Linux基础——Ollama安装
# Ollama安装
#Deepseek版本及下载
下载最新Deepseek最新版本：https://ollama.com/library/deepseek-r1
#ollama
Model模型（可测试安装）：https://ollama.com/search
参考文档：https://docs.ollama.com/linux
下载文档：https://ollama.com/download/linux
Docker文档：https://docs.ollama.com/docker
Troubleshooting文档：https://docs.ollama.com/troubleshooting
验证是否存在：https://ollama.com/library/deepseek-r1
#宿主机bcoe21.10系统依赖包
https://easysoftware.openeuler.org/zh/search?q=gcc&type=rpmpkg&sort=all
#宿主机bcoe21.10系统在线yum
dnf config-manager --add-repo https://repo.openeuler.org/openEuler-20.03-LTS-SP4/everything/x86_64
dnf clean all && dnf makecache
#安装Deepseek gcc依赖包
dnf install gcc-toolset-10-gcc gcc-toolset-10-libstdc++
# Install
To install Ollama, run the following command:
curl -fsSL https://ollama.com/install.sh | sh
# Manual install
If you are upgrading from a prior version, you should remove the old libraries with sudo rm -rf /usr/lib/ollama first.
Download and extract the package:
curl -fsSL https://ollama.com/download/ollama-linux-amd64.tar.zst | sudo tar x -C /usr
# Start Ollama:
ollama serve
# In another terminal, verify that Ollama is running:
ollama -v
# Adding Ollama as a startup service (recommended)
# Create a user and group for Ollama:
sudo useradd -r -s /bin/false -U -m -d /usr/share/ollama ollama
sudo usermod -a -G ollama $(whoami)
# Create a service file in /etc/systemd/system/ollama.service:
[root@Sealer tmp]# cat /etc/systemd/system/ollama.service
[Unit]
Description=Ollama Service
After=network-online.target
[Service]
ExecStart=/usr/local/bin/ollama serve
User=ollama
Group=ollama
Restart=always
RestartSec=3
Environment="PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/root/bin"
Environment="OLLAMA_HOST=0.0.0.0:11434"
[Install]
WantedBy=default.target
# Then start the service:
sudo systemctl daemon-reload
sudo systemctl enable ollama
# Customizing
# To customize the installation of Ollama, you can edit the systemd service file or the environment variables by running:
sudo systemctl edit ollama
Alternatively, create an override file manually in /etc/systemd/system/ollama.service.d/override.conf:
[Service]
Environment="OLLAMA_DEBUG=1"
# Updating
Update Ollama by running the install script again:
curl -fsSL https://ollama.com/install.sh | sh
Or by re-downloading Ollama:
curl -fsSL https://ollama.com/download/ollama-linux-amd64.tar.zst \
| sudo tar x -C /usr
# Installing specific versions
Use OLLAMA_VERSION environment variable with the install script to install a specific version of Ollama, including pre-releases. You can find the version numbers in the releases page.
For example:
curl -fsSL https://ollama.com/install.sh | OLLAMA_VERSION=0.5.7 sh
# Viewing logs
To view logs of Ollama running as a startup service, run:
journalctl -e -u ollama
# Uninstall
# Remove the ollama service:
sudo systemctl stop ollama
sudo systemctl disable ollama
sudo rm /etc/systemd/system/ollama.service
Remove ollama libraries from your lib directory (either /usr/local/lib, /usr/lib, or /lib):
sudo rm -r $(which ollama | tr 'bin' 'lib')
Remove the ollama binary from your bin directory (either /usr/local/bin, /usr/bin, or /bin):
sudo rm $(which ollama)
Remove the downloaded models and Ollama service user and group:
sudo userdel ollama
sudo groupdel ollama
sudo rm -r /usr/share/ollama
问题1：启动deepseek-r1:14b失败，由于/usr/lib64/libstdc++.so.6没有GLIBCXX_3.4.25
[root@Sealer ~]# ollama run deepseek-r1:14b
Error: 500 Internal Server Error: llama-server process has terminated: exit status 1
# 根因定位（从日志截图）
```
/usr/local/lib/ollama/llama‑server: /usr/lib64/libstdc++.so.6: version `GLIBCXX_3.4.25' not found
```
**不是内存 OOM，是 openEuler 系统 C++ 标准库版本过低，ollama 内置的 libllama.so.0 需要 `GLIBCXX_3.4.25`，当前系统缺少这个符号，llama‑server 直接启动崩溃 exit status 1**。
>
> 你的机器内存 total=15.1GiB，就算库版本修好，14B 模型也跑不动，只能跑 7B。
#排查
[root@Sealer ~]# strings /usr/lib64/libstdc++.so.6 | grep GLIBCXX
