[GH-ISSUE #8349] can't use gpu after update #67410

New Issue

GiteaMirror · 2026-05-04T10:16:02-05:00

GiteaMirror commented

2026-05-04 10:16:02 -05:00

Originally created by @Heart-eartH on GitHub (Jan 8, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8349

Originally assigned to: @dhiltgen on GitHub.

What is the issue?

Previously, it run on the GPU by default, but after the update, it run on the CPU. What can I do to change it

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.5.4

Originally created by @Heart-eartH on GitHub (Jan 8, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8349 Originally assigned to: @dhiltgen on GitHub. ### What is the issue? Previously, it run on the GPU by default, but after the update, it run on the CPU. What can I do to change it ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.5.4

GiteaMirror added the bug nvidia gpu needs more info labels 2026-05-04 10:16:04 -05:00

GiteaMirror closed this issue

2026-05-04 10:16:08 -05:00

GiteaMirror commented

2026-05-04 10:16:10 -05:00

@rick-github commented on GitHub (Jan 8, 2025):

Server logs will help in debugging.

@rick-github commented on GitHub (Jan 8, 2025): [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) will help in debugging.

GiteaMirror commented

2026-05-04 10:16:12 -05:00

@Heart-eartH commented on GitHub (Jan 9, 2025):

Today I deleted Olama and reinstalled it, and now I can use GPU. It seems that this issue may occur when using update, but it will not occur when installing directly
upgrade.log

@Heart-eartH commented on GitHub (Jan 9, 2025): Today I deleted Olama and reinstalled it, and now I can use GPU. It seems that this issue may occur when using update, but it will not occur when installing directly [upgrade.log](https://github.com/user-attachments/files/18362405/upgrade.log)

GiteaMirror commented

2026-05-04 10:16:16 -05:00

@peperunas commented on GitHub (Jan 14, 2025):

I am getting a somewhat similar issue after upgrading to 0.5.4. I am on NixOS.

