github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #7005] Docker not use GPU after idle #50949

New Issue
Closed
opened 2026-04-28 17:38:57 -05:00 by GiteaMirror · 7 comments
GiteaMirror commented 2026-04-28 17:38:57 -05:00
Owner
Copy Link

Originally created by @phukrit7171 on GitHub (Sep 27, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7005

Originally assigned to: @dhiltgen on GitHub.

What is the issue?

After the model is cleared from the graphics card RAM, when it is run again, the model is not loaded to the graphics card RAM but runs on the CPU instead, which slows it down a lot. You have to do docker stop ollama and docker start ollama to get it to run again with the graphics card.

OS

Linux, Docker

GPU

Nvidia

CPU

Intel

Ollama version

0.3.12

Originally created by @phukrit7171 on GitHub (Sep 27, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7005 Originally assigned to: @dhiltgen on GitHub. ### What is the issue? After the model is cleared from the graphics card RAM, when it is run again, the model is not loaded to the graphics card RAM but runs on the CPU instead, which slows it down a lot. You have to do docker stop ollama and docker start ollama to get it to run again with the graphics card. ### OS Linux, Docker ### GPU Nvidia ### CPU Intel ### Ollama version 0.3.12
GiteaMirror added the dockerneeds more infobugnvidia labels 2026-04-28 17:38:58 -05:00
GiteaMirror commented 2026-04-28 17:39:05 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Sep 28, 2024):

Server logs will help in debugging.

<!-- gh-comment-id:2380305786 --> @rick-github commented on GitHub (Sep 28, 2024): [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) will help in debugging.
GiteaMirror commented 2026-04-28 17:39:08 -05:00
Author
Owner
Copy Link

@phukrit7171 commented on GitHub (Sep 28, 2024):

There is a part that shows "msg=gpu VRAM usage didn't recover within timeout" and "error looking up nvidia GPU memory"

<!-- gh-comment-id:2380735483 --> @phukrit7171 commented on GitHub (Sep 28, 2024): There is a part that shows "msg=gpu VRAM usage didn't recover within timeout" and "error looking up nvidia GPU memory"
GiteaMirror commented 2026-04-28 17:39:09 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Sep 28, 2024):

@phukrit7171 we'll need a more complete server log to be able to understand what's going wrong. There's likely some error being returned by one of the cuda APIs we call. If you're unable to share the log, please try the troubleshooting steps for nvidia and hopefully one of them resolves your problem - https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#linux-nvidia-troubleshooting

<!-- gh-comment-id:2381017110 --> @dhiltgen commented on GitHub (Sep 28, 2024): @phukrit7171 we'll need a more complete server log to be able to understand what's going wrong. There's likely some error being returned by one of the cuda APIs we call. If you're unable to share the log, please try the troubleshooting steps for nvidia and hopefully one of them resolves your problem - https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#linux-nvidia-troubleshooting
GiteaMirror commented 2026-04-28 17:39:10 -05:00
Author
Owner
Copy Link

@phukrit7171 commented on GitHub (Sep 29, 2024):