GLIBCXX_3.4
GLIBCXX_3.4.1
GLIBCXX_3.4.2
GLIBCXX_3.4.3
GLIBCXX_3.4.4
GLIBCXX_3.4.5
GLIBCXX_3.4.6
GLIBCXX_3.4.7
GLIBCXX_3.4.8
GLIBCXX_3.4.9
GLIBCXX_3.4.10
GLIBCXX_3.4.11
GLIBCXX_3.4.12
GLIBCXX_3.4.13
GLIBCXX_3.4.14
GLIBCXX_3.4.15
GLIBCXX_3.4.16
GLIBCXX_3.4.17
GLIBCXX_3.4.18
GLIBCXX_3.4.19
GLIBCXX_3.4.20
GLIBCXX_3.4.21
GLIBCXX_3.4.22
GLIBCXX_3.4.23
GLIBCXX_3.4.24
GLIBCXX_DEBUG_MESSAGE_LENGTH
[root@Sealer yum.repos.d]# find / -name "gcc"
/opt/openEuler/gcc-toolset-10/root/usr/libexec/gcc
/opt/openEuler/gcc-toolset-10/root/usr/bin/gcc
/opt/openEuler/gcc-toolset-10/root/usr/lib/gcc
/usr/libexec/gcc
/usr/src/kernels/4.19.90-2107.6.0.0192.8.oe1.bclinux.x86_64/include/config/have/gcc
/usr/src/kernels/4.19.90-2107.6.0.0192.8.oe1.bclinux.x86_64/include/config/gcc
/usr/share/bash-completion/completions/gcc
/usr/bin/gcc
/usr/lib/gcc
[root@Sealer yum.repos.d]# /opt/openEuler/gcc-toolset-10/root/usr/bin/gcc -v
Using built-in specs.
COLLECT_GCC=/opt/openEuler/gcc-toolset-10/root/usr/bin/gcc
COLLECT_LTO_WRAPPER=/opt/openEuler/gcc-toolset-10/root/usr/libexec/gcc/x86_64-linux-gnu/10/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../configure --disable-bootstrap --with-isl=/root/rpmbuild/BUILD/gcc-10.3.0/obj-x86_64-linux-gnu/isl-install --enable-languages=c,c++,fortran,lto --prefix=/opt/openEuler/gcc-toolset-10/root/usr --mandir=/opt/openEuler/gcc-toolset-10/root/usr/share/man --infodir=/opt/openEuler/gcc-toolset-10/root/usr/share/info --with-bugurl=https://gitee.com/src-openeuler/gcc/issues --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --enable-libstdcxx-backtrace --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --enable-gnu-indirect-function --with-tune=generic --disable-multilib --with-arch_32=x86-64 --build=x86_64-linux-gnu --program-suffix=
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 10.3.0 (GCC)
[root@Sealer yum.repos.d]# ls /opt/openEuler/gcc-toolset-10/
root
[root@Sealer yum.repos.d]# ls /opt/openEuler/gcc-toolset-10/root/usr/
bin/ lib/ lib64/ libexec/ share/
[root@Sealer yum.repos.d]# ls /opt/openEuler/gcc-toolset-10/root/usr/lib64
libasan.so.6 libasan.so.6.0.0 libgcc_s-10.so.1 libgcc_s.so.1 libgomp.so.1 libgomp.so.1.0.0 libstdc++.so.6 libstdc++.so.6.0.28
[root@Sealer yum.repos.d]# strings /opt/openEuler/gcc-toolset-10/root/usr/lib64/libstdc++.so.6.0.28 | grep GLIBCXX_3.4.25
GLIBCXX_3.4.25
[root@Sealer yum.repos.d]# cp -r /opt/openEuler/gcc-toolset-10/root/usr/lib64 /usr/lib64/gcc10/
[root@Sealer yum.repos.d]# ls /usr/lib64/gcc10
libasan.so.6 libasan.so.6.0.0 libgcc_s-10.so.1 libgcc_s.so.1 libgomp.so.1 libgomp.so.1.0.0 libstdc++.so.6 libstdc++.so.6.0.28
[root@Sealer yum.repos.d]# echo "/usr/lib64/gcc10" > /etc/ld.so.conf.d/gcc10.conf
[root@Sealer yum.repos.d]# ldconfig
[root@Sealer yum.repos.d]# strings /usr/lib64/gcc10/libstdc++.so.6 | grep GLIBCXX_3.4.25
GLIBCXX_3.4.25
[root@Sealer yum.repos.d]#
#启动日志
[root@Sealer ~]# tail -f /var/log/messages
Aug 21 14:27:42 Sealer ollama[8664]: resolve_fused_ops: resolving fused DeepSeek V4 HC support:
Aug 21 14:27:42 Sealer ollama[8664]: resolve_fused_ops: fused DeepSeek V4 HC pre enabled
Aug 21 14:27:42 Sealer ollama[8664]: resolve_fused_ops: fused DeepSeek V4 HC comb enabled
Aug 21 14:27:42 Sealer ollama[8664]: resolve_fused_ops: fused DeepSeek V4 HC post enabled
Aug 21 14:27:42 Sealer ollama[8664]: sched_reserve: CPU compute buffer size = 125.01 MiB
Aug 21 14:27:42 Sealer ollama[8664]: sched_reserve: graph nodes = 1638
Aug 21 14:27:42 Sealer ollama[8664]: sched_reserve: graph splits = 1
Aug 21 14:27:42 Sealer ollama[8664]: sched_reserve: reserve took 34.08 ms, sched copies = 1
Aug 21 14:27:42 Sealer ollama[8664]: cmn init: llama threadpool init, n_threads = 2
Aug 21 14:27:42 Sealer ollama[8664]: cmn common_init_: warming up the model with an empty run - please wait ... (--no-warmup to disable)