Jan 14 16:41:34 architect ollama[16241]: 2025/01/14 16:41:34 routes.go:1259: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:true OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE:q8_0 OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/var/lib/ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:2 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: http_proxy: https_proxy: no_proxy:]"
Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.305Z level=INFO source=images.go:757 msg="total blobs: 47"
Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.305Z level=INFO source=images.go:764 msg="total unused blobs removed: 0"
Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.306Z level=INFO source=routes.go:1310 msg="Listening on 127.0.0.1:11434 (version 0.5.4)"
Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.306Z level=INFO source=routes.go:1339 msg="Dynamic LLM libraries" runners=[cpu]
Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.306Z level=INFO source=gpu.go:226 msg="looking for compatible GPUs"
Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.532Z level=INFO source=types.go:131 msg="inference compute" id=GPU-9fd65da0-a21e-786b-bfca-f857fa6d5662 library=cuda variant=v12 compute=8.9 driver=12.7 name="NVIDIA GeForce RTX 4090" total="23.5 GiB" available="23.1 GiB"
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.132Z level=INFO source=server.go:104 msg="system memory" total="62.7 GiB" free="57.8 GiB" free_swap="64.0 GiB"
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.277Z level=INFO source=memory.go:356 msg="offload to cuda" layers.requested=-1 layers.model=65 layers.offload=57 layers.split="" memory.available="[23.1 GiB]" memory.gpu_overhead="0 B" memory.required.full="25.9 GiB" memory.required.partial="22.9 GiB" memory.required.kv="3.9 GiB" memory.required.allocations="[22.9 GiB]" memory.weights.total="21.4 GiB" memory.weights.repeating="20.8 GiB" memory.weights.nonrepeating="609.1 MiB" memory.graph.full="2.5 GiB" memory.graph.partial="3.1 GiB"
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.277Z level=INFO source=server.go:223 msg="enabling flash attention"
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.278Z level=INFO source=server.go:376 msg="starting llama server" cmd="/nix/store/y7jkic6a3ybzh2n2cxhwas2f1hhmcvsb-ollama-0.5.4/bin/.ollama-wrapped runner --model /var/lib/ollama/models/blobs/sha256-ac3d1ba8aa77755dab3806d9024e9c385ea0d5b412d6bdf9157f8a4a7e9fc0d9 --ctx-size 32000 --batch-size 512 --n-gpu-layers 57 --threads 12 --flash-attn --kv-cache-type q8_0 --parallel 2 --port 39273"
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.278Z level=INFO source=sched.go:449 msg="loaded runners" count=1
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.278Z level=INFO source=server.go:555 msg="waiting for llama runner to start responding"
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.278Z level=INFO source=server.go:589 msg="waiting for server to become available" status="llm server error"
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.289Z level=INFO source=runner.go:945 msg="starting go runner"
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.290Z level=INFO source=runner.go:946 msg=system info="CPU : LLAMAFILE = 1 | AARCH64_REPACK = 1 | CPU : LLAMAFILE = 1 | AARCH64_REPACK = 1 | cgo(gcc)" threads=12
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.290Z level=INFO source=runner.go:1004 msg="Server listening on 127.0.0.1:39273"
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: loaded meta data with 34 key-value pairs and 771 tensors from /var/lib/ollama/models/blobs/sha256-ac3d1ba8aa77755dab3806d9024e9c385ea0d5b412d6bdf9157f8a4a7e9fc0d9 (version GGUF V3 (latest))
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   0:                       general.architecture str              = qwen2
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   1:                               general.type str              = model
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   2:                               general.name str              = Qwen2.5 Coder 32B Instruct
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   3:                           general.finetune str              = Instruct
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   4:                           general.basename str              = Qwen2.5-Coder
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   5:                         general.size_label str              = 32B
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   6:                            general.license str              = apache-2.0
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   7:                       general.license.link str              = https://huggingface.co/Qwen/Qwen2.5-C...
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   8:                   general.base_model.count u32              = 1
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv   9:                  general.base_model.0.name str              = Qwen2.5 Coder 32B
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  10:          general.base_model.0.organization str              = Qwen
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  11:              general.base_model.0.repo_url str              = https://huggingface.co/Qwen/Qwen2.5-C...
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  12:                               general.tags arr[str,6]       = ["code", "codeqwen", "chat", "qwen", ...
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  13:                          general.languages arr[str,1]       = ["en"]
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  14:                          qwen2.block_count u32              = 64
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  15:                       qwen2.context_length u32              = 32768
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  16:                     qwen2.embedding_length u32              = 5120
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  17:                  qwen2.feed_forward_length u32              = 27648
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  18:                 qwen2.attention.head_count u32              = 40
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  19:              qwen2.attention.head_count_kv u32              = 8
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  20:                       qwen2.rope.freq_base f32              = 1000000.000000
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  21:     qwen2.attention.layer_norm_rms_epsilon f32              = 0.000001
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  22:                          general.file_type u32              = 15
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  23:                       tokenizer.ggml.model str              = gpt2
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  24:                         tokenizer.ggml.pre str              = qwen2
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  25:                      tokenizer.ggml.tokens arr[str,152064]  = ["!", "\"", "#", "$", "%", "&", "'", ...
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  26:                  tokenizer.ggml.token_type arr[i32,152064]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  27:                      tokenizer.ggml.merges arr[str,151387]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  28:                tokenizer.ggml.eos_token_id u32              = 151645
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  29:            tokenizer.ggml.padding_token_id u32              = 151643
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  30:                tokenizer.ggml.bos_token_id u32              = 151643
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  31:               tokenizer.ggml.add_bos_token bool             = false
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  32:                    tokenizer.chat_template str              = {%- if tools %}\n    {{- '<|im_start|>...
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv  33:               general.quantization_version u32              = 2
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - type  f32:  321 tensors
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - type q4_K:  385 tensors
Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - type q6_K:   65 tensors
Jan 14 16:42:20 architect ollama[16241]: llm_load_vocab: special tokens cache size = 22
Jan 14 16:42:20 architect ollama[16241]: llm_load_vocab: token to piece cache size = 0.9310 MB
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: format           = GGUF V3 (latest)
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: arch             = qwen2
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: vocab type       = BPE
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_vocab          = 152064
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_merges         = 151387
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: vocab_only       = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_ctx_train      = 32768
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd           = 5120
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_layer          = 64
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_head           = 40
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_head_kv        = 8
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_rot            = 128
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_swa            = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd_head_k    = 128
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd_head_v    = 128
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_gqa            = 5
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd_k_gqa     = 1024
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd_v_gqa     = 1024
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_norm_eps       = 0.0e+00
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_norm_rms_eps   = 1.0e-06
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_clamp_kqv      = 0.0e+00
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_max_alibi_bias = 0.0e+00
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_logit_scale    = 0.0e+00
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_ff             = 27648
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_expert         = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_expert_used    = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: causal attn      = 1
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: pooling type     = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: rope type        = 2
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: rope scaling     = linear
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: freq_base_train  = 1000000.0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: freq_scale_train = 1
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_ctx_orig_yarn  = 32768
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: rope_finetuned   = unknown
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_d_conv       = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_d_inner      = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_d_state      = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_dt_rank      = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_dt_b_c_rms   = 0