OK This is full log
`
2024/09/27 16:53:08 routes.go:1153: 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:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/root/.ollama/models 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://] OLLAMA_SCHED_SPREAD:false OLLAMA_TMPDIR: ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]"
time=2024-09-27T16:53:08.216Z level=INFO source=images.go:753 msg="total blobs: 5"
time=2024-09-27T16:53:08.216Z level=INFO source=images.go:760 msg="total unused blobs removed: 0"
time=2024-09-27T16:53:08.216Z level=INFO source=routes.go:1200 msg="Listening on [::]:11434 (version 0.3.12)"
time=2024-09-27T16:53:08.216Z level=INFO source=common.go:49 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11 cuda_v12]"
time=2024-09-27T16:53:08.216Z level=INFO source=gpu.go:199 msg="looking for compatible GPUs"
time=2024-09-27T16:53:08.568Z level=INFO source=types.go:107 msg="inference compute" id=GPU-5b501d39-1388-7ef8-1dd4-4c618162b80d library=cuda variant=v12 compute=8.6 driver=12.6 name="NVIDIA GeForce RTX 3070 Laptop GPU" total="7.7 GiB" available="6.5 GiB"
[GIN] 2024/09/27 - 16:53:26 | 200 | 40.736µs | 127.0.0.1 | HEAD "/"
[GIN] 2024/09/27 - 16:53:29 | 200 | 2.783019705s | 127.0.0.1 | POST "/api/pull"
time=2024-09-27T16:53:43.051Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
time=2024-09-27T16:53:43.237Z level=INFO source=server.go:103 msg="system memory" total="31.1 GiB" free="26.1 GiB" free_swap="8.0 GiB"
time=2024-09-27T16:53:43.238Z level=INFO source=memory.go:326 msg="offload to cuda" layers.requested=-1 layers.model=33 layers.offload=32 layers.split="" memory.available="[6.5 GiB]" memory.gpu_overhead="0 B" memory.required.full="6.6 GiB" memory.required.partial="6.2 GiB" memory.required.kv="512.0 MiB" memory.required.allocations="[6.2 GiB]" memory.weights.total="4.4 GiB" memory.weights.repeating="4.0 GiB" memory.weights.nonrepeating="411.0 MiB" memory.graph.full="296.0 MiB" memory.graph.partial="677.5 MiB"
time=2024-09-27T16:53:43.239Z level=INFO source=server.go:388 msg="starting llama server" cmd="/usr/lib/ollama/runners/cuda_v12/ollama_llama_server --model /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc --ctx-size 4096 --batch-size 512 --embedding --log-disable --n-gpu-layers 32 --mmproj /root/.ollama/models/blobs/sha256-eb569aba7d65cf3da1d0369610eb6869f4a53ee369992a804d5810a80e9fa035 --parallel 1 --port 44039"
time=2024-09-27T16:53:43.239Z level=INFO source=sched.go:449 msg="loaded runners" count=1
time=2024-09-27T16:53:43.239Z level=INFO source=server.go:587 msg="waiting for llama runner to start responding"
time=2024-09-27T16:53:43.239Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server error"
INFO [main] build info | build=10 commit="3f6ec33" tid="124849675476992" timestamp=1727456023
INFO [main] system info | n_threads=8 n_threads_batch=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="124849675476992" timestamp=1727456023 total_threads=16
INFO [main] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="44039" tid="124849675476992" timestamp=1727456023
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce RTX 3070 Laptop GPU, compute capability 8.6, VMM: yes
time=2024-09-27T16:53:43.491Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server loading model"
llama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv 0: general.architecture str = llama
llama_model_loader: - kv 1: general.name str = Downloads
llama_model_loader: - kv 2: llama.vocab_size u32 = 128256
llama_model_loader: - kv 3: llama.context_length u32 = 8192
llama_model_loader: - kv 4: llama.embedding_length u32 = 4096
llama_model_loader: - kv 5: llama.block_count u32 = 32
llama_model_loader: - kv 6: llama.feed_forward_length u32 = 14336
llama_model_loader: - kv 7: llama.rope.dimension_count u32 = 128
llama_model_loader: - kv 8: llama.attention.head_count u32 = 32
llama_model_loader: - kv 9: llama.attention.head_count_kv u32 = 8
llama_model_loader: - kv 10: llama.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 11: llama.rope.freq_base f32 = 500000.000000
llama_model_loader: - kv 12: general.file_type u32 = 15
llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 14: tokenizer.ggml.tokens arr[str,128256] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 15: tokenizer.ggml.scores arr[f32,128256] = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000
llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128001
llama_model_loader: - kv 20: tokenizer.chat_template str = {% set loop_messages = messages %}{% ...
llama_model_loader: - kv 21: general.quantization_version u32 = 2
llama_model_loader: - type f32: 65 tensors
llama_model_loader: - type q4_K: 193 tensors
llama_model_loader: - type q6_K: 33 tensors
llm_load_vocab: missing or unrecognized pre-tokenizer type, using: 'default'
llm_load_vocab: special tokens cache size = 256
llm_load_vocab: token to piece cache size = 0.8000 MB
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = llama
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 = 8192
llm_load_print_meta: n_embd = 4096
llm_load_print_meta: n_layer = 32
llm_load_print_meta: n_head = 32
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 = 4
llm_load_print_meta: n_embd_k_gqa = 1024
llm_load_print_meta: n_embd_v_gqa = 1024
llm_load_print_meta: f_norm_eps = 0.0e+00
llm_load_print_meta: f_norm_rms_eps = 1.0e-05
llm_load_print_meta: f_clamp_kqv = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale = 0.0e+00
llm_load_print_meta: n_ff = 14336
llm_load_print_meta: n_expert = 0
llm_load_print_meta: n_expert_used = 0
llm_load_print_meta: causal attn = 1
llm_load_print_meta: pooling type = 0
llm_load_print_meta: rope type = 0
llm_load_print_meta: rope scaling = linear
llm_load_print_meta: freq_base_train = 500000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn = 8192
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 = 8B
llm_load_print_meta: model ftype = Q4_K - Medium
llm_load_print_meta: model params = 8.03 B
llm_load_print_meta: model size = 4.58 GiB (4.89 BPW)
llm_load_print_meta: general.name = Downloads
llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>'
llm_load_print_meta: EOS token = 128001 '<|end_of_text|>'
llm_load_print_meta: LF token = 128 'Ä'
llm_load_print_meta: EOT token = 128009 '<|eot_id|>'
llm_load_print_meta: max token length = 256
llm_load_tensors: ggml ctx size = 0.27 MiB
llm_load_tensors: offloading 32 repeating layers to GPU
llm_load_tensors: offloaded 32/33 layers to GPU
llm_load_tensors: CPU buffer size = 4685.30 MiB
llm_load_tensors: CUDA0 buffer size = 3992.50 MiB
llama_new_context_with_model: n_ctx = 4096
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_kv_cache_init: CUDA0 KV buffer size = 512.00 MiB
llama_new_context_with_model: KV self size = 512.00 MiB, K (f16): 256.00 MiB, V (f16): 256.00 MiB
llama_new_context_with_model: CUDA_Host output buffer size = 0.50 MiB
llama_new_context_with_model: CUDA0 compute buffer size = 669.48 MiB
llama_new_context_with_model: CUDA_Host compute buffer size = 16.01 MiB
llama_new_context_with_model: graph nodes = 1030
llama_new_context_with_model: graph splits = 4
INFO [main] model loaded | tid="124849675476992" timestamp=1727456024
time=2024-09-27T16:53:44.996Z level=INFO source=server.go:626 msg="llama runner started in 1.76 seconds"
[GIN] 2024/09/27 - 16:53:46 | 200 | 2.997732217s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:53:58.095Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:53:59 | 200 | 1.5877281s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:54:14.003Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:54:15 | 200 | 1.85722756s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:54:29.837Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:54:31 | 200 | 1.697664331s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:55:08.819Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:55:14 | 200 | 5.223275133s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:55:56.326Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:56:03 | 200 | 7.24924078s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:57:16.592Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:57:18 | 200 | 2.37922811s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:57:30.667Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:57:33 | 200 | 2.544740166s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:57:42.142Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:57:44 | 200 | 2.738373745s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:57:51.989Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:57:55 | 200 | 3.131983296s | 172.17.0.1 | POST "/api/chat"