Aug 21 14:27:46 Sealer ollama[8664]: srv load_model: initializing, n_slots = 1, n_ctx_slot = 4096, kv_unified = 'false'
Aug 21 14:27:46 Sealer ollama[8664]: spec common_specu: no implementations specified for speculative decoding
Aug 21 14:27:46 Sealer ollama[8664]: slot load_model: id 0 | task -1 | new slot, n_ctx = 4096
Aug 21 14:27:46 Sealer ollama[8664]: srv load_model: prompt cache is enabled, size limit: 8192 MiB
Aug 21 14:27:46 Sealer ollama[8664]: srv load_model: use `--cache-ram 0` to disable the prompt cache
Aug 21 14:27:46 Sealer ollama[8664]: srv load_model: for more info see https://github.com/ggml-org/llama.cpp/pull/16391
Aug 21 14:27:46 Sealer ollama[8664]: srv load_model: context checkpoints enabled, max = 32, min spacing = 8192
Aug 21 14:27:46 Sealer ollama[8664]: srv init: idle slots will be saved to prompt cache upon starting a new task
Aug 21 14:27:46 Sealer ollama[8664]: srv init: init: chat template, example_format: 'You are a helpful assistant<｜User｜>Hello<｜Assistant｜>Hi there<｜end▁of▁sentence｜><｜User｜>How are you?<｜Assistant｜>'
Aug 21 14:27:46 Sealer ollama[8664]: srv init: init: chat template, thinking = 1
Aug 21 14:27:46 Sealer ollama[8664]: srv llama_server: model loaded
Aug 21 14:27:46 Sealer ollama[8664]: srv llama_server: listening on http://127.0.0.1:45771
Aug 21 14:27:46 Sealer ollama[8664]: srv update_slots: all slots are idle
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.199+08:00 level=INFO source=llama_server.go:1362 msg="llama-server started in 90.32 seconds"
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.361+08:00 level=INFO source=images.go:381 msg="template selection" model=registry.ollama.ai/library/deepseek-r1:14b selected=gguf_chat_template renderer="" parser="" go_template="[completion thinking]" chat_template="[tools thinking completion]" harmony=null renderer_parser=null
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.368+08:00 level=INFO source=sched.go:728 msg="loaded runners" count=1
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.369+08:00 level=INFO source=llama_server.go:1295 msg="waiting for llama-server to start responding"
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.371+08:00 level=INFO source=llama_server.go:1362 msg="llama-server started in 90.49 seconds"
Aug 21 14:27:46 Sealer ollama[8664]: [GIN] 2026/08/21 - 14:27:46 | 200 | 1m30s | 127.0.0.1 | POST "/api/generate"
[root@Sealer ~]# journalctl -u ollama -f
-- Logs begin at Fri 2026-08-21 08:51:15 CST. --
Aug 21 14:27:46 Sealer ollama[8664]: srv init: init: chat template, thinking = 1
Aug 21 14:27:46 Sealer ollama[8664]: srv llama_server: model loaded
Aug 21 14:27:46 Sealer ollama[8664]: srv llama_server: listening on http://127.0.0.1:45771
Aug 21 14:27:46 Sealer ollama[8664]: srv update_slots: all slots are idle
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.199+08:00 level=INFO source=llama_server.go:1362 msg="llama-server started in 90.32 seconds"
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.361+08:00 level=INFO source=images.go:381 msg="template selection" model=registry.ollama.ai/library/deepseek-r1:14b selected=gguf_chat_template renderer="" parser="" go_template="[completion thinking]" chat_template="[tools thinking completion]" harmony=null renderer_parser=null
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.368+08:00 level=INFO source=sched.go:728 msg="loaded runners" count=1
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.369+08:00 level=INFO source=llama_server.go:1295 msg="waiting for llama-server to start responding"
Aug 21 14:27:46 Sealer ollama[8664]: time=2026-08-21T14:27:46.371+08:00 level=INFO source=llama_server.go:1362 msg="llama-server started in 90.49 seconds"
Aug 21 14:27:46 Sealer ollama[8664]: [GIN] 2026/08/21 - 14:27:46 | 200 | 1m30s | 127.0.0.1 | POST "/api/generate"