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: model type       = 32B
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: model ftype      = Q4_K - Medium
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: model params     = 32.76 B
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: model size       = 18.48 GiB (4.85 BPW)
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: general.name     = Qwen2.5 Coder 32B Instruct
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: BOS token        = 151643 '<|endoftext|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOS token        = 151645 '<|im_end|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOT token        = 151645 '<|im_end|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: PAD token        = 151643 '<|endoftext|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: LF token         = 148848 'ÄĬ'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM PRE token    = 151659 '<|fim_prefix|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM SUF token    = 151661 '<|fim_suffix|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM MID token    = 151660 '<|fim_middle|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM PAD token    = 151662 '<|fim_pad|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM REP token    = 151663 '<|repo_name|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM SEP token    = 151664 '<|file_sep|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token        = 151643 '<|endoftext|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token        = 151645 '<|im_end|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token        = 151662 '<|fim_pad|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token        = 151663 '<|repo_name|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token        = 151664 '<|file_sep|>'
Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: max token length = 256
Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.529Z level=INFO source=server.go:589 msg="waiting for server to become available" status="llm server loading model"
Jan 14 16:42:21 architect ollama[16241]: llm_load_tensors:   CPU_Mapped model buffer size = 18926.01 MiB
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_seq_max     = 2
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_ctx         = 32000
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_ctx_per_seq = 16000
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_batch       = 1024
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_ubatch      = 512
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: flash_attn    = 1
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: freq_base     = 1000000.0
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: freq_scale    = 1
Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_ctx_per_seq (16000) < n_ctx_train (32768) -- the full capacity of the model will not be utilized
Jan 14 16:42:23 architect ollama[16241]: llama_kv_cache_init:        CPU KV buffer size =  4250.00 MiB
Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model: KV self size  = 4250.00 MiB, K (q8_0): 2125.00 MiB, V (q8_0): 2125.00 MiB
Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model:        CPU  output buffer size =     1.20 MiB
Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model:        CPU compute buffer size =   307.00 MiB
Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model: graph nodes  = 1991
Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model: graph splits = 1
Jan 14 16:42:23 architect ollama[16241]: time=2025-01-14T16:42:23.288Z level=INFO source=server.go:594 msg="llama runner started in 3.01 seconds"
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: loaded meta data with 34 key-value pairs and 771 tensors from /var/lib/ollama/models/blobs/sha256-ac3d1ba8aa77755dab3806d9024e9c385ea0d5b412d6bdf9157f8a4a7e9fc0d9 (version GGUF V3 (latest))
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   0:                       general.architecture str              = qwen2
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   1:                               general.type str              = model
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   2:                               general.name str              = Qwen2.5 Coder 32B Instruct
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   3:                           general.finetune str              = Instruct
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   4:                           general.basename str              = Qwen2.5-Coder
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   5:                         general.size_label str              = 32B
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   6:                            general.license str              = apache-2.0
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   7:                       general.license.link str              = https://huggingface.co/Qwen/Qwen2.5-C...
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   8:                   general.base_model.count u32              = 1
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv   9:                  general.base_model.0.name str              = Qwen2.5 Coder 32B
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  10:          general.base_model.0.organization str              = Qwen
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  11:              general.base_model.0.repo_url str              = https://huggingface.co/Qwen/Qwen2.5-C...
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  12:                               general.tags arr[str,6]       = ["code", "codeqwen", "chat", "qwen", ...
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  13:                          general.languages arr[str,1]       = ["en"]
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  14:                          qwen2.block_count u32              = 64
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  15:                       qwen2.context_length u32              = 32768
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  16:                     qwen2.embedding_length u32              = 5120
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  17:                  qwen2.feed_forward_length u32              = 27648
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  18:                 qwen2.attention.head_count u32              = 40
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  19:              qwen2.attention.head_count_kv u32              = 8
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  20:                       qwen2.rope.freq_base f32              = 1000000.000000
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  21:     qwen2.attention.layer_norm_rms_epsilon f32              = 0.000001
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  22:                          general.file_type u32              = 15
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  23:                       tokenizer.ggml.model str              = gpt2
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  24:                         tokenizer.ggml.pre str              = qwen2
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  25:                      tokenizer.ggml.tokens arr[str,152064]  = ["!", "\"", "#", "$", "%", "&", "'", ...
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  26:                  tokenizer.ggml.token_type arr[i32,152064]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  27:                      tokenizer.ggml.merges arr[str,151387]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  28:                tokenizer.ggml.eos_token_id u32              = 151645
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  29:            tokenizer.ggml.padding_token_id u32              = 151643
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  30:                tokenizer.ggml.bos_token_id u32              = 151643
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  31:               tokenizer.ggml.add_bos_token bool             = false
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  32:                    tokenizer.chat_template str              = {%- if tools %}\n    {{- '<|im_start|>...
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv  33:               general.quantization_version u32              = 2
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - type  f32:  321 tensors
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - type q4_K:  385 tensors
Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - type q6_K:   65 tensors
Jan 14 16:42:23 architect ollama[16241]: llm_load_vocab: special tokens cache size = 22
Jan 14 16:42:23 architect ollama[16241]: llm_load_vocab: token to piece cache size = 0.9310 MB
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: format           = GGUF V3 (latest)
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: arch             = qwen2
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: vocab type       = BPE
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: n_vocab          = 152064
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: n_merges         = 151387
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: vocab_only       = 1
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: model type       = ?B
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: model ftype      = all F32
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: model params     = 32.76 B
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: model size       = 18.48 GiB (4.85 BPW)
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: general.name     = Qwen2.5 Coder 32B Instruct
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: BOS token        = 151643 '<|endoftext|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOS token        = 151645 '<|im_end|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOT token        = 151645 '<|im_end|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: PAD token        = 151643 '<|endoftext|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: LF token         = 148848 'ÄĬ'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM PRE token    = 151659 '<|fim_prefix|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM SUF token    = 151661 '<|fim_suffix|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM MID token    = 151660 '<|fim_middle|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM PAD token    = 151662 '<|fim_pad|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM REP token    = 151663 '<|repo_name|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM SEP token    = 151664 '<|file_sep|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token        = 151643 '<|endoftext|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token        = 151645 '<|im_end|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token        = 151662 '<|fim_pad|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token        = 151663 '<|repo_name|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token        = 151664 '<|file_sep|>'
Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: max token length = 256
Jan 14 16:42:23 architect ollama[16241]: llama_model_load: vocab only - skipping tensors