time=2024-09-27T16:58:11.778Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
[GIN] 2024/09/27 - 16:58:15 | 200 | 3.414680324s | 172.17.0.1 | POST "/api/chat"
cuda driver library failed to get device context 800time=2024-09-27T17:03:15.195Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:15.472Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:15.704Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:15.954Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:16.203Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:16.452Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:16.701Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:16.952Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:17.203Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:17.452Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:17.701Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:17.950Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:18.201Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:18.452Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:18.701Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:18.951Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:19.201Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:19.455Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:19.703Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
cuda driver library failed to get device context 800time=2024-09-27T17:03:19.952Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
time=2024-09-27T17:03:20.196Z level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.018335295 model=/root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc
cuda driver library failed to get device context 800time=2024-09-27T17:03:20.202Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
time=2024-09-27T17:03:20.446Z level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.268763433 model=/root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc
cuda driver library failed to get device context 800time=2024-09-27T17:03:20.456Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
time=2024-09-27T17:03:20.695Z level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.518129721 model=/root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc
time=2024-09-27T17:08:41.164Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
cuda driver library failed to get device context 800time=2024-09-27T17:08:41.197Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory"
time=2024-09-27T17:08:41.239Z level=INFO source=server.go:103 msg="system memory" total="31.1 GiB" free="25.1 GiB" free_swap="8.0 GiB"
time=2024-09-27T17:08:41.240Z level=INFO source=memory.go:326 msg="offload to cuda" layers.requested=-1 layers.model=33 layers.offload=32 layers.split="" memory.available="[6.5 GiB]" memory.gpu_overhead="0 B" memory.required.full="6.6 GiB" memory.required.partial="6.2 GiB" memory.required.kv="512.0 MiB" memory.required.allocations="[6.2 GiB]" memory.weights.total="4.4 GiB" memory.weights.repeating="4.0 GiB" memory.weights.nonrepeating="411.0 MiB" memory.graph.full="296.0 MiB" memory.graph.partial="677.5 MiB"
time=2024-09-27T17:08:41.242Z level=INFO source=server.go:388 msg="starting llama server" cmd="/usr/lib/ollama/runners/cuda_v12/ollama_llama_server --model /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc --ctx-size 4096 --batch-size 512 --embedding --log-disable --n-gpu-layers 32 --mmproj /root/.ollama/models/blobs/sha256-eb569aba7d65cf3da1d0369610eb6869f4a53ee369992a804d5810a80e9fa035 --parallel 1 --port 40321"
time=2024-09-27T17:08:41.242Z level=INFO source=sched.go:449 msg="loaded runners" count=1
time=2024-09-27T17:08:41.242Z level=INFO source=server.go:587 msg="waiting for llama runner to start responding"
time=2024-09-27T17:08:41.242Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server error"
INFO [main] build info | build=10 commit="3f6ec33" tid="126518886834176" timestamp=1727456921
INFO [main] system info | n_threads=8 n_threads_batch=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="126518886834176" timestamp=1727456921 total_threads=16
INFO [main] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="40321" tid="126518886834176" timestamp=1727456921
ggml_cuda_init: failed to initialize CUDA: no CUDA-capable device is detected
ggml_backend_cuda_init: invalid device 0
time=2024-09-27T17:08:41.492Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server loading model"
llama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv 0: general.architecture str = llama
llama_model_loader: - kv 1: general.name str = Downloads
llama_model_loader: - kv 2: llama.vocab_size u32 = 128256
llama_model_loader: - kv 3: llama.context_length u32 = 8192
llama_model_loader: - kv 4: llama.embedding_length u32 = 4096
llama_model_loader: - kv 5: llama.block_count u32 = 32
llama_model_loader: - kv 6: llama.feed_forward_length u32 = 14336
llama_model_loader: - kv 7: llama.rope.dimension_count u32 = 128
llama_model_loader: - kv 8: llama.attention.head_count u32 = 32
llama_model_loader: - kv 9: llama.attention.head_count_kv u32 = 8
llama_model_loader: - kv 10: llama.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 11: llama.rope.freq_base f32 = 500000.000000
llama_model_loader: - kv 12: general.file_type u32 = 15
llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 14: tokenizer.ggml.tokens arr[str,128256] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 15: tokenizer.ggml.scores arr[f32,128256] = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000
llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128001
llama_model_loader: - kv 20: tokenizer.chat_template str = {% set loop_messages = messages %}{% ...
llama_model_loader: - kv 21: general.quantization_version u32 = 2
llama_model_loader: - type f32: 65 tensors
llama_model_loader: - type q4_K: 193 tensors
llama_model_loader: - type q6_K: 33 tensors
llm_load_vocab: missing or unrecognized pre-tokenizer type, using: 'default'
llm_load_vocab: special tokens cache size = 256
llm_load_vocab: token to piece cache size = 0.8000 MB
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = llama
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 = 8192
llm_load_print_meta: n_embd = 4096
llm_load_print_meta: n_layer = 32
llm_load_print_meta: n_head = 32
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 = 4
llm_load_print_meta: n_embd_k_gqa = 1024
llm_load_print_meta: n_embd_v_gqa = 1024
llm_load_print_meta: f_norm_eps = 0.0e+00
llm_load_print_meta: f_norm_rms_eps = 1.0e-05
llm_load_print_meta: f_clamp_kqv = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale = 0.0e+00
llm_load_print_meta: n_ff = 14336
llm_load_print_meta: n_expert = 0
llm_load_print_meta: n_expert_used = 0
llm_load_print_meta: causal attn = 1
llm_load_print_meta: pooling type = 0
llm_load_print_meta: rope type = 0
llm_load_print_meta: rope scaling = linear
llm_load_print_meta: freq_base_train = 500000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn = 8192
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 = 8B
llm_load_print_meta: model ftype = Q4_K - Medium
llm_load_print_meta: model params = 8.03 B
llm_load_print_meta: model size = 4.58 GiB (4.89 BPW)
llm_load_print_meta: general.name = Downloads
llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>'
llm_load_print_meta: EOS token = 128001 '<|end_of_text|>'
llm_load_print_meta: LF token = 128 'Ä'
llm_load_print_meta: EOT token = 128009 '<|eot_id|>'
llm_load_print_meta: max token length = 256
llm_load_tensors: ggml ctx size = 0.14 MiB
llm_load_tensors: offloading 32 repeating layers to GPU
llm_load_tensors: offloaded 32/33 layers to GPU
llm_load_tensors: CPU buffer size = 4685.30 MiB
llama_new_context_with_model: n_ctx = 4096
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
