.877435s |       127.0.0.1 | POST     "/api/chat"
[GIN] 2025/03/27 - 10:35:48 | 200 |          1m9s |       127.0.0.1 | POST     "/api/chat"
[GIN] 2025/03/27 - 10:38:28 | 200 |         3m47s |       127.0.0.1 | POST     "/api/chat"
[GIN] 2025/03/27 - 10:39:05 | 200 |         4m25s |       127.0.0.1 | POST     "/api/chat"
[GIN] 2025/03/27 - 10:39:18 | 200 |        49.6µs |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:39:18 | 200 |      4.2162ms |       127.0.0.1 | POST     "/api/generate"
[GIN] 2025/03/27 - 10:39:23 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:39:23 | 200 |     30.8407ms |       127.0.0.1 | POST     "/api/show"
time=2025-03-27T10:39:23.738+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet"
time=2025-03-27T10:39:23.827+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128
time=2025-03-27T10:39:23.827+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128
time=2025-03-27T10:39:23.830+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128
time=2025-03-27T10:39:23.830+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128
time=2025-03-27T10:39:23.836+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128
time=2025-03-27T10:39:23.836+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128
time=2025-03-27T10:39:23.837+01:00 level=INFO source=sched.go:731 msg="new model will fit in available VRAM, loading" model=C:\Users\devops01\.ollama\models\blobs\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 library=cuda parallel=1 required="76.8 GiB"
time=2025-03-27T10:39:23.866+01:00 level=INFO source=server.go:97 msg="system memory" total="256.0 GiB" free="249.8 GiB" free_swap="284.6 GiB"
time=2025-03-27T10:39:23.869+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128
time=2025-03-27T10:39:23.869+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128
time=2025-03-27T10:39:23.870+01:00 level=INFO source=server.go:130 msg=offload library=cuda layers.requested=-1 layers.model=101 layers.offload=101 layers.split=51,50 memory.available="[42.9 GiB 42.9 GiB]" memory.gpu_overhead="0 B" memory.required.full="76.8 GiB" memory.required.partial="76.8 GiB" memory.required.kv="1.6 GiB" memory.required.allocations="[41.1 GiB 35.7 GiB]" memory.weights.total="67.0 GiB" memory.weights.repeating="66.2 GiB" memory.weights.nonrepeating="822.0 MiB" memory.graph.full="1.1 GiB" memory.graph.partial="1.1 GiB" projector.weights="1.9 GiB" projector.graph="2.8 GiB"
time=2025-03-27T10:39:23.876+01:00 level=INFO source=server.go:380 msg="starting llama server" cmd="C:\\Users\\devops01\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --model C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 --ctx-size 2048 --batch-size 512 --n-gpu-layers 101 --mmproj C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-6b6c374d159e097509b33e9fda648c178c903959fc0c7dbfae487cc8d958093e --threads 8 --no-mmap --parallel 1 --tensor-split 51,50 --port 62665"
time=2025-03-27T10:39:23.883+01:00 level=INFO source=sched.go:450 msg="loaded runners" count=1
time=2025-03-27T10:39:23.883+01:00 level=INFO source=server.go:557 msg="waiting for llama runner to start responding"
time=2025-03-27T10:39:23.884+01:00 level=INFO source=server.go:591 msg="waiting for server to become available" status="llm server error"
time=2025-03-27T10:39:23.921+01:00 level=INFO source=runner.go:932 msg="starting go runner"
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 CUDA devices:
  Device 0: NVIDIA A40-48Q, compute capability 8.6, VMM: no
  Device 1: NVIDIA A40-48Q, compute capability 8.6, VMM: no
load_backend: loaded CUDA backend from C:\Users\devops01\AppData\Local\Programs\Ollama\lib\ollama\cuda_v11\ggml-cuda.dll
load_backend: loaded CPU backend from C:\Users\devops01\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-icelake.dll
time=2025-03-27T10:39:24.073+01:00 level=INFO source=runner.go:935 msg=system info="CPU : LLAMAFILE = 1 | CPU : LLAMAFILE = 1 | CUDA : USE_GRAPHS = 1 | PEER_MAX_BATCH_SIZE = 128 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | cgo(clang)" threads=8
time=2025-03-27T10:39:24.073+01:00 level=INFO source=runner.go:993 msg="Server listening on 127.0.0.1:62665"
time=2025-03-27T10:39:24.136+01:00 level=INFO source=server.go:591 msg="waiting for server to become available" status="llm server loading model"
llama_load_model_from_file: using device CUDA0 (NVIDIA A40-48Q) - 43830 MiB free
llama_load_model_from_file: using device CUDA1 (NVIDIA A40-48Q) - 43830 MiB free
llama_model_loader: loaded meta data with 27 key-value pairs and 984 tensors from C:\Users\devops01\.ollama\models\blobs\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 (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              = mllama
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Model