Aug 21 14:30:20 Sealer ollama[8664]: srv server_strea: conv_id= (empty=1)
Aug 21 14:30:20 Sealer ollama[8664]: srv operator(): chat format: peg-native
Aug 21 14:30:21 Sealer ollama[8664]: slot get_availabl: id 0 | task -1 | - skipping, slot is empty
Aug 21 14:30:21 Sealer ollama[8664]: slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = -1
Aug 21 14:30:21 Sealer ollama[8664]: srv get_availabl: updating prompt cache
Aug 21 14:30:21 Sealer ollama[8664]: srv load: - looking for better prompt, base f_keep = -1.000, f_sim = 0.000
Aug 21 14:30:21 Sealer ollama[8664]: srv update: - cache state: 0 prompts, 0.000 MiB (limits: 8192.000 MiB, 4096 tokens, 8589934592 est)
Aug 21 14:30:21 Sealer ollama[8664]: srv get_availabl: prompt cache update took 0.01 ms
Aug 21 14:30:21 Sealer ollama[8664]: slot launch_slot_: id 0 | task -1 | sampler chain: logits -> ?penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist
Aug 21 14:30:21 Sealer ollama[8664]: slot launch_slot_: id 0 | task -1 | sampler params:
Aug 21 14:30:21 Sealer ollama[8664]: repeat_last_n = 64, repeat_penalty = 1.000, frequency_penalty = 0.000, presence_penalty = 0.000
Aug 21 14:30:21 Sealer ollama[8664]: dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64
Aug 21 14:30:21 Sealer ollama[8664]: top_k = 40, top_p = 0.900, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.800
Aug 21 14:30:21 Sealer ollama[8664]: mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900
Aug 21 14:30:21 Sealer ollama[8664]: slot launch_slot_: id 0 | task 0 | processing task, is_child = 0
Aug 21 14:30:21 Sealer ollama[8664]: slot operator(): id 0 | task 0 | new prompt, n_ctx_slot = 4096, n_keep = 4, task.n_tokens = 17
Aug 21 14:30:21 Sealer ollama[8664]: slot operator(): id 0 | task 0 | cached n_tokens = 0, memory_seq_rm [0, end)
Aug 21 14:30:21 Sealer ollama[8664]: slot init_sampler: id 0 | task 0 | init sampler, took 0.01 ms, tokens: text = 17, total = 17
Aug 21 14:31:48 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 100, tg = 1.22 t/s, tg_3s = 1.23 t/s
Aug 21 14:31:51 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 104, tg = 1.22 t/s, tg_3s = 1.25 t/s
Aug 21 14:31:54 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 108, tg = 1.22 t/s, tg_3s = 1.23 t/s
Aug 21 14:31:57 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 112, tg = 1.22 t/s, tg_3s = 1.27 t/s
Aug 21 14:32:00 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 116, tg = 1.22 t/s, tg_3s = 1.27 t/s
Aug 21 14:32:04 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 120, tg = 1.22 t/s, tg_3s = 1.11 t/s
Aug 21 14:32:08 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 124, tg = 1.22 t/s, tg_3s = 1.12 t/s
Aug 21 14:32:11 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 128, tg = 1.22 t/s, tg_3s = 1.19 t/s
Aug 21 14:32:14 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 132, tg = 1.22 t/s, tg_3s = 1.22 t/s
Aug 21 14:32:18 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 136, tg = 1.21 t/s, tg_3s = 1.16 t/s
Aug 21 14:32:21 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 140, tg = 1.21 t/s, tg_3s = 1.21 t/s
Aug 21 14:32:24 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 144, tg = 1.21 t/s, tg_3s = 1.20 t/s
Aug 21 14:32:28 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 148, tg = 1.21 t/s, tg_3s = 1.21 t/s
Aug 21 14:32:31 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 152, tg = 1.22 t/s, tg_3s = 1.29 t/s
Aug 21 14:32:34 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 156, tg = 1.22 t/s, tg_3s = 1.29 t/s
Aug 21 14:32:37 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 160, tg = 1.22 t/s, tg_3s = 1.21 t/s
Aug 21 14:32:40 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 164, tg = 1.22 t/s, tg_3s = 1.27 t/s
Aug 21 14:32:43 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 168, tg = 1.22 t/s, tg_3s = 1.27 t/s