ggml_cuda_host_malloc: failed to allocate 512.00 MiB of pinned memory: no CUDA-capable device is detected
llama_kv_cache_init: CPU KV buffer size = 512.00 MiB
llama_new_context_with_model: KV self size = 512.00 MiB, K (f16): 256.00 MiB, V (f16): 256.00 MiB
ggml_cuda_host_malloc: failed to allocate 0.50 MiB of pinned memory: no CUDA-capable device is detected
llama_new_context_with_model: CPU output buffer size = 0.50 MiB
ggml_cuda_host_malloc: failed to allocate 296.01 MiB of pinned memory: no CUDA-capable device is detected
llama_new_context_with_model: CUDA_Host compute buffer size = 296.01 MiB
llama_new_context_with_model: graph nodes = 1030
llama_new_context_with_model: graph splits = 1
INFO [main] model loaded | tid="126518886834176" timestamp=1727456922
time=2024-09-27T17:08:42.497Z level=INFO source=server.go:626 msg="llama runner started in 1.25 seconds"
[GIN] 2024/09/27 - 17:09:02 | 200 | 21.067207572s | 172.17.0.1 | POST "/api/chat"
2024/09/27 17:09:27 routes.go:1153: 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:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/root/.ollama/models 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://*] OLLAMA_SCHED_SPREAD:false OLLAMA_TMPDIR: ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]"
time=2024-09-27T17:09:27.643Z level=INFO source=images.go:753 msg="total blobs: 5"
time=2024-09-27T17:09:27.643Z level=INFO source=images.go:760 msg="total unused blobs removed: 0"
time=2024-09-27T17:09:27.643Z level=INFO source=routes.go:1200 msg="Listening on [::]:11434 (version 0.3.12)"
time=2024-09-27T17:09:27.643Z level=INFO source=common.go:49 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11 cuda_v12]"
time=2024-09-27T17:09:27.643Z level=INFO source=gpu.go:199 msg="looking for compatible GPUs"
time=2024-09-27T17:09:27.992Z level=INFO source=types.go:107 msg="inference compute" id=GPU-5b501d39-1388-7ef8-1dd4-4c618162b80d library=cuda variant=v12 compute=8.6 driver=12.6 name="NVIDIA GeForce RTX 3070 Laptop GPU" total="7.7 GiB" available="6.1 GiB"
time=2024-09-27T17:09:32.464Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet"
time=2024-09-27T17:09:32.657Z level=INFO source=server.go:103 msg="system memory" total="31.1 GiB" free="25.0 GiB" free_swap="8.0 GiB"
time=2024-09-27T17:09:32.658Z level=INFO source=memory.go:326 msg="offload to cuda" layers.requested=-1 layers.model=33 layers.offload=30 layers.split="" memory.available="[6.1 GiB]" memory.gpu_overhead="0 B" memory.required.full="6.6 GiB" memory.required.partial="5.9 GiB" memory.required.kv="512.0 MiB" memory.required.allocations="[5.9 GiB]" memory.weights.total="4.4 GiB" memory.weights.repeating="4.0 GiB" memory.weights.nonrepeating="411.0 MiB" memory.graph.full="296.0 MiB" memory.graph.partial="677.5 MiB"
time=2024-09-27T17:09:32.659Z level=INFO source=server.go:388 msg="starting llama server" cmd="/usr/lib/ollama/runners/cuda_v12/ollama_llama_server --model /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc --ctx-size 4096 --batch-size 512 --embedding --log-disable --n-gpu-layers 30 --mmproj /root/.ollama/models/blobs/sha256-eb569aba7d65cf3da1d0369610eb6869f4a53ee369992a804d5810a80e9fa035 --parallel 1 --port 42495"
time=2024-09-27T17:09:32.659Z level=INFO source=sched.go:449 msg="loaded runners" count=1
time=2024-09-27T17:09:32.659Z level=INFO source=server.go:587 msg="waiting for llama runner to start responding"
time=2024-09-27T17:09:32.659Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server error"
INFO [main] build info | build=10 commit="3f6ec33" tid="140507753418752" timestamp=1727456972
INFO [main] system info | n_threads=8 n_threads_batch=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="140507753418752" timestamp=1727456972 total_threads=16
INFO [main] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="42495" tid="140507753418752" timestamp=1727456972
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce RTX 3070 Laptop GPU, compute capability 8.6, VMM: yes
time=2024-09-27T17:09:32.911Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server loading model"
llama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv 0: general.architecture str = llama
llama_model_loader: - kv 1: general.name str = Downloads
llama_model_loader: - kv 2: llama.vocab_size u32 = 128256
llama_model_loader: - kv 3: llama.context_length u32 = 8192
llama_model_loader: - kv 4: llama.embedding_length u32 = 4096
llama_model_loader: - kv 5: llama.block_count u32 = 32
llama_model_loader: - kv 6: llama.feed_forward_length u32 = 14336
llama_model_loader: - kv 7: llama.rope.dimension_count u32 = 128
llama_model_loader: - kv 8: llama.attention.head_count u32 = 32
llama_model_loader: - kv 9: llama.attention.head_count_kv u32 = 8
llama_model_loader: - kv 10: llama.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 11: llama.rope.freq_base f32 = 500000.000000
llama_model_loader: - kv 12: general.file_type u32 = 15
llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 14: tokenizer.ggml.tokens arr[str,128256] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 15: tokenizer.ggml.scores arr[f32,128256] = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000
llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128001
llama_model_loader: - kv 20: tokenizer.chat_template str = {% set loop_messages = messages %}{% ...
llama_model_loader: - kv 21: general.quantization_version u32 = 2
llama_model_loader: - type f32: 65 tensors
llama_model_loader: - type q4_K: 193 tensors
llama_model_loader: - type q6_K: 33 tensors
llm_load_vocab: missing or unrecognized pre-tokenizer type, using: 'default'
llm_load_vocab: special tokens cache size = 256
llm_load_vocab: token to piece cache size = 0.8000 MB
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = llama
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 = 8192
llm_load_print_meta: n_embd = 4096
llm_load_print_meta: n_layer = 32
llm_load_print_meta: n_head = 32
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 = 4
llm_load_print_meta: n_embd_k_gqa = 1024
llm_load_print_meta: n_embd_v_gqa = 1024
llm_load_print_meta: f_norm_eps = 0.0e+00
llm_load_print_meta: f_norm_rms_eps = 1.0e-05
llm_load_print_meta: f_clamp_kqv = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale = 0.0e+00
llm_load_print_meta: n_ff = 14336
llm_load_print_meta: n_expert = 0
llm_load_print_meta: n_expert_used = 0
llm_load_print_meta: causal attn = 1
llm_load_print_meta: pooling type = 0
llm_load_print_meta: rope type = 0
llm_load_print_meta: rope scaling = linear
llm_load_print_meta: freq_base_train = 500000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn = 8192
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 = 8B
llm_load_print_meta: model ftype = Q4_K - Medium
llm_load_print_meta: model params = 8.03 B
llm_load_print_meta: model size = 4.58 GiB (4.89 BPW)
llm_load_print_meta: general.name = Downloads
llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>'
llm_load_print_meta: EOS token = 128001 '<|end_of_text|>'
llm_load_print_meta: LF token = 128 'Ä'
llm_load_print_meta: EOT token = 128009 '<|eot_id|>'
llm_load_print_meta: max token length = 256
llm_load_tensors: ggml ctx size = 0.27 MiB
llm_load_tensors: offloading 30 repeating layers to GPU
llm_load_tensors: offloaded 30/33 layers to GPU
llm_load_tensors: CPU buffer size = 4685.30 MiB
llm_load_tensors: CUDA0 buffer size = 3727.50 MiB
llama_new_context_with_model: n_ctx = 4096
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_kv_cache_init: CUDA_Host KV buffer size = 32.00 MiB
llama_kv_cache_init: CUDA0 KV buffer size = 480.00 MiB
llama_new_context_with_model: KV self size = 512.00 MiB, K (f16): 256.00 MiB, V (f16): 256.00 MiB
llama_new_context_with_model: CUDA_Host output buffer size = 0.50 MiB
llama_new_context_with_model: CUDA0 compute buffer size = 669.48 MiB
llama_new_context_with_model: CUDA_Host compute buffer size = 16.01 MiB
llama_new_context_with_model: graph nodes = 1030
llama_new_context_with_model: graph splits = 26
INFO [main] model loaded | tid="140507753418752" timestamp=1727456974
time=2024-09-27T17:09:34.418Z level=INFO source=server.go:626 msg="llama runner started in 1.76 seconds"
[GIN] 2024/09/27 - 17:09:38 | 200 | 5.737602399s | 172.17.0.1 | POST "/api/chat"