As you can see, it detects the GPU, but then it falls back to CPU-only. I didn't have this behavior before.

Specs:

OS: NixOS 24.11.20250114.62b598a (Vicuna) x86_64
Kernel: 6.6.71
CPU: AMD Ryzen 9 5900X (24) @ 3.700GHz
GPU: NVIDIA GeForce RTX 4090
Memory: 64202MiB

@peperunas commented on GitHub (Jan 14, 2025): I am getting a somewhat similar issue after upgrading to 0.5.4. I am on NixOS. ``` Jan 14 16:41:34 architect ollama[16241]: 2025/01/14 16:41:34 routes.go:1259: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:true OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE:q8_0 OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/var/lib/ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:2 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: http_proxy: https_proxy: no_proxy:]" Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.305Z level=INFO source=images.go:757 msg="total blobs: 47" Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.305Z level=INFO source=images.go:764 msg="total unused blobs removed: 0" Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.306Z level=INFO source=routes.go:1310 msg="Listening on 127.0.0.1:11434 (version 0.5.4)" Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.306Z level=INFO source=routes.go:1339 msg="Dynamic LLM libraries" runners=[cpu] Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.306Z level=INFO source=gpu.go:226 msg="looking for compatible GPUs" Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.532Z level=INFO source=types.go:131 msg="inference compute" id=GPU-9fd65da0-a21e-786b-bfca-f857fa6d5662 library=cuda variant=v12 compute=8.9 driver=12.7 name="NVIDIA GeForce RTX 4090" total="23.5 GiB" available="23.1 GiB" Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.132Z level=INFO source=server.go:104 msg="system memory" total="62.7 GiB" free="57.8 GiB" free_swap="64.0 GiB" Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.277Z level=INFO source=memory.go:356 msg="offload to cuda" layers.requested=-1 layers.model=65 layers.offload=57 layers.split="" memory.available="[23.1 GiB]" memory.gpu_overhead="0 B" memory.required.full="25.9 GiB" memory.required.partial="22.9 GiB" memory.required.kv="3.9 GiB" memory.required.allocations="[22.9 GiB]" memory.weights.total="21.4 GiB" memory.weights.repeating="20.8 GiB" memory.weights.nonrepeating="609.1 MiB" memory.graph.full="2.5 GiB" memory.graph.partial="3.1 GiB" Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.277Z level=INFO source=server.go:223 msg="enabling flash attention" Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.278Z level=INFO source=server.go:376 msg="starting llama server" cmd="/nix/store/y7jkic6a3ybzh2n2cxhwas2f1hhmcvsb-ollama-0.5.4/bin/.ollama-wrapped runner --model /var/lib/ollama/models/blobs/sha256-ac3d1ba8aa77755dab3806d9024e9c385ea0d5b412d6bdf9157f8a4a7e9fc0d9 --ctx-size 32000 --batch-size 512 --n-gpu-layers 57 --threads 12 --flash-attn --kv-cache-type q8_0 --parallel 2 --port 39273" Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.278Z level=INFO source=sched.go:449 msg="loaded runners" count=1 Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.278Z level=INFO source=server.go:555 msg="waiting for llama runner to start responding" Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.278Z level=INFO source=server.go:589 msg="waiting for server to become available" status="llm server error" Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.289Z level=INFO source=runner.go:945 msg="starting go runner" Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.290Z level=INFO source=runner.go:946 msg=system info="CPU : LLAMAFILE = 1 | AARCH64_REPACK = 1 | CPU : LLAMAFILE = 1 | AARCH64_REPACK = 1 | cgo(gcc)" threads=12 Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.290Z level=INFO source=runner.go:1004 msg="Server listening on 127.0.0.1:39273" Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: loaded meta data with 34 key-value pairs and 771 tensors from /var/lib/ollama/models/blobs/sha256-ac3d1ba8aa77755dab3806d9024e9c385ea0d5b412d6bdf9157f8a4a7e9fc0d9 (version GGUF V3 (latest)) Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 0: general.architecture str = qwen2 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 1: general.type str = model Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 2: general.name str = Qwen2.5 Coder 32B Instruct Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 3: general.finetune str = Instruct Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 4: general.basename str = Qwen2.5-Coder Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 5: general.size_label str = 32B Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 6: general.license str = apache-2.0 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 7: general.license.link str = https://huggingface.co/Qwen/Qwen2.5-C... Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 8: general.base_model.count u32 = 1 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 9: general.base_model.0.name str = Qwen2.5 Coder 32B Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 10: general.base_model.0.organization str = Qwen Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/Qwen/Qwen2.5-C... Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 12: general.tags arr[str,6] = ["code", "codeqwen", "chat", "qwen", ... Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 13: general.languages arr[str,1] = ["en"] Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 14: qwen2.block_count u32 = 64 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 15: qwen2.context_length u32 = 32768 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 16: qwen2.embedding_length u32 = 5120 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 17: qwen2.feed_forward_length u32 = 27648 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 18: qwen2.attention.head_count u32 = 40 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 19: qwen2.attention.head_count_kv u32 = 8 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 20: qwen2.rope.freq_base f32 = 1000000.000000 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 21: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 22: general.file_type u32 = 15 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 23: tokenizer.ggml.model str = gpt2 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 24: tokenizer.ggml.pre str = qwen2 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 25: tokenizer.ggml.tokens arr[str,152064] = ["!", "\"", "#", "$", "%", "&", "'", ... Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 26: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 27: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 28: tokenizer.ggml.eos_token_id u32 = 151645 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 29: tokenizer.ggml.padding_token_id u32 = 151643 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 30: tokenizer.ggml.bos_token_id u32 = 151643 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 31: tokenizer.ggml.add_bos_token bool = false Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 32: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>... Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - kv 33: general.quantization_version u32 = 2 Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - type f32: 321 tensors Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - type q4_K: 385 tensors Jan 14 16:42:20 architect ollama[16241]: llama_model_loader: - type q6_K: 65 tensors Jan 14 16:42:20 architect ollama[16241]: llm_load_vocab: special tokens cache size = 22 Jan 14 16:42:20 architect ollama[16241]: llm_load_vocab: token to piece cache size = 0.9310 MB Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: format = GGUF V3 (latest) Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: arch = qwen2 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: vocab type = BPE Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_vocab = 152064 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_merges = 151387 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: vocab_only = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_ctx_train = 32768 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd = 5120 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_layer = 64 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_head = 40 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_head_kv = 8 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_rot = 128 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_swa = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd_head_k = 128 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd_head_v = 128 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_gqa = 5 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd_k_gqa = 1024 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_embd_v_gqa = 1024 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_norm_eps = 0.0e+00 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_norm_rms_eps = 1.0e-06 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_clamp_kqv = 0.0e+00 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_max_alibi_bias = 0.0e+00 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: f_logit_scale = 0.0e+00 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_ff = 27648 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_expert = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_expert_used = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: causal attn = 1 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: pooling type = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: rope type = 2 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: rope scaling = linear Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: freq_base_train = 1000000.0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: freq_scale_train = 1 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: n_ctx_orig_yarn = 32768 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: rope_finetuned = unknown Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_d_conv = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_d_inner = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_d_state = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_dt_rank = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: ssm_dt_b_c_rms = 0 Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: model type = 32B Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: model ftype = Q4_K - Medium Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: model params = 32.76 B Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: model size = 18.48 GiB (4.85 BPW) Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: general.name = Qwen2.5 Coder 32B Instruct Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: BOS token = 151643 '<|endoftext|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOS token = 151645 '<|im_end|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOT token = 151645 '<|im_end|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: PAD token = 151643 '<|endoftext|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: LF token = 148848 'ÄĬ' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM PRE token = 151659 '<|fim_prefix|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM SUF token = 151661 '<|fim_suffix|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM MID token = 151660 '<|fim_middle|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM PAD token = 151662 '<|fim_pad|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM REP token = 151663 '<|repo_name|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: FIM SEP token = 151664 '<|file_sep|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token = 151643 '<|endoftext|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token = 151645 '<|im_end|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token = 151662 '<|fim_pad|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token = 151663 '<|repo_name|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: EOG token = 151664 '<|file_sep|>' Jan 14 16:42:20 architect ollama[16241]: llm_load_print_meta: max token length = 256 Jan 14 16:42:20 architect ollama[16241]: time=2025-01-14T16:42:20.529Z level=INFO source=server.go:589 msg="waiting for server to become available" status="llm server loading model" Jan 14 16:42:21 architect ollama[16241]: llm_load_tensors: CPU_Mapped model buffer size = 18926.01 MiB Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_seq_max = 2 Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_ctx = 32000 Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_ctx_per_seq = 16000 Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_batch = 1024 Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_ubatch = 512 Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: flash_attn = 1 Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: freq_base = 1000000.0 Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: freq_scale = 1 Jan 14 16:42:21 architect ollama[16241]: llama_new_context_with_model: n_ctx_per_seq (16000) < n_ctx_train (32768) -- the full capacity of the model will not be utilized Jan 14 16:42:23 architect ollama[16241]: llama_kv_cache_init: CPU KV buffer size = 4250.00 MiB Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model: KV self size = 4250.00 MiB, K (q8_0): 2125.00 MiB, V (q8_0): 2125.00 MiB Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model: CPU output buffer size = 1.20 MiB Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model: CPU compute buffer size = 307.00 MiB Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model: graph nodes = 1991 Jan 14 16:42:23 architect ollama[16241]: llama_new_context_with_model: graph splits = 1 Jan 14 16:42:23 architect ollama[16241]: time=2025-01-14T16:42:23.288Z level=INFO source=server.go:594 msg="llama runner started in 3.01 seconds" Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: loaded meta data with 34 key-value pairs and 771 tensors from /var/lib/ollama/models/blobs/sha256-ac3d1ba8aa77755dab3806d9024e9c385ea0d5b412d6bdf9157f8a4a7e9fc0d9 (version GGUF V3 (latest)) Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 0: general.architecture str = qwen2 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 1: general.type str = model Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 2: general.name str = Qwen2.5 Coder 32B Instruct Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 3: general.finetune str = Instruct Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 4: general.basename str = Qwen2.5-Coder Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 5: general.size_label str = 32B Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 6: general.license str = apache-2.0 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 7: general.license.link str = https://huggingface.co/Qwen/Qwen2.5-C... Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 8: general.base_model.count u32 = 1 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 9: general.base_model.0.name str = Qwen2.5 Coder 32B Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 10: general.base_model.0.organization str = Qwen Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/Qwen/Qwen2.5-C... Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 12: general.tags arr[str,6] = ["code", "codeqwen", "chat", "qwen", ... Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 13: general.languages arr[str,1] = ["en"] Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 14: qwen2.block_count u32 = 64 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 15: qwen2.context_length u32 = 32768 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 16: qwen2.embedding_length u32 = 5120 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 17: qwen2.feed_forward_length u32 = 27648 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 18: qwen2.attention.head_count u32 = 40 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 19: qwen2.attention.head_count_kv u32 = 8 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 20: qwen2.rope.freq_base f32 = 1000000.000000 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 21: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 22: general.file_type u32 = 15 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 23: tokenizer.ggml.model str = gpt2 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 24: tokenizer.ggml.pre str = qwen2 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 25: tokenizer.ggml.tokens arr[str,152064] = ["!", "\"", "#", "$", "%", "&", "'", ... Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 26: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 27: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 28: tokenizer.ggml.eos_token_id u32 = 151645 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 29: tokenizer.ggml.padding_token_id u32 = 151643 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 30: tokenizer.ggml.bos_token_id u32 = 151643 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 31: tokenizer.ggml.add_bos_token bool = false Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 32: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>... Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - kv 33: general.quantization_version u32 = 2 Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - type f32: 321 tensors Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - type q4_K: 385 tensors Jan 14 16:42:23 architect ollama[16241]: llama_model_loader: - type q6_K: 65 tensors Jan 14 16:42:23 architect ollama[16241]: llm_load_vocab: special tokens cache size = 22 Jan 14 16:42:23 architect ollama[16241]: llm_load_vocab: token to piece cache size = 0.9310 MB Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: format = GGUF V3 (latest) Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: arch = qwen2 Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: vocab type = BPE Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: n_vocab = 152064 Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: n_merges = 151387 Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: vocab_only = 1 Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: model type = ?B Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: model ftype = all F32 Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: model params = 32.76 B Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: model size = 18.48 GiB (4.85 BPW) Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: general.name = Qwen2.5 Coder 32B Instruct Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: BOS token = 151643 '<|endoftext|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOS token = 151645 '<|im_end|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOT token = 151645 '<|im_end|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: PAD token = 151643 '<|endoftext|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: LF token = 148848 'ÄĬ' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM PRE token = 151659 '<|fim_prefix|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM SUF token = 151661 '<|fim_suffix|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM MID token = 151660 '<|fim_middle|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM PAD token = 151662 '<|fim_pad|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM REP token = 151663 '<|repo_name|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: FIM SEP token = 151664 '<|file_sep|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token = 151643 '<|endoftext|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token = 151645 '<|im_end|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token = 151662 '<|fim_pad|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token = 151663 '<|repo_name|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: EOG token = 151664 '<|file_sep|>' Jan 14 16:42:23 architect ollama[16241]: llm_load_print_meta: max token length = 256 Jan 14 16:42:23 architect ollama[16241]: llama_model_load: vocab only - skipping tensors ``` As you can see, it detects the GPU, but then it falls back to CPU-only. I didn't have this behavior before. Specs: ``` OS: NixOS 24.11.20250114.62b598a (Vicuna) x86_64 Kernel: 6.6.71 CPU: AMD Ryzen 9 5900X (24) @ 3.700GHz GPU: NVIDIA GeForce RTX 4090 Memory: 64202MiB ```