llama_model_loader: - kv   3:                         general.size_label str              = 88B
llama_model_loader: - kv   4:                         mllama.block_count u32              = 100
llama_model_loader: - kv   5:                      mllama.context_length u32              = 131072
llama_model_loader: - kv   6:                    mllama.embedding_length u32              = 8192
llama_model_loader: - kv   7:                 mllama.feed_forward_length u32              = 28672
llama_model_loader: - kv   8:                mllama.attention.head_count u32              = 64
llama_model_loader: - kv   9:             mllama.attention.head_count_kv u32              = 8
llama_model_loader: - kv  10:                      mllama.rope.freq_base f32              = 500000.000000
llama_model_loader: - kv  11:    mllama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  12:                          general.file_type u32              = 18
llama_model_loader: - kv  13:                          mllama.vocab_size u32              = 128256
llama_model_loader: - kv  14:                mllama.rope.dimension_count u32              = 128
llama_model_loader: - kv  15:    mllama.attention.cross_attention_layers arr[i32,20]      = [3, 8, 13, 18, 23, 28, 33, 38, 43, 48...
llama_model_loader: - kv  16:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  17:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  18:                         tokenizer.ggml.pre str              = llama-bpe
llama_model_loader: - kv  19:                      tokenizer.ggml.tokens arr[str,128257]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  20:                  tokenizer.ggml.token_type arr[i32,128257]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  21:                      tokenizer.ggml.merges arr[str,280147]  = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv  22:                tokenizer.ggml.bos_token_id u32              = 128000
llama_model_loader: - kv  23:                tokenizer.ggml.eos_token_id u32              = 128009
llama_model_loader: - kv  24:            tokenizer.ggml.padding_token_id u32              = 128004
llama_model_loader: - kv  25:                    tokenizer.chat_template str              = {{- bos_token }}\n{%- if custom_tools ...
llama_model_loader: - kv  26:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:  282 tensors
llama_model_loader: - type q6_K:  702 tensors
llm_load_vocab: special tokens cache size = 257
llm_load_vocab: token to piece cache size = 0.7999 MB
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = mllama
llm_load_print_meta: vocab type       = BPE
llm_load_print_meta: n_vocab          = 128256
llm_load_print_meta: n_merges         = 280147
llm_load_print_meta: vocab_only       = 0
llm_load_print_meta: n_ctx_train      = 131072
llm_load_print_meta: n_embd           = 8192
llm_load_print_meta: n_layer          = 100
llm_load_print_meta: n_head           = 64
llm_load_print_meta: n_head_kv        = 8
llm_load_print_meta: n_rot            = 128
llm_load_print_meta: n_swa            = 0
llm_load_print_meta: n_embd_head_k    = 128
llm_load_print_meta: n_embd_head_v    = 128
llm_load_print_meta: n_gqa            = 8
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             = 28672
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  = 500000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn  = 131072
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: ssm_dt_b_c_rms   = 0
llm_load_print_meta: model type       = ?B
llm_load_print_meta: model ftype      = Q6_K
llm_load_print_meta: model params     = 87.67 B
llm_load_print_meta: model size       = 66.98 GiB (6.56 BPW) 
llm_load_print_meta: general.name     = Model
llm_load_print_meta: BOS token        = 128000 '<|begin_of_text|>'
llm_load_print_meta: EOS token        = 128009 '<|eot_id|>'
llm_load_print_meta: EOT token        = 128009 '<|eot_id|>'
llm_load_print_meta: EOM token        = 128008 '<|eom_id|>'
llm_load_print_meta: PAD token        = 128004 '<|finetune_right_pad_id|>'
llm_load_print_meta: LF token         = 128 'Ä'
llm_load_print_meta: EOG token        = 128008 '<|eom_id|>'
llm_load_print_meta: EOG token        = 128009 '<|eot_id|>'
llm_load_print_meta: max token length = 256
llama_model_load: vocab mismatch 128256 !- 128257 ...
llm_load_tensors: offloading 100 repeating layers to GPU
llm_load_tensors: offloading output layer to GPU
llm_load_tensors: offloaded 101/101 layers to GPU
llm_load_tensors:          CPU model buffer size =   822.00 MiB
llm_load_tensors:        CUDA0 model buffer size = 34141.32 MiB
llm_load_tensors:        CUDA1 model buffer size = 33624.43 MiB
[GIN] 2025/03/27 - 10:39:29 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/27 - 10:39:29 | 200 |            0s |       127.0.0.1 | GET      "/api/ps"