`

<!-- gh-comment-id:2381121453 --> @phukrit7171 commented on GitHub (Sep 29, 2024): OK This is full log ` 2024/09/27 16:53:08 routes.go:1153: 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:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/root/.ollama/models 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://*] OLLAMA_SCHED_SPREAD:false OLLAMA_TMPDIR: ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]" time=2024-09-27T16:53:08.216Z level=INFO source=images.go:753 msg="total blobs: 5" time=2024-09-27T16:53:08.216Z level=INFO source=images.go:760 msg="total unused blobs removed: 0" time=2024-09-27T16:53:08.216Z level=INFO source=routes.go:1200 msg="Listening on [::]:11434 (version 0.3.12)" time=2024-09-27T16:53:08.216Z level=INFO source=common.go:49 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11 cuda_v12]" time=2024-09-27T16:53:08.216Z level=INFO source=gpu.go:199 msg="looking for compatible GPUs" time=2024-09-27T16:53:08.568Z level=INFO source=types.go:107 msg="inference compute" id=GPU-5b501d39-1388-7ef8-1dd4-4c618162b80d library=cuda variant=v12 compute=8.6 driver=12.6 name="NVIDIA GeForce RTX 3070 Laptop GPU" total="7.7 GiB" available="6.5 GiB" [GIN] 2024/09/27 - 16:53:26 | 200 | 40.736µs | 127.0.0.1 | HEAD "/" [GIN] 2024/09/27 - 16:53:29 | 200 | 2.783019705s | 127.0.0.1 | POST "/api/pull" time=2024-09-27T16:53:43.051Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" time=2024-09-27T16:53:43.237Z level=INFO source=server.go:103 msg="system memory" total="31.1 GiB" free="26.1 GiB" free_swap="8.0 GiB" time=2024-09-27T16:53:43.238Z level=INFO source=memory.go:326 msg="offload to cuda" layers.requested=-1 layers.model=33 layers.offload=32 layers.split="" memory.available="[6.5 GiB]" memory.gpu_overhead="0 B" memory.required.full="6.6 GiB" memory.required.partial="6.2 GiB" memory.required.kv="512.0 MiB" memory.required.allocations="[6.2 GiB]" memory.weights.total="4.4 GiB" memory.weights.repeating="4.0 GiB" memory.weights.nonrepeating="411.0 MiB" memory.graph.full="296.0 MiB" memory.graph.partial="677.5 MiB" time=2024-09-27T16:53:43.239Z level=INFO source=server.go:388 msg="starting llama server" cmd="/usr/lib/ollama/runners/cuda_v12/ollama_llama_server --model /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc --ctx-size 4096 --batch-size 512 --embedding --log-disable --n-gpu-layers 32 --mmproj /root/.ollama/models/blobs/sha256-eb569aba7d65cf3da1d0369610eb6869f4a53ee369992a804d5810a80e9fa035 --parallel 1 --port 44039" time=2024-09-27T16:53:43.239Z level=INFO source=sched.go:449 msg="loaded runners" count=1 time=2024-09-27T16:53:43.239Z level=INFO source=server.go:587 msg="waiting for llama runner to start responding" time=2024-09-27T16:53:43.239Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server error" INFO [main] build info | build=10 commit="3f6ec33" tid="124849675476992" timestamp=1727456023 INFO [main] system info | n_threads=8 n_threads_batch=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="124849675476992" timestamp=1727456023 total_threads=16 INFO [main] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="44039" tid="124849675476992" timestamp=1727456023 ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 1 CUDA devices: Device 0: NVIDIA GeForce RTX 3070 Laptop GPU, compute capability 8.6, VMM: yes time=2024-09-27T16:53:43.491Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server loading model" llama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc (version GGUF V3 (latest)) llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. llama_model_loader: - kv 0: general.architecture str = llama llama_model_loader: - kv 1: general.name str = Downloads llama_model_loader: - kv 2: llama.vocab_size u32 = 128256 llama_model_loader: - kv 3: llama.context_length u32 = 8192 llama_model_loader: - kv 4: llama.embedding_length u32 = 4096 llama_model_loader: - kv 5: llama.block_count u32 = 32 llama_model_loader: - kv 6: llama.feed_forward_length u32 = 14336 llama_model_loader: - kv 7: llama.rope.dimension_count u32 = 128 llama_model_loader: - kv 8: llama.attention.head_count u32 = 32 llama_model_loader: - kv 9: llama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 10: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 11: llama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 12: general.file_type u32 = 15 llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 14: tokenizer.ggml.tokens arr[str,128256] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 15: tokenizer.ggml.scores arr[f32,128256] = [0.000000, 0.000000, 0.000000, 0.0000... llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128001 llama_model_loader: - kv 20: tokenizer.chat_template str = {% set loop_messages = messages %}{% ... llama_model_loader: - kv 21: general.quantization_version u32 = 2 llama_model_loader: - type f32: 65 tensors llama_model_loader: - type q4_K: 193 tensors llama_model_loader: - type q6_K: 33 tensors llm_load_vocab: missing or unrecognized pre-tokenizer type, using: 'default' llm_load_vocab: special tokens cache size = 256 llm_load_vocab: token to piece cache size = 0.8000 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = llama 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 = 8192 llm_load_print_meta: n_embd = 4096 llm_load_print_meta: n_layer = 32 llm_load_print_meta: n_head = 32 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 = 4 llm_load_print_meta: n_embd_k_gqa = 1024 llm_load_print_meta: n_embd_v_gqa = 1024 llm_load_print_meta: f_norm_eps = 0.0e+00 llm_load_print_meta: f_norm_rms_eps = 1.0e-05 llm_load_print_meta: f_clamp_kqv = 0.0e+00 llm_load_print_meta: f_max_alibi_bias = 0.0e+00 llm_load_print_meta: f_logit_scale = 0.0e+00 llm_load_print_meta: n_ff = 14336 llm_load_print_meta: n_expert = 0 llm_load_print_meta: n_expert_used = 0 llm_load_print_meta: causal attn = 1 llm_load_print_meta: pooling type = 0 llm_load_print_meta: rope type = 0 llm_load_print_meta: rope scaling = linear llm_load_print_meta: freq_base_train = 500000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_ctx_orig_yarn = 8192 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 = 8B llm_load_print_meta: model ftype = Q4_K - Medium llm_load_print_meta: model params = 8.03 B llm_load_print_meta: model size = 4.58 GiB (4.89 BPW) llm_load_print_meta: general.name = Downloads llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128001 '<|end_of_text|>' llm_load_print_meta: LF token = 128 'Ä' llm_load_print_meta: EOT token = 128009 '<|eot_id|>' llm_load_print_meta: max token length = 256 llm_load_tensors: ggml ctx size = 0.27 MiB llm_load_tensors: offloading 32 repeating layers to GPU llm_load_tensors: offloaded 32/33 layers to GPU llm_load_tensors: CPU buffer size = 4685.30 MiB llm_load_tensors: CUDA0 buffer size = 3992.50 MiB llama_new_context_with_model: n_ctx = 4096 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_kv_cache_init: CUDA0 KV buffer size = 512.00 MiB llama_new_context_with_model: KV self size = 512.00 MiB, K (f16): 256.00 MiB, V (f16): 256.00 MiB llama_new_context_with_model: CUDA_Host output buffer size = 0.50 MiB llama_new_context_with_model: CUDA0 compute buffer size = 669.48 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 16.01 MiB llama_new_context_with_model: graph nodes = 1030 llama_new_context_with_model: graph splits = 4 INFO [main] model loaded | tid="124849675476992" timestamp=1727456024 time=2024-09-27T16:53:44.996Z level=INFO source=server.go:626 msg="llama runner started in 1.76 seconds" [GIN] 2024/09/27 - 16:53:46 | 200 | 2.997732217s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:53:58.095Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:53:59 | 200 | 1.5877281s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:54:14.003Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:54:15 | 200 | 1.85722756s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:54:29.837Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:54:31 | 200 | 1.697664331s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:55:08.819Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:55:14 | 200 | 5.223275133s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:55:56.326Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:56:03 | 200 | 7.24924078s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:57:16.592Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:57:18 | 200 | 2.37922811s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:57:30.667Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:57:33 | 200 | 2.544740166s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:57:42.142Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:57:44 | 200 | 2.738373745s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:57:51.989Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:57:55 | 200 | 3.131983296s | 172.17.0.1 | POST "/api/chat" time=2024-09-27T16:58:11.778Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" [GIN] 2024/09/27 - 16:58:15 | 200 | 3.414680324s | 172.17.0.1 | POST "/api/chat" cuda driver library failed to get device context 800time=2024-09-27T17:03:15.195Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:15.472Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:15.704Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:15.954Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:16.203Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:16.452Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:16.701Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:16.952Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:17.203Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:17.452Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:17.701Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:17.950Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:18.201Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:18.452Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:18.701Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:18.951Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:19.201Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:19.455Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:19.703Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" cuda driver library failed to get device context 800time=2024-09-27T17:03:19.952Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" time=2024-09-27T17:03:20.196Z level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.018335295 model=/root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc cuda driver library failed to get device context 800time=2024-09-27T17:03:20.202Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" time=2024-09-27T17:03:20.446Z level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.268763433 model=/root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc cuda driver library failed to get device context 800time=2024-09-27T17:03:20.456Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" time=2024-09-27T17:03:20.695Z level=WARN source=sched.go:646 msg="gpu VRAM usage didn't recover within timeout" seconds=5.518129721 model=/root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc time=2024-09-27T17:08:41.164Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" cuda driver library failed to get device context 800time=2024-09-27T17:08:41.197Z level=WARN source=gpu.go:400 msg="error looking up nvidia GPU memory" time=2024-09-27T17:08:41.239Z level=INFO source=server.go:103 msg="system memory" total="31.1 GiB" free="25.1 GiB" free_swap="8.0 GiB" time=2024-09-27T17:08:41.240Z level=INFO source=memory.go:326 msg="offload to cuda" layers.requested=-1 layers.model=33 layers.offload=32 layers.split="" memory.available="[6.5 GiB]" memory.gpu_overhead="0 B" memory.required.full="6.6 GiB" memory.required.partial="6.2 GiB" memory.required.kv="512.0 MiB" memory.required.allocations="[6.2 GiB]" memory.weights.total="4.4 GiB" memory.weights.repeating="4.0 GiB" memory.weights.nonrepeating="411.0 MiB" memory.graph.full="296.0 MiB" memory.graph.partial="677.5 MiB" time=2024-09-27T17:08:41.242Z level=INFO source=server.go:388 msg="starting llama server" cmd="/usr/lib/ollama/runners/cuda_v12/ollama_llama_server --model /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc --ctx-size 4096 --batch-size 512 --embedding --log-disable --n-gpu-layers 32 --mmproj /root/.ollama/models/blobs/sha256-eb569aba7d65cf3da1d0369610eb6869f4a53ee369992a804d5810a80e9fa035 --parallel 1 --port 40321" time=2024-09-27T17:08:41.242Z level=INFO source=sched.go:449 msg="loaded runners" count=1 time=2024-09-27T17:08:41.242Z level=INFO source=server.go:587 msg="waiting for llama runner to start responding" time=2024-09-27T17:08:41.242Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server error" INFO [main] build info | build=10 commit="3f6ec33" tid="126518886834176" timestamp=1727456921 INFO [main] system info | n_threads=8 n_threads_batch=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="126518886834176" timestamp=1727456921 total_threads=16 INFO [main] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="40321" tid="126518886834176" timestamp=1727456921 ggml_cuda_init: failed to initialize CUDA: no CUDA-capable device is detected ggml_backend_cuda_init: invalid device 0 time=2024-09-27T17:08:41.492Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server loading model" llama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc (version GGUF V3 (latest)) llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. llama_model_loader: - kv 0: general.architecture str = llama llama_model_loader: - kv 1: general.name str = Downloads llama_model_loader: - kv 2: llama.vocab_size u32 = 128256 llama_model_loader: - kv 3: llama.context_length u32 = 8192 llama_model_loader: - kv 4: llama.embedding_length u32 = 4096 llama_model_loader: - kv 5: llama.block_count u32 = 32 llama_model_loader: - kv 6: llama.feed_forward_length u32 = 14336 llama_model_loader: - kv 7: llama.rope.dimension_count u32 = 128 llama_model_loader: - kv 8: llama.attention.head_count u32 = 32 llama_model_loader: - kv 9: llama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 10: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 11: llama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 12: general.file_type u32 = 15 llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 14: tokenizer.ggml.tokens arr[str,128256] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 15: tokenizer.ggml.scores arr[f32,128256] = [0.000000, 0.000000, 0.000000, 0.0000... llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128001 llama_model_loader: - kv 20: tokenizer.chat_template str = {% set loop_messages = messages %}{% ... llama_model_loader: - kv 21: general.quantization_version u32 = 2 llama_model_loader: - type f32: 65 tensors llama_model_loader: - type q4_K: 193 tensors llama_model_loader: - type q6_K: 33 tensors llm_load_vocab: missing or unrecognized pre-tokenizer type, using: 'default' llm_load_vocab: special tokens cache size = 256 llm_load_vocab: token to piece cache size = 0.8000 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = llama 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 = 8192 llm_load_print_meta: n_embd = 4096 llm_load_print_meta: n_layer = 32 llm_load_print_meta: n_head = 32 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 = 4 llm_load_print_meta: n_embd_k_gqa = 1024 llm_load_print_meta: n_embd_v_gqa = 1024 llm_load_print_meta: f_norm_eps = 0.0e+00 llm_load_print_meta: f_norm_rms_eps = 1.0e-05 llm_load_print_meta: f_clamp_kqv = 0.0e+00 llm_load_print_meta: f_max_alibi_bias = 0.0e+00 llm_load_print_meta: f_logit_scale = 0.0e+00 llm_load_print_meta: n_ff = 14336 llm_load_print_meta: n_expert = 0 llm_load_print_meta: n_expert_used = 0 llm_load_print_meta: causal attn = 1 llm_load_print_meta: pooling type = 0 llm_load_print_meta: rope type = 0 llm_load_print_meta: rope scaling = linear llm_load_print_meta: freq_base_train = 500000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_ctx_orig_yarn = 8192 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 = 8B llm_load_print_meta: model ftype = Q4_K - Medium llm_load_print_meta: model params = 8.03 B llm_load_print_meta: model size = 4.58 GiB (4.89 BPW) llm_load_print_meta: general.name = Downloads llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128001 '<|end_of_text|>' llm_load_print_meta: LF token = 128 'Ä' llm_load_print_meta: EOT token = 128009 '<|eot_id|>' llm_load_print_meta: max token length = 256 llm_load_tensors: ggml ctx size = 0.14 MiB llm_load_tensors: offloading 32 repeating layers to GPU llm_load_tensors: offloaded 32/33 layers to GPU llm_load_tensors: CPU buffer size = 4685.30 MiB llama_new_context_with_model: n_ctx = 4096 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 ggml_cuda_host_malloc: failed to allocate 512.00 MiB of pinned memory: no CUDA-capable device is detected llama_kv_cache_init: CPU KV