GiteaMirror commented

2026-05-04 10:16:18 -05:00

@rick-github commented on GitHub (Jan 15, 2025):

Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.306Z level=INFO source=routes.go:1339 msg="Dynamic LLM libraries" runners=[cpu]

It's using CPU because that's the only runner available. The package maintainer would need to address this.

@rick-github commented on GitHub (Jan 15, 2025): ``` Jan 14 16:41:34 architect ollama[16241]: time=2025-01-14T16:41:34.306Z level=INFO source=routes.go:1339 msg="Dynamic LLM libraries" runners=[cpu] ``` It's using CPU because that's the only runner available. The package maintainer would need to address this.

GiteaMirror commented

2026-05-04 10:16:20 -05:00

@scd31 commented on GitHub (Jan 15, 2025):

FYI this definitely seems to be an issue on the NixOS side, so this is probably safe to close. I have a PR up (in nixpkgs) which fixes this

@scd31 commented on GitHub (Jan 15, 2025): FYI this definitely seems to be an issue on the NixOS side, so this is probably safe to close. I have a PR up (in nixpkgs) which fixes this

GiteaMirror commented

2026-05-04 10:16:21 -05:00

@pbella commented on GitHub (Jan 22, 2025):

I'm also experiencing the inability to use my nVidia GPU after upgrading from 0.3.12->0.5.7 on Arch Linux. I've attached the ollama server output...

ollama-serve-output.txt

@pbella commented on GitHub (Jan 22, 2025): I'm also experiencing the inability to use my nVidia GPU after upgrading from 0.3.12->0.5.7 on Arch Linux. I've attached the ollama server output... [ollama-serve-output.txt](https://github.com/user-attachments/files/18499714/ollama-serve-output.txt)

GiteaMirror commented

2026-05-04 10:16:22 -05:00

@rick-github commented on GitHub (Jan 22, 2025):

time=2025-01-21T22:49:44.412-05:00 level=INFO source=routes.go:1267 msg="Dynamic LLM libraries" runners="[cpu_avx cpu_avx2 cpu]"

It's using CPU because that's the only runners available. The package maintainer would need to address this.

@rick-github commented on GitHub (Jan 22, 2025): ``` time=2025-01-21T22:49:44.412-05:00 level=INFO source=routes.go:1267 msg="Dynamic LLM libraries" runners="[cpu_avx cpu_avx2 cpu]" ``` It's using CPU because that's the only runners available. The package maintainer would need to address this.

Sign in to join this conversation.