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.5.12

Originally created by @forReason on GitHub (Mar 27, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10011 ### What is the issue? OLLAMA_NUM_PARALLEL is not working. It just loads one instance. ![Image](https://github.com/user-attachments/assets/6dc26074-bda6-4853-ae66-1d7eaf61b079) ![Image](https://github.com/user-attachments/assets/1b11c5f8-4708-4e9e-911f-e2615a6ec7e9) ### Relevant log output ```shell 2025/03/27 08:03:56 routes.go:1205: INFO server config env="map[CUDA_VISIBLE_DEVICES:0,1 GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\\Users\\devops01\\.ollama\\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]" time=2025-03-27T08:03:56.510+01:00 level=INFO source=images.go:432 msg="total blobs: 12" time=2025-03-27T08:03:56.511+01:00 level=INFO source=images.go:439 msg="total unused blobs removed: 0" time=2025-03-27T08:03:56.512+01:00 level=INFO source=routes.go:1256 msg="Listening on [::]:11434 (version 0.5.12)" time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu.go:217 msg="looking for compatible GPUs" time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:167 msg=packages count=4 time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:214 msg="" package=0 cores=2 efficiency=0 threads=2 time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:214 msg="" package=1 cores=2 efficiency=0 threads=2 time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:214 msg="" package=2 cores=2 efficiency=0 threads=2 time=2025-03-27T08:03:56.512+01:00 level=INFO source=gpu_windows.go:214 msg="" package=3 cores=2 efficiency=0 threads=2 time=2025-03-27T08:03:56.910+01:00 level=INFO source=gpu.go:319 msg="detected OS VRAM overhead" id=GPU-cae06f2c-27fb-11b2-9c39-98c46e11c22e library=cuda compute=8.6 driver=11.4 name="NVIDIA A40-48Q" overhead="907.2 MiB" time=2025-03-27T08:03:57.222+01:00 level=INFO source=gpu.go:319 msg="detected OS VRAM overhead" id=GPU-cdb4fbea-27fb-11b2-847d-42035a3548d7 library=cuda compute=8.6 driver=11.4 name="NVIDIA A40-48Q" overhead="932.5 MiB" time=2025-03-27T08:03:57.226+01:00 level=INFO source=types.go:130 msg="inference compute" id=GPU-cae06f2c-27fb-11b2-9c39-98c46e11c22e library=cuda variant=v11 compute=8.6 driver=11.4 name="NVIDIA A40-48Q" total="48.0 GiB" available="42.9 GiB" time=2025-03-27T08:03:57.226+01:00 level=INFO source=types.go:130 msg="inference compute" id=GPU-cdb4fbea-27fb-11b2-847d-42035a3548d7 library=cuda variant=v11 compute=8.6 driver=11.4 name="NVIDIA A40-48Q" total="48.0 GiB" available="42.9 GiB" [GIN] 2025/03/27 - 08:03:57 | 200 | 0s | 127.0.0.1 | HEAD "/" time=2025-03-27T08:03:57.229+01:00 level=WARN source=routes.go:901 msg="bad manifest filepath" name=registry.ollama.ai/library/llama_vision_fp6_custom:latest error="open C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-069235addef1211ea9efd620985e88763b48096fffe53803d9e61a39af9fc866: The system cannot find the file specified." [GIN] 2025/03/27 - 08:03:57 | 200 | 2.2242ms | 127.0.0.1 | GET "/api/tags" [GIN] 2025/03/27 - 08:04:01 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 08:05:22 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 08:05:22 | 200 | 33.1633ms | 127.0.0.1 | POST "/api/generate" [GIN] 2025/03/27 - 08:05:22 | 200 | 22.717ms | 127.0.0.1 | DELETE "/api/delete" [GIN] 2025/03/27 - 10:24:46 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:24:46 | 200 | 19.0364ms | 127.0.0.1 | GET "/api/tags" [GIN] 2025/03/27 - 10:24:57 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:24:57 | 404 | 2.786ms | 127.0.0.1 | POST "/api/show" time=2025-03-27T10:25:09.966+01:00 level=INFO source=images.go:669 msg="request failed: Get \"https://registry.ollama.ai/v2/library/llama_vision_fp6_custom/manifests/latest\": dial tcp: lookup registry.ollama.ai: no such host" [GIN] 2025/03/27 - 10:25:09 | 200 | 12.0138758s | 127.0.0.1 | POST "/api/pull" [GIN] 2025/03/27 - 10:26:41 | 200 | 65.3µs | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:26:41 | 404 | 4.4727ms | 127.0.0.1 | POST "/api/show" time=2025-03-27T10:26:53.952+01:00 level=INFO source=images.go:669 msg="request failed: Get \"https://registry.ollama.ai/v2/library/llama_vision_fp6_custom/manifests/latest\": dial tcp: lookup registry.ollama.ai: no such host" [GIN] 2025/03/27 - 10:26:53 | 200 | 12.0130116s | 127.0.0.1 | POST "/api/pull" [GIN] 2025/03/27 - 10:32:12 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:32:12 | 200 | 74.9073ms | 127.0.0.1 | POST "/api/show" time=2025-03-27T10:32:12.953+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" time=2025-03-27T10:32:13.040+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128 time=2025-03-27T10:32:13.040+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128 time=2025-03-27T10:32:13.043+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128 time=2025-03-27T10:32:13.043+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128 