buffer size = 512.00 MiB llama_new_context_with_model: KV self size = 512.00 MiB, K (f16): 256.00 MiB, V (f16): 256.00 MiB ggml_cuda_host_malloc: failed to allocate 0.50 MiB of pinned memory: no CUDA-capable device is detected llama_new_context_with_model: CPU output buffer size = 0.50 MiB ggml_cuda_host_malloc: failed to allocate 296.01 MiB of pinned memory: no CUDA-capable device is detected llama_new_context_with_model: CUDA_Host compute buffer size = 296.01 MiB llama_new_context_with_model: graph nodes = 1030 llama_new_context_with_model: graph splits = 1 INFO [main] model loaded | tid="126518886834176" timestamp=1727456922 time=2024-09-27T17:08:42.497Z level=INFO source=server.go:626 msg="llama runner started in 1.25 seconds" [GIN] 2024/09/27 - 17:09:02 | 200 | 21.067207572s | 172.17.0.1 | POST "/api/chat" 2024/09/27 17:09:27 routes.go:1153: 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:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/root/.ollama/models 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://*] OLLAMA_SCHED_SPREAD:false OLLAMA_TMPDIR: ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]" time=2024-09-27T17:09:27.643Z level=INFO source=images.go:753 msg="total blobs: 5" time=2024-09-27T17:09:27.643Z level=INFO source=images.go:760 msg="total unused blobs removed: 0" time=2024-09-27T17:09:27.643Z level=INFO source=routes.go:1200 msg="Listening on [::]:11434 (version 0.3.12)" time=2024-09-27T17:09:27.643Z level=INFO source=common.go:49 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11 cuda_v12]" time=2024-09-27T17:09:27.643Z level=INFO source=gpu.go:199 msg="looking for compatible GPUs" time=2024-09-27T17:09:27.992Z level=INFO source=types.go:107 msg="inference compute" id=GPU-5b501d39-1388-7ef8-1dd4-4c618162b80d library=cuda variant=v12 compute=8.6 driver=12.6 name="NVIDIA GeForce RTX 3070 Laptop GPU" total="7.7 GiB" available="6.1 GiB" time=2024-09-27T17:09:32.464Z level=WARN source=sched.go:137 msg="multimodal models don't support parallel requests yet" time=2024-09-27T17:09:32.657Z level=INFO source=server.go:103 msg="system memory" total="31.1 GiB" free="25.0 GiB" free_swap="8.0 GiB" time=2024-09-27T17:09:32.658Z level=INFO source=memory.go:326 msg="offload to cuda" layers.requested=-1 layers.model=33 layers.offload=30 layers.split="" memory.available="[6.1 GiB]" memory.gpu_overhead="0 B" memory.required.full="6.6 GiB" memory.required.partial="5.9 GiB" memory.required.kv="512.0 MiB" memory.required.allocations="[5.9 GiB]" memory.weights.total="4.4 GiB" memory.weights.repeating="4.0 GiB" memory.weights.nonrepeating="411.0 MiB" memory.graph.full="296.0 MiB" memory.graph.partial="677.5 MiB" time=2024-09-27T17:09:32.659Z level=INFO source=server.go:388 msg="starting llama server" cmd="/usr/lib/ollama/runners/cuda_v12/ollama_llama_server --model /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc --ctx-size 4096 --batch-size 512 --embedding --log-disable --n-gpu-layers 30 --mmproj /root/.ollama/models/blobs/sha256-eb569aba7d65cf3da1d0369610eb6869f4a53ee369992a804d5810a80e9fa035 --parallel 1 --port 42495" time=2024-09-27T17:09:32.659Z level=INFO source=sched.go:449 msg="loaded runners" count=1 time=2024-09-27T17:09:32.659Z level=INFO source=server.go:587 msg="waiting for llama runner to start responding" time=2024-09-27T17:09:32.659Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server error" INFO [main] build info | build=10 commit="3f6ec33" tid="140507753418752" timestamp=1727456972 INFO [main] system info | n_threads=8 n_threads_batch=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="140507753418752" timestamp=1727456972 total_threads=16 INFO [main] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="42495" tid="140507753418752" timestamp=1727456972 ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 1 CUDA devices: Device 0: NVIDIA GeForce RTX 3070 Laptop GPU, compute capability 8.6, VMM: yes time=2024-09-27T17:09:32.911Z level=INFO source=server.go:621 msg="waiting for server to become available" status="llm server loading model" llama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from /root/.ollama/models/blobs/sha256-b6e1d703db0da8227fdb7127d8716bbc5049c9bf17ca2bb345be9470d217f3fc (version GGUF V3 (latest)) llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. llama_model_loader: - kv 0: general.architecture str = llama llama_model_loader: - kv 1: general.name str = Downloads llama_model_loader: - kv 2: llama.vocab_size u32 = 128256 llama_model_loader: - kv 3: llama.context_length u32 = 8192 llama_model_loader: - kv 4: llama.embedding_length u32 = 4096 llama_model_loader: - kv 5: llama.block_count u32 = 32 llama_model_loader: - kv 6: llama.feed_forward_length u32 = 14336 llama_model_loader: - kv 7: llama.rope.dimension_count u32 = 128 llama_model_loader: - kv 8: llama.attention.head_count u32 = 32 llama_model_loader: - kv 9: llama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 10: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 11: llama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 12: general.file_type u32 = 15 llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 14: tokenizer.ggml.tokens arr[str,128256] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 15: tokenizer.ggml.scores arr[f32,128256] = [0.000000, 0.000000, 0.000000, 0.0000... llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128001 llama_model_loader: - kv 20: tokenizer.chat_template str = {% set loop_messages = messages %}{% ... llama_model_loader: - kv 21: general.quantization_version u32 = 2 llama_model_loader: - type f32: 65 tensors llama_model_loader: - type q4_K: 193 tensors llama_model_loader: - type q6_K: 33 tensors llm_load_vocab: missing or unrecognized pre-tokenizer type, using: 'default' llm_load_vocab: special tokens cache size = 256 llm_load_vocab: token to piece cache size = 0.8000 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = llama 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 = 8192 llm_load_print_meta: n_embd = 4096 llm_load_print_meta: n_layer = 32 llm_load_print_meta: n_head = 32 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 = 4 llm_load_print_meta: n_embd_k_gqa = 1024 llm_load_print_meta: n_embd_v_gqa = 1024 llm_load_print_meta: f_norm_eps = 0.0e+00 llm_load_print_meta: f_norm_rms_eps = 1.0e-05 llm_load_print_meta: f_clamp_kqv = 0.0e+00 llm_load_print_meta: f_max_alibi_bias = 0.0e+00 llm_load_print_meta: f_logit_scale = 0.0e+00 llm_load_print_meta: n_ff = 14336 llm_load_print_meta: n_expert = 0 llm_load_print_meta: n_expert_used = 0 llm_load_print_meta: causal attn = 1 llm_load_print_meta: pooling type = 0 llm_load_print_meta: rope type = 0 llm_load_print_meta: rope scaling = linear llm_load_print_meta: freq_base_train = 500000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_ctx_orig_yarn = 8192 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 = 8B llm_load_print_meta: model ftype = Q4_K - Medium llm_load_print_meta: model params = 8.03 B llm_load_print_meta: model size = 4.58 GiB (4.89 BPW) llm_load_print_meta: general.name = Downloads llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128001 '<|end_of_text|>' llm_load_print_meta: LF token = 128 'Ä' llm_load_print_meta: EOT token = 128009 '<|eot_id|>' llm_load_print_meta: max token length = 256 llm_load_tensors: ggml ctx size = 0.27 MiB llm_load_tensors: offloading 30 repeating layers to GPU llm_load_tensors: offloaded 30/33 layers to GPU llm_load_tensors: CPU buffer size = 4685.30 MiB llm_load_tensors: CUDA0 buffer size = 3727.50 MiB llama_new_context_with_model: n_ctx = 4096 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_kv_cache_init: CUDA_Host KV buffer size = 32.00 MiB llama_kv_cache_init: CUDA0 KV buffer size = 480.00 MiB llama_new_context_with_model: KV self size = 512.00 MiB, K (f16): 256.00 MiB, V (f16): 256.00 MiB llama_new_context_with_model: CUDA_Host output buffer size = 0.50 MiB llama_new_context_with_model: CUDA0 compute buffer size = 669.48 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 16.01 MiB llama_new_context_with_model: graph nodes = 1030 llama_new_context_with_model: graph splits = 26 INFO [main] model loaded | tid="140507753418752" timestamp=1727456974 time=2024-09-27T17:09:34.418Z level=INFO source=server.go:626 msg="llama runner started in 1.76 seconds" [GIN] 2024/09/27 - 17:09:38 | 200 | 5.737602399s | 172.17.0.1 | POST "/api/chat" `
GiteaMirror commented 2026-04-28 17:39:12 -05:00
Author
Owner
Copy Link