Aug 21 14:32:46 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 172, tg = 1.22 t/s, tg_3s = 1.28 t/s
Aug 21 14:32:50 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 176, tg = 1.22 t/s, tg_3s = 1.28 t/s
Aug 21 14:32:53 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 180, tg = 1.22 t/s, tg_3s = 1.28 t/s
Aug 21 14:32:56 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 184, tg = 1.22 t/s, tg_3s = 1.28 t/s
Aug 21 14:32:59 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 188, tg = 1.23 t/s, tg_3s = 1.29 t/s
Aug 21 14:33:02 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 192, tg = 1.23 t/s, tg_3s = 1.23 t/s
Aug 21 14:33:06 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 196, tg = 1.22 t/s, tg_3s = 1.18 t/s
Aug 21 14:33:09 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 200, tg = 1.23 t/s, tg_3s = 1.24 t/s
Aug 21 14:33:12 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 204, tg = 1.23 t/s, tg_3s = 1.29 t/s
Aug 21 14:33:15 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 208, tg = 1.23 t/s, tg_3s = 1.21 t/s
Aug 21 14:33:18 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 212, tg = 1.23 t/s, tg_3s = 1.28 t/s
Aug 21 14:33:21 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 216, tg = 1.23 t/s, tg_3s = 1.27 t/s
Aug 21 14:33:25 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 220, tg = 1.23 t/s, tg_3s = 1.27 t/s
Aug 21 14:33:28 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 224, tg = 1.23 t/s, tg_3s = 1.27 t/s
Aug 21 14:33:31 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 228, tg = 1.23 t/s, tg_3s = 1.28 t/s
Aug 21 14:33:34 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 232, tg = 1.23 t/s, tg_3s = 1.27 t/s
Aug 21 14:33:37 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 236, tg = 1.23 t/s, tg_3s = 1.25 t/s
Aug 21 14:33:40 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 240, tg = 1.23 t/s, tg_3s = 1.26 t/s
Aug 21 14:33:44 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 244, tg = 1.23 t/s, tg_3s = 1.22 t/s
Aug 21 14:33:47 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 248, tg = 1.23 t/s, tg_3s = 1.23 t/s
Aug 21 14:33:50 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 252, tg = 1.23 t/s, tg_3s = 1.24 t/s
Aug 21 14:33:53 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 256, tg = 1.23 t/s, tg_3s = 1.24 t/s
Aug 21 14:33:57 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 260, tg = 1.23 t/s, tg_3s = 1.25 t/s
Aug 21 14:34:00 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 264, tg = 1.23 t/s, tg_3s = 1.24 t/s
Aug 21 14:34:03 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 268, tg = 1.23 t/s, tg_3s = 1.19 t/s
Aug 21 14:34:07 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 272, tg = 1.23 t/s, tg_3s = 1.22 t/s
Aug 21 14:34:10 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 276, tg = 1.23 t/s, tg_3s = 1.24 t/s
Aug 21 14:34:13 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 280, tg = 1.23 t/s, tg_3s = 1.23 t/s
Aug 21 14:34:16 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 284, tg = 1.23 t/s, tg_3s = 1.24 t/s
Aug 21 14:34:19 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 288, tg = 1.23 t/s, tg_3s = 1.24 t/s
Aug 21 14:34:23 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 292, tg = 1.23 t/s, tg_3s = 1.25 t/s
Aug 21 14:34:26 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 296, tg = 1.23 t/s, tg_3s = 1.22 t/s
Aug 21 14:34:29 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 300, tg = 1.23 t/s, tg_3s = 1.22 t/s
Aug 21 14:34:32 Sealer ollama[8664]: slot print_timing: id 0 | task 0 | n_gen = 304, tg = 1.23 t/s, tg_3s = 1.21 t/s
Aug 21 14:34:33 Sealer ollama[8664]: [GIN] 2026/08/21 - 14:34:33 | 200 | 4m12s | 127.0.0.1 | POST "/api/chat"
Aug 21 14:34:33 Sealer ollama[8664]: srv stop: cancel task, id_task = 0