time=2025-03-27T10:32:13.051+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128 time=2025-03-27T10:32:13.051+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128 time=2025-03-27T10:32:13.053+01:00 level=INFO source=sched.go:731 msg="new model will fit in available VRAM, loading" model=C:\Users\devops01\.ollama\models\blobs\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 library=cuda parallel=1 required="76.8 GiB" time=2025-03-27T10:32:13.084+01:00 level=INFO source=server.go:97 msg="system memory" total="256.0 GiB" free="250.0 GiB" free_swap="284.7 GiB" time=2025-03-27T10:32:13.087+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128 time=2025-03-27T10:32:13.087+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128 time=2025-03-27T10:32:13.087+01:00 level=INFO source=server.go:130 msg=offload library=cuda layers.requested=-1 layers.model=101 layers.offload=101 layers.split=51,50 memory.available="[42.9 GiB 42.8 GiB]" memory.gpu_overhead="0 B" memory.required.full="76.8 GiB" memory.required.partial="76.8 GiB" memory.required.kv="1.6 GiB" memory.required.allocations="[41.1 GiB 35.7 GiB]" memory.weights.total="67.0 GiB" memory.weights.repeating="66.2 GiB" memory.weights.nonrepeating="822.0 MiB" memory.graph.full="1.1 GiB" memory.graph.partial="1.1 GiB" projector.weights="1.9 GiB" projector.graph="2.8 GiB" time=2025-03-27T10:32:13.095+01:00 level=INFO source=server.go:380 msg="starting llama server" cmd="C:\\Users\\devops01\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --model C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 --ctx-size 2048 --batch-size 512 --n-gpu-layers 101 --mmproj C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-6b6c374d159e097509b33e9fda648c178c903959fc0c7dbfae487cc8d958093e --threads 8 --no-mmap --parallel 1 --tensor-split 51,50 --port 62648" time=2025-03-27T10:32:13.102+01:00 level=INFO source=sched.go:450 msg="loaded runners" count=1 time=2025-03-27T10:32:13.102+01:00 level=INFO source=server.go:557 msg="waiting for llama runner to start responding" time=2025-03-27T10:32:13.103+01:00 level=INFO source=server.go:591 msg="waiting for server to become available" status="llm server error" time=2025-03-27T10:32:13.139+01:00 level=INFO source=runner.go:932 msg="starting go runner" ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 2 CUDA devices: Device 0: NVIDIA A40-48Q, compute capability 8.6, VMM: no Device 1: NVIDIA A40-48Q, compute capability 8.6, VMM: no load_backend: loaded CUDA backend from C:\Users\devops01\AppData\Local\Programs\Ollama\lib\ollama\cuda_v11\ggml-cuda.dll load_backend: loaded CPU backend from C:\Users\devops01\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-icelake.dll time=2025-03-27T10:32:14.460+01:00 level=INFO source=runner.go:935 msg=system info="CPU : LLAMAFILE = 1 | CPU : LLAMAFILE = 1 | CUDA : USE_GRAPHS = 1 | PEER_MAX_BATCH_SIZE = 128 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | cgo(clang)" threads=8 time=2025-03-27T10:32:14.462+01:00 level=INFO source=runner.go:993 msg="Server listening on 127.0.0.1:62648" time=2025-03-27T10:32:14.611+01:00 level=INFO source=server.go:591 msg="waiting for server to become available" status="llm server loading model" llama_load_model_from_file: using device CUDA0 (NVIDIA A40-48Q) - 43830 MiB free llama_load_model_from_file: using device CUDA1 (NVIDIA A40-48Q) - 43830 MiB free llama_model_loader: loaded meta data with 27 key-value pairs and 984 tensors from C:\Users\devops01\.ollama\models\blobs\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 (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 = mllama llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = Model llama_model_loader: - kv 3: general.size_label str = 88B llama_model_loader: - kv 4: mllama.block_count u32 = 100 llama_model_loader: - kv 5: mllama.context_length u32 = 131072 llama_model_loader: - kv 6: mllama.embedding_length u32 = 8192 llama_model_loader: - kv 7: mllama.feed_forward_length u32 = 28672 llama_model_loader: - kv 8: mllama.attention.head_count u32 = 64 llama_model_loader: - kv 9: mllama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 10: mllama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 11: mllama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 12: general.file_type u32 = 18 llama_model_loader: - kv 13: mllama.vocab_size u32 = 128256 llama_model_loader: - kv 14: mllama.rope.dimension_count u32 = 128 llama_model_loader: - kv 15: mllama.attention.cross_attention_layers arr[i32,20] = [3, 8, 13, 18, 23, 28, 33, 38, 43, 48... llama_model_loader: - kv 16: tokenizer.ggml.add_bos_token bool = true llama_model_loader: - kv 17: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 18: tokenizer.ggml.pre str = llama-bpe llama_model_loader: - kv 19: tokenizer.ggml.tokens arr[str,128257] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 20: tokenizer.ggml.token_type