Branches Tags

main

dhiltgen/ci

parth-launch-plan-gating

hoyyeva/anthropic-reference-images-path

parth-anthropic-reference-images-path

brucemacd/download-before-remove

hoyyeva/editor-config-repair

parth-mlx-decode-checkpoints

parth-launch-codex-app

hoyyeva/fix-codex-model-metadata-warning

hoyyeva/qwen

parth/hide-claude-desktop-till-release

hoyyeva/opencode-image-modality

parth-add-claude-code-autoinstall

release_v0.22.0

pdevine/manifest-list

codex/fix-codex-model-metadata-warning

pdevine/addressable-manifest

brucemacd/launch-fetch-reccomended

jmorganca/llama-compat

launch-copilot-cli

hoyyeva/opencode-thinking

release_v0.20.7

parth-auto-save-backup

parth-test

jmorganca/gemma4-audio-replacements

fix-manifest-digest-on-pull

hoyyeva/vscode-improve

brucemacd/install-server-wait

parth/update-claude-docs

brucemac/start-ap-install

pdevine/mlx-update

pdevine/qwen35_vision

drifkin/api-show-fallback

mintlify/image-generation-1773352582

hoyyeva/server-context-length-local-config

jmorganca/faster-reptition-penalties

jmorganca/convert-nemotron

parth-pi-thinking

pdevine/sampling-penalties

jmorganca/fix-create-quantization-memory

dongchen/resumable_transfer_fix

pdevine/sampling-cache-error

jessegross/mlx-usage

hoyyeva/openclaw-config

hoyyeva/app-html

pdevine/qwen3next

brucemacd/sign-sh-install

brucemacd/tui-update

brucemacd/usage-api

jmorganca/launch-empty

fix-app-dist-embed

mxyng/mlx-compile

mxyng/mlx-quant

mxyng/mlx-glm4.7

mxyng/mlx

brucemacd/simplify-model-picker

jmorganca/qwen3-concurrent

fix-glm-4.7-flash-mla-config

drifkin/qwen3-coder-opening-tag

brucemacd/usage-cli

fix-cuda12-fattn-shmem

ollama-imagegen-docs

parth/fix-multiline-inputs

brucemacd/config-docs

mxyng/model-files

mxyng/simple-execute

fix-imagegen-ollama-models

mxyng/async-upload

jmorganca/lazy-no-dtype-changes

imagegen-auto-detect-create

parth/decrease-concurrent-download-hf

fix-mlx-quantize-init

jmorganca/x-cleanup

usage

imagegen-readme

jmorganca/glm-image

mlx-gpu-cd

jmorganca/imagegen-modelfile

parth/agent-skills

parth/agent-allowlist

parth/signed-in-offline

parth/agents

parth/fix-context-chopping

improve-cloud-flow

parth/add-models-websearch

parth/prompt-renderer-mcp

jmorganca/native-settings

jmorganca/download-stream-hash

jmorganca/client2-rebased

brucemacd/oai-chat-req-multipart

jessegross/multi_chunk_reserve

grace/additional-omit-empty

grace/mistral-3-large

mxyng/tokenizer2

mxyng/tokenizer

jessegross/flash

hoyyeva/windows-nacked-app

mxyng/cleanup-attention

grace/deepseek-parser

hoyyeva/remember-unsent-prompt

parth/add-lfs-pointer-error-conversion

parth/olmo2-test2

hoyyeva/ollama-launchagent-plist

nicole/olmo-model

parth/olmo-test

mxyng/remove-embedded

parth/render-template

jmorganca/intellect-3

parth/remove-prealloc-linter

jmorganca/cmd-eval

nicole/nomic-embed-text-fix

mxyng/lint-2

hoyyeva/add-gemini-3-pro-preview

hoyyeva/load-model-list

mxyng/expand-path

mxyng/environ-2

hoyyeva/deeplink-json-encoding

parth/improve-tool-calling-tests

hoyyeva/conversation

hoyyeva/assistant-edit-response

hoyyeva/thinking

origin/brucemacd/invalid-char-i-err

parth/improve-tool-calling

jmorganca/required-omitempty

grace/qwen3-vl-tests

mxyng/iter-client

parth/docs-readme

nicole/embed-test

pdevine/integration-benchstat

parth/remove-generate-cmd

parth/add-toolcall-id

mxyng/server-tests

jmorganca/glm-4.6

jmorganca/gin-h-compat

drifkin/stable-tool-args

pdevine/qwen3-more-thinking

parth/add-websearch-client

nicole/websearch_local

jmorganca/qwen3-coder-updates

grace/deepseek-v3-migration-tests

mxyng/fix-create

jmorganca/cloud-errors

pdevine/parser-tidy

revert-12233-parth/simplify-entrypoints-runner

parth/enable-so-gpt-oss

brucemacd/qwen3vl

jmorganca/readme-simplify

parth/gpt-oss-structured-outputs

revert-12039-jmorganca/tools-braces

mxyng/embeddings

mxyng/gguf

mxyng/benchmark

mxyng/types-null

parth/move-parsing

mxyng/gemma2

jmorganca/docs

mxyng/16-bit

mxyng/create-stdin

pdevine/authorizedkeys

mxyng/quant

parth/opt-in-error-context-window

brucemacd/cache-models

brucemacd/runner-completion

jmorganca/llama-update-6

brucemacd/benchmark-list

brucemacd/partial-read-caps

parth/deepseek-r1-tools

mxyng/omit-array

parth/tool-prefix-temp

brucemacd/runner-test

jmorganca/qwen25vl

brucemacd/model-forward-test-ext

parth/python-function-parsing

jmorganca/cuda-compression-none

drifkin/num-parallel

drifkin/chat-truncation-fix

jmorganca/sync

parth/python-tools-calling