@Qualzz commented on GitHub (Sep 29, 2024):

I would like to add that I'm "sometimes" experiencing the same issue.
ollama ps show 100% GPU, but in reality, the model is running on the CPU ( No VRAM usage, model fully loaded on RAM )
Docker restart is enough to solve the issue. But I still don't know what's causing it. ( ubuntu + ollama on docker )

<!-- gh-comment-id:2381464012 --> @Qualzz commented on GitHub (Sep 29, 2024): I would like to add that I'm "sometimes" experiencing the same issue. ollama ps show 100% GPU, but in reality, the model is running on the CPU ( No VRAM usage, model fully loaded on RAM ) Docker restart is enough to solve the issue. But I still don't know what's causing it. ( ubuntu + ollama on docker )
GiteaMirror commented 2026-04-28 17:39:15 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Sep 30, 2024):

cuda driver library failed to get device context 800

Also observed in https://github.com/ollama/ollama/issues/6928. Currently a restart of the docker container temporarily fixes the problem. If you experience it again, server logs from ollama and system logs from the machine (dmesg, /var/log/gpu-manager.log, /var/log/syslog, /var/log/messages, /var/log/kern.log, etc) may help in resolving the problem.

<!-- gh-comment-id:2381925540 --> @rick-github commented on GitHub (Sep 30, 2024): ``` cuda driver library failed to get device context 800 ``` Also observed in https://github.com/ollama/ollama/issues/6928. Currently a restart of the docker container temporarily fixes the problem. If you experience it again, server logs from ollama and system logs from the machine (dmesg, /var/log/gpu-manager.log, /var/log/syslog, /var/log/messages, /var/log/kern.log, etc) may help in resolving the problem.
GiteaMirror commented 2026-04-28 17:39:16 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Sep 30, 2024):

This does look like a dup of the intermittent error code 800 error in containers. Lets track this via #6928. If you see any "interesting" logs from the kernel/drivers when this 800 error crop up, please share them on that issue.

<!-- gh-comment-id:2383566474 --> @dhiltgen commented on GitHub (Sep 30, 2024): This does look like a dup of the intermittent error code 800 error in containers. Lets track this via #6928. If you see any "interesting" logs from the kernel/drivers when this 800 error crop up, please share them on that issue.
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 docker needs more info nvidia
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#50949
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?