arr[i32,128257] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 21: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 22: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 23: tokenizer.ggml.eos_token_id u32 = 128009 llama_model_loader: - kv 24: tokenizer.ggml.padding_token_id u32 = 128004 llama_model_loader: - kv 25: tokenizer.chat_template str = {{- bos_token }}\n{%- if custom_tools ... llama_model_loader: - kv 26: general.quantization_version u32 = 2 llama_model_loader: - type f32: 282 tensors llama_model_loader: - type q6_K: 702 tensors llm_load_vocab: special tokens cache size = 257 llm_load_vocab: token to piece cache size = 0.7999 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = mllama llm_load_print_meta: vocab type = BPE llm_load_print_meta: n_vocab = 128256 llm_load_print_meta: n_merges = 280147 llm_load_print_meta: vocab_only = 0 llm_load_print_meta: n_ctx_train = 131072 llm_load_print_meta: n_embd = 8192 llm_load_print_meta: n_layer = 100 llm_load_print_meta: n_head = 64 llm_load_print_meta: n_head_kv = 8 llm_load_print_meta: n_rot = 128 llm_load_print_meta: n_swa = 0 llm_load_print_meta: n_embd_head_k = 128 llm_load_print_meta: n_embd_head_v = 128 llm_load_print_meta: n_gqa = 8 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 = 28672 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 = 500000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_ctx_orig_yarn = 131072 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: ssm_dt_b_c_rms = 0 llm_load_print_meta: model type = ?B llm_load_print_meta: model ftype = Q6_K llm_load_print_meta: model params = 87.67 B llm_load_print_meta: model size = 66.98 GiB (6.56 BPW) llm_load_print_meta: general.name = Model llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128009 '<|eot_id|>' llm_load_print_meta: EOT token = 128009 '<|eot_id|>' llm_load_print_meta: EOM token = 128008 '<|eom_id|>' llm_load_print_meta: PAD token = 128004 '<|finetune_right_pad_id|>' llm_load_print_meta: LF token = 128 'Ä' llm_load_print_meta: EOG token = 128008 '<|eom_id|>' llm_load_print_meta: EOG token = 128009 '<|eot_id|>' llm_load_print_meta: max token length = 256 llama_model_load: vocab mismatch 128256 !- 128257 ... llm_load_tensors: offloading 100 repeating layers to GPU llm_load_tensors: offloading output layer to GPU llm_load_tensors: offloaded 101/101 layers to GPU llm_load_tensors: CPU model buffer size = 822.00 MiB llm_load_tensors: CUDA0 model buffer size = 34141.32 MiB llm_load_tensors: CUDA1 model buffer size = 33624.43 MiB llama_new_context_with_model: n_seq_max = 1 llama_new_context_with_model: n_ctx = 2048 llama_new_context_with_model: n_ctx_per_seq = 2048 llama_new_context_with_model: n_batch = 512 llama_new_context_with_model: n_ubatch = 512 llama_new_context_with_model: flash_attn = 0 llama_new_context_with_model: freq_base = 500000.0 llama_new_context_with_model: freq_scale = 1 llama_new_context_with_model: n_ctx_per_seq (2048) < n_ctx_train (131072) -- the full capacity of the model will not be utilized llama_kv_cache_init: kv_size = 2048, offload = 1, type_k = 'f16', type_v = 'f16', n_layer = 100, can_shift = 1 llama_kv_cache_init: CUDA0 KV buffer size = 828.31 MiB llama_kv_cache_init: CUDA1 KV buffer size = 812.31 MiB llama_new_context_with_model: KV self size = 1640.62 MiB, K (f16): 820.31 MiB, V (f16): 820.31 MiB llama_new_context_with_model: CUDA_Host output buffer size = 0.52 MiB llama_new_context_with_model: pipeline parallelism enabled (n_copies=4) llama_new_context_with_model: CUDA0 compute buffer size = 400.01 MiB llama_new_context_with_model: CUDA1 compute buffer size = 400.02 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 32.02 MiB llama_new_context_with_model: graph nodes = 2566 llama_new_context_with_model: graph splits = 3 mllama_model_load: model name: Llama-3.2-90B-Vision-Instruct mllama_model_load: description: vision encoder for Mllama mllama_model_load: GGUF version: 3 mllama_model_load: alignment: 32 mllama_model_load: n_tensors: 512 mllama_model_load: n_kv: 17 mllama_model_load: ftype: f16 mllama_model_load: mllama_model_load: mllama_model_load: using CUDA0 backend mllama_model_load: compute allocated memory: 2853.34 MB time=2025-03-27T10:32:49.460+01:00 level=INFO source=server.go:596 msg="llama runner started in 36.36 seconds" [GIN] 2025/03/27 - 10:32:49 | 200 | 36.5347842s | 127.0.0.1 | POST "/api/generate" [GIN] 2025/03/27 - 10:33:07 | 200 | 85.6µs | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:33:07 | 200 | 635.5µs | 127.0.0.1 | GET "/api/ps" [GIN] 2025/03/27 - 10:33:34 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:33:34 | 200 | 33.761ms | 127.0.0.1 | POST "/api/show" time=2025-03-27T10:33:34.857+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" [GIN] 2025/03/27 - 10:33:34 | 200 | 37.4451ms | 127.0.0.1 | POST "/api/generate" [GIN] 2025/03/27 - 10:34:00 