drifkin/array-head-count

brucemacd/create-no-loop

parth/server-enable-content-stream-with-tools

qwen25omni

mxyng/v3

brucemacd/ropeconfig

jmorganca/silence-tokenizer

parth/sample-so-test

parth/sampling-structured-outputs

brucemacd/doc-go-engine

parth/constrained-sampling-json

jmorganca/mistral-wip

brucemacd/mistral-small-convert

parth/sample-unmarshal-json-for-params

brucemacd/jomorganca/mistral

pdevine/bfloat16

jmorganca/mistral

brucemacd/mistral

pdevine/logging

parth/sample-correctness-fix

parth/sample-fix-sorting

jmorgan/sample-fix-sorting-extras

jmorganca/temp-0-images

brucemacd/parallel-embed-models

brucemacd/shim-grammar

jmorganca/fix-gguf-error

bmizerany/nameswork

jmorganca/faster-releases

bmizerany/validatenames

brucemacd/err-no-vocab

brucemacd/rope-config

brucemacd/err-hint

brucemacd/qwen2_5

brucemacd/logprobs

brucemacd/new_runner_graph_bench

progress-flicker

brucemacd/forward-test

brucemacd/go_qwen2

pdevine/gemma2

jmorganca/add-missing-symlink-eval

mxyng/next-debug

parth/set-context-size-openai

brucemacd/next-bpe-bench

brucemacd/next-bpe-test

brucemacd/new_runner_e2e

brucemacd/new_runner_qwen2

pdevine/convert-cohere2

brucemacd/convert-cli

parth/log-probs

mxyng/next-mlx

mxyng/cmd-history

parth/templating

parth/tokenize-detokenize

brucemacd/check-key-register

bmizerany/grammar

jmorganca/vendor-081b29bd

mxyng/func-checks

jmorganca/fix-null-format

parth/fix-default-to-warn-json

jmorganca/qwen2vl

jmorganca/no-concat

parth/cmd-cleanup-SO

brucemacd/check-key-register-structured-err

parth/openai-stream-usage

parth/fix-referencing-so

stream-tools-stop

jmorganca/degin-1

brucemacd/install-path-clean

brucemacd/push-name-validation

brucemacd/browser-key-register

jmorganca/openai-fix-first-message

jmorganca/fix-proxy

jessegross/sample

parth/disallow-streaming-tools

dhiltgen/remove_submodule

jmorganca/ga

jmorganca/mllama

pdevine/newlines

pdevine/geems-2b

jmorganca/llama-bump

mxyng/modelname-7

mxyng/gin-slog

mxyng/modelname-6

jyan/convert-prog

jyan/quant5

paligemma-support

pdevine/import-docs

jmorganca/openai-context

jyan/paligemma

jyan/p2

jyan/palitest

bmizerany/embedspeedup

jmorganca/llama-vit

brucemacd/allow-ollama

royh/ep-methods

royh/whisper

mxyng/api-models

mxyng/fix-memory

jyan/q4_4/8

jyan/ollama-v

royh/stream-tools

roy-embed-parallel

bmizerany/hrm

revert-5963-revert-5924-mxyng/llama3.1-rope

royh/embed-viz

jyan/local2

jyan/auth

jyan/local

jyan/parse-temp

jmorganca/template-mistral

jyan/reord-g

royh-openai-suffixdocs

royh-imgembed

royh-embed-parallel

jyan/quant4

royh-precision

jyan/progress

pdevine/fix-template

jyan/quant3

pdevine/ggla

mxyng/update-registry-domain

jmorganca/ggml-static

mxyng/create-context

jyan/v0.146

mxyng/layers-from-files

build_dist

bmizerany/noseek

royh-ls

royh-name

timeout

mxyng/server-timestamp

bmizerany/nosillyggufslurps

royh-params

jmorganca/llama-cpp-7c26775

royh-openai-delete

royh-show-rigid

jmorganca/enable-fa

jmorganca/no-error-template

jyan/format

royh-testdelete

bmizerany/fastverify

language_support

pdevine/ps-glitches

brucemacd/tokenize

bruce/iq-quants

bmizerany/filepathwithcoloninhost

mxyng/split-bin

bmizerany/client-registry

jmorganca/if-none-match

native

jmorganca/native

jmorganca/batch-embeddings

jmorganca/initcmake

jmorganca/mm

pdevine/showggmlinfo

modenameenforcealphanum

bmizerany/modenameenforcealphanum

jmorganca/done-reason

jmorganca/llama-cpp-8960fe8

ollama.com

bmizerany/filepathnobuild

bmizerany/types/model/defaultfix

rmdisplaylong

nogogen

bmizerany/x

modelfile-readme

bmizerany/replacecolon

jmorganca/limit

jmorganca/execstack

jmorganca/replace-assets

mxyng/tune-concurrency

jmorganca/testing

whitespace-detection

jmorganca/options

upgrade-all

scratch

cuda-search

mattw/airenamer

mattw/allmodelsonhuggingface

mattw/quantcontext

mattw/whatneedstorun

brucemacd/llama-mem-calc

mattw/faq-context

mattw/communitylinks

mattw/noprune

mattw/python-functioncalling

rename

mxyng/install

pulse

remove-first

editor

mattw/selfqueryingretrieval

cgo

mattw/howtoquant

api

matt/streamingapi

format-config

mxyng/extra-args

shell

update-nous-hermes

cp-model

upload-progress

fix-unknown-model

fix-model-names

delete-fix

insecure-registry

ls

deletemodels

progressbar

readme-updates

license-layers

skip-list

list-models

modelpath

matt/examplemodelfiles

distribution

go-opts

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/ollama#67410