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:34:00 | 200 | 32.5477ms | 127.0.0.1 | POST "/api/show" time=2025-03-27T10:34:00.626+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" [GIN] 2025/03/27 - 10:34:00 | 200 | 30.9911ms | 127.0.0.1 | POST "/api/generate" [GIN] 2025/03/27 - 10:34:16 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:34:16 | 200 | 33.7433ms | 127.0.0.1 | POST "/api/show" time=2025-03-27T10:34:16.250+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" [GIN] 2025/03/27 - 10:34:16 | 200 | 33.6178ms | 127.0.0.1 | POST "/api/generate" time=2025-03-27T10:34:38.089+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" time=2025-03-27T10:34:39.562+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" time=2025-03-27T10:34:40.185+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" time=2025-03-27T10:34:40.590+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" [GIN] 2025/03/27 - 10:35:07 | 200 | 29.877435s | 127.0.0.1 | POST "/api/chat" [GIN] 2025/03/27 - 10:35:48 | 200 | 1m9s | 127.0.0.1 | POST "/api/chat" [GIN] 2025/03/27 - 10:38:28 | 200 | 3m47s | 127.0.0.1 | POST "/api/chat" [GIN] 2025/03/27 - 10:39:05 | 200 | 4m25s | 127.0.0.1 | POST "/api/chat" [GIN] 2025/03/27 - 10:39:18 | 200 | 49.6µs | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:39:18 | 200 | 4.2162ms | 127.0.0.1 | POST "/api/generate" [GIN] 2025/03/27 - 10:39:23 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:39:23 | 200 | 30.8407ms | 127.0.0.1 | POST "/api/show" time=2025-03-27T10:39:23.738+01:00 level=WARN source=sched.go:138 msg="mllama doesn't support parallel requests yet" time=2025-03-27T10:39:23.827+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128 time=2025-03-27T10:39:23.827+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128 time=2025-03-27T10:39:23.830+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128 time=2025-03-27T10:39:23.830+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128 time=2025-03-27T10:39:23.836+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128 time=2025-03-27T10:39:23.836+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128 time=2025-03-27T10:39:23.837+01:00 level=INFO source=sched.go:731 msg="new model will fit in available VRAM, loading" model=C:\Users\devops01\.ollama\models\blobs\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 library=cuda parallel=1 required="76.8 GiB" time=2025-03-27T10:39:23.866+01:00 level=INFO source=server.go:97 msg="system memory" total="256.0 GiB" free="249.8 GiB" free_swap="284.6 GiB" time=2025-03-27T10:39:23.869+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.key_length default=128 time=2025-03-27T10:39:23.869+01:00 level=WARN source=ggml.go:132 msg="key not found" key=mllama.attention.value_length default=128 time=2025-03-27T10:39:23.870+01:00 level=INFO source=server.go:130 msg=offload library=cuda layers.requested=-1 layers.model=101 layers.offload=101 layers.split=51,50 memory.available="[42.9 GiB 42.9 GiB]" memory.gpu_overhead="0 B" memory.required.full="76.8 GiB" memory.required.partial="76.8 GiB" memory.required.kv="1.6 GiB" memory.required.allocations="[41.1 GiB 35.7 GiB]" memory.weights.total="67.0 GiB" memory.weights.repeating="66.2 GiB" memory.weights.nonrepeating="822.0 MiB" memory.graph.full="1.1 GiB" memory.graph.partial="1.1 GiB" projector.weights="1.9 GiB" projector.graph="2.8 GiB" time=2025-03-27T10:39:23.876+01:00 level=INFO source=server.go:380 msg="starting llama server" cmd="C:\\Users\\devops01\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --model C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 --ctx-size 2048 --batch-size 512 --n-gpu-layers 101 --mmproj C:\\Users\\devops01\\.ollama\\models\\blobs\\sha256-6b6c374d159e097509b33e9fda648c178c903959fc0c7dbfae487cc8d958093e --threads 8 --no-mmap --parallel 1 --tensor-split 51,50 --port 62665" time=2025-03-27T10:39:23.883+01:00 level=INFO source=sched.go:450 msg="loaded runners" count=1 time=2025-03-27T10:39:23.883+01:00 level=INFO source=server.go:557 msg="waiting for llama runner to start responding" time=2025-03-27T10:39:23.884+01:00 level=INFO source=server.go:591 msg="waiting for server to become available" status="llm server error" time=2025-03-27T10:39:23.921+01:00 level=INFO source=runner.go:932 msg="starting go runner" ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 2 CUDA devices: Device 0: NVIDIA A40-48Q, compute capability 8.6, VMM: no Device 1: NVIDIA A40-48Q, compute capability 8.6, VMM: no load_backend: loaded CUDA backend from C:\Users\devops01\AppData\Local\Programs\Ollama\lib\ollama\cuda_v11\ggml-cuda.dll load_backend: loaded CPU backend from C:\Users\devops01\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-icelake.dll time=2025-03-27T10:39:24.073+01:00 level=INFO source=runner.go:935 msg=system info="CPU : LLAMAFILE = 1 | CPU : LLAMAFILE = 1 | CUDA : USE_GRAPHS = 1 | PEER_MAX_BATCH_SIZE = 128 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | cgo(clang)" threads=8 time=2025-03-27T10:39:24.073+01:00 level=INFO source=runner.go:993 msg="Server listening on 127.0.0.1:62665" time=2025-03-27T10:39:24.136+01:00 level=INFO source=server.go:591 msg="waiting for server to become available" status="llm server loading model" llama_load_model_from_file: using device CUDA0 (NVIDIA A40-48Q) - 43830 MiB free llama_load_model_from_file: using device CUDA1 (NVIDIA A40-48Q) - 43830 MiB free llama_model_loader: loaded meta data with 27 key-value pairs and 984 tensors from C:\Users\devops01\.ollama\models\blobs\sha256-808f35a8f5569dc354fd7531d50b09889ef84d11cc570e83c0467268f9faf135 (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 = mllama llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = Model llama_model_loader: - kv 3: general.size_label str = 88B llama_model_loader: - kv 4: mllama.block_count u32 = 100 llama_model_loader: - kv 5: mllama.context_length u32 = 131072 llama_model_loader: - kv 6: mllama.embedding_length u32 = 8192 llama_model_loader: - kv 7: mllama.feed_forward_length u32 = 28672 llama_model_loader: - kv 8: mllama.attention.head_count u32 = 64 llama_model_loader: - kv 9: mllama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 10: mllama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 11: mllama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 12: general.file_type u32 = 18 llama_model_loader: - kv 13: mllama.vocab_size u32 = 128256 llama_model_loader: - kv 14: mllama.rope.dimension_count u32 = 128 llama_model_loader: - kv 15: mllama.attention.cross_attention_layers arr[i32,20] = [3, 8, 13, 18, 23, 28, 33, 38, 43, 48... llama_model_loader: - kv 16: tokenizer.ggml.add_bos_token bool = true llama_model_loader: - kv 17: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 18: tokenizer.ggml.pre str = llama-bpe llama_model_loader: - kv 19: tokenizer.ggml.tokens arr[str,128257] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 20: tokenizer.ggml.token_type arr[i32,128257] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 21: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 22: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 23: tokenizer.ggml.eos_token_id u32 = 128009 llama_model_loader: - kv 24: tokenizer.ggml.padding_token_id u32 = 128004 llama_model_loader: - kv 25: tokenizer.chat_template str = {{- bos_token }}\n{%- if custom_tools ... llama_model_loader: - kv 26: general.quantization_version u32 = 2 llama_model_loader: - type f32: 282 tensors llama_model_loader: - type q6_K: 702 tensors llm_load_vocab: special tokens cache size = 257 llm_load_vocab: token to piece cache size = 0.7999 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = mllama llm_load_print_meta: vocab type = BPE llm_load_print_meta: n_vocab = 128256 llm_load_print_meta: n_merges = 280147 llm_load_print_meta: vocab_only = 0 llm_load_print_meta: n_ctx_train = 131072 llm_load_print_meta: n_embd = 8192 llm_load_print_meta: n_layer = 100 llm_load_print_meta: n_head = 64 llm_load_print_meta: n_head_kv = 8 llm_load_print_meta: n_rot = 128 llm_load_print_meta: n_swa = 0 llm_load_print_meta: n_embd_head_k = 128 llm_load_print_meta: n_embd_head_v = 128 llm_load_print_meta: n_gqa = 8 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 = 28672 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 = 500000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_ctx_orig_yarn = 131072 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: ssm_dt_b_c_rms = 0 llm_load_print_meta: model type = ?B llm_load_print_meta: model ftype = Q6_K llm_load_print_meta: model params = 87.67 B llm_load_print_meta: model size = 66.98 GiB (6.56 BPW) llm_load_print_meta: general.name = Model llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128009 '<|eot_id|>' llm_load_print_meta: EOT token = 128009 '<|eot_id|>' llm_load_print_meta: EOM token = 128008 '<|eom_id|>' llm_load_print_meta: PAD token = 128004 '<|finetune_right_pad_id|>' llm_load_print_meta: LF token = 128 'Ä' llm_load_print_meta: EOG token = 128008 '<|eom_id|>' llm_load_print_meta: EOG token = 128009 '<|eot_id|>' llm_load_print_meta: max token length = 256 llama_model_load: vocab mismatch 128256 !- 128257 ... llm_load_tensors: offloading 100 repeating layers to GPU llm_load_tensors: offloading output layer to GPU llm_load_tensors: offloaded 101/101 layers to GPU llm_load_tensors: CPU model buffer size = 822.00 MiB llm_load_tensors: CUDA0 model buffer size = 34141.32 MiB llm_load_tensors: CUDA1 model buffer size = 33624.43 MiB [GIN] 2025/03/27 - 10:39:29 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/03/27 - 10:39:29 | 200 | 0s | 127.0.0.1 | GET "/api/ps" ``` ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.5.12
GiteaMirror added the bug label 2026-04-12 18:11:28 -05:00
GiteaMirror commented 2026-04-12 18:11:30 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 27, 2025):

OLLAMA_NUM_PARALLEL creates multiple instances of the KV cache (the working area for inference) but has only one copy of the model weights.

However, in your log:

OLLAMA_NUM_PARALLEL:0

it is unset. Have you restarted the server?

<!-- gh-comment-id:2757502511 --> @rick-github commented on GitHub (Mar 27, 2025): `OLLAMA_NUM_PARALLEL` creates multiple instances of the KV cache (the working area for inference) but has only one copy of the model weights. However, in your log: ``` OLLAMA_NUM_PARALLEL:0 ``` it is unset. Have you restarted the server?
GiteaMirror commented 2026-04-12 18:11:30 -05:00
Author
Owner
Copy Link

@forReason commented on GitHub (Mar 27, 2025):

yes. But it seems parallel not supported for vision models, is that correct?

<!-- gh-comment-id:2757608106 --> @forReason commented on GitHub (Mar 27, 2025): yes. But it seems parallel not supported for vision models, is that correct?
GiteaMirror commented 2026-04-12 18:11:31 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 27, 2025):

https://github.com/ollama/ollama/issues/9564#issuecomment-2707382182

<!-- gh-comment-id:2757620834 --> @rick-github commented on GitHub (Mar 27, 2025): https://github.com/ollama/ollama/issues/9564#issuecomment-2707382182
GiteaMirror commented 2026-04-12 18:11:31 -05:00
Author
Owner
Copy Link

@sieveLau commented on GitHub (Mar 27, 2025):

System variables need a reboot to take effect, but I'm not sure if a logout and login can work. If you change the variable for the account, a new shell will take effect.

<!-- gh-comment-id:2757742494 --> @sieveLau commented on GitHub (Mar 27, 2025): System variables need a reboot to take effect, but I'm not sure if a logout and login can work. If you change the variable for the account, a new shell will take effect.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#6563
Delete Branch "%!s()"

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