Aug 21 14:34:34 Sealer ollama[8664]: slot release: id 0 | task 0 | stop processing: n_tokens = 322, truncated = 0
Aug 21 14:34:34 Sealer ollama[8664]: srv update_slots: all slots are idle
Aug 21 14:40:24 Sealer ollama[8664]: time=2026-08-21T14:40:24.613+08:00 level=INFO source=sched.go:1147 msg="disabling mmap for llama-server load by default" model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e reason=cpu
Aug 21 14:40:24 Sealer ollama[8664]: time=2026-08-21T14:40:24.613+08:00 level=INFO source=server.go:109 msg="using llama-server for model" model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e
Aug 21 14:40:24 Sealer ollama[8664]: time=2026-08-21T14:40:24.636+08:00 level=INFO source=llama_server.go:433 msg="starting llama-server" cmd="/usr/local/lib/ollama/llama-server --model /usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e --port 33837 --host 127.0.0.1 --no-webui --offline -c 4096 -np 1 --log-verbosity 4 --no-log-prefix --no-log-timestamps --load-mode none --flash-attn auto -b 512 -ub 512 --context-shift --keep 4"
Aug 21 14:40:24 Sealer ollama[8664]: time=2026-08-21T14:40:24.638+08:00 level=INFO source=sched.go:613 msg="system memory" total="15.1 GiB" free="14.3 GiB" free_swap="3.0 GiB"
Aug 21 14:40:24 Sealer ollama[8664]: time=2026-08-21T14:40:24.638+08:00 level=INFO source=llama_server.go:1048 msg="loading model via llama-server" model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e
Aug 21 14:40:24 Sealer ollama[8664]: time=2026-08-21T14:40:24.639+08:00 level=INFO source=llama_server.go:1295 msg="waiting for llama-server to start responding"
Aug 21 14:40:24 Sealer ollama[8664]: time=2026-08-21T14:40:24.641+08:00 level=INFO source=llama_server.go:1350 msg="waiting for llama-server to become available" status="llm server not responding"
Aug 21 14:40:24 Sealer ollama[8664]: cmn common_param: common_params_print_info: build 1 (9d77fa172) with GNU 13.3.1 for Linux x86_64
Aug 21 14:40:24 Sealer ollama[8664]: cmn common_param: common_params_print_info: verbosity = 4 (adjust with the `-lv N` CLI arg)
Aug 21 14:40:24 Sealer ollama[8664]: cmn common_param: device_info:
Aug 21 14:40:24 Sealer ollama[8664]: cmn common_param: - CPU : AMD Ryzen 7 5700U with Radeon Graphics (15499 MiB, 15499 MiB free)
Aug 21 14:40:24 Sealer ollama[8664]: cmn common_param: system_info: n_threads = 2 (n_threads_batch = 2) / 2 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | LLAMAFILE = 1 | REPACK = 1 |
Aug 21 14:40:24 Sealer ollama[8664]: srv init: using 5 threads for HTTP server
Aug 21 14:40:24 Sealer ollama[8664]: srv init: The UI is disabled
Aug 21 14:40:24 Sealer ollama[8664]: srv init: Use --ui/--no-ui (or deprecated --webui/--no-webui) to enable/disable
Aug 21 14:40:24 Sealer ollama[8664]: srv llama_server: -----------------
Aug 21 14:40:24 Sealer ollama[8664]: srv llama_server: CORS is set to allow all origins ('*') and no API key is set
Aug 21 14:40:24 Sealer ollama[8664]: srv llama_server: this can be a security risk (cross-origin attacks)
Aug 21 14:40:24 Sealer ollama[8664]: srv llama_server: more info: https://github.com/ggml-org/llama.cpp/pull/25655
Aug 21 14:40:24 Sealer ollama[8664]: srv llama_server: -----------------
Aug 21 14:40:24 Sealer ollama[8664]: srv start: binding port with default address family
Aug 21 14:40:24 Sealer ollama[8664]: srv load_model: loading model '/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e'
Aug 21 14:40:24 Sealer ollama[8664]: srv load_model: local path '/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e'
Aug 21 14:40:24 Sealer ollama[8664]: cmn common_init_: fitting params to device memory ...
Aug 21 14:40:24 Sealer ollama[8664]: cmn common_init_: (for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on)
Aug 21 14:40:24 Sealer ollama[8664]: common_params_fit_impl: getting device memory data for initial parameters:
Aug 21 14:40:24 Sealer ollama[8664]: time=2026-08-21T14:40:24.895+08:00 level=INFO source=llama_server.go:1350 msg="waiting for llama-server to become available" status="llm server loading model"
Aug 21 14:40:25 Sealer ollama[8664]: common_memory_breakdown_print: | memory breakdown [MiB] | total free self model context compute unaccounted |
Aug 21 14:40:25 Sealer ollama[8664]: common_memory_breakdown_print: | - Host | 3350 = 2457 + 768 + 125 |
Aug 21 14:40:25 Sealer ollama[8664]: common_memory_breakdown_print: | - CPU_REPACK | 6108 = 6108 + 0 + 0 |
Aug 21 14:40:25 Sealer ollama[8664]: common_params_fit_impl: projected to use 3350 MiB of host memory vs. 15499 MiB of total host memory
Aug 21 14:40:25 Sealer ollama[8664]: common_params_fit_impl: will leave 12149 >= 1024 MiB of system memory, no changes needed
Aug 21 14:40:25 Sealer ollama[8664]: common_fit_params: successfully fit params to free device memory
Aug 21 14:40:25 Sealer ollama[8664]: common_fit_params: fitting params to free memory took 0.53 seconds
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: loaded meta data with 26 key-value pairs and 579 tensors from /usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e (version GGUF V3 (latest))
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 0: general.architecture str = qwen2
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 1: general.type str = model
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 2: general.name str = DeepSeek R1 Distill Qwen 14B
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 3: general.basename str = DeepSeek-R1-Distill-Qwen
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 4: general.size_label str = 14B
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 5: qwen2.block_count u32 = 48
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 6: qwen2.context_length u32 = 131072
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 7: qwen2.embedding_length u32 = 5120
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 8: qwen2.feed_forward_length u32 = 13824
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 9: qwen2.attention.head_count u32 = 40
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 10: qwen2.attention.head_count_kv u32 = 8
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 11: qwen2.rope.freq_base f32 = 1000000.000000
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 12: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000010
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 13: general.file_type u32 = 15
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 14: tokenizer.ggml.model str = gpt2
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 15: tokenizer.ggml.pre str = qwen2
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 16: tokenizer.ggml.tokens arr[str,152064] = ["!", "\"", "#", "$", "%", "&", "'", ...
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 17: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 18: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 19: tokenizer.ggml.bos_token_id u32 = 151646
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 151643
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 151643
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 22: tokenizer.ggml.add_bos_token bool = true
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 23: tokenizer.ggml.add_eos_token bool = false
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 24: tokenizer.chat_template str = {% if not add_generation_prompt is de...
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - kv 25: general.quantization_version u32 = 2
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - type f32: 241 tensors
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - type q4_K: 289 tensors
Aug 21 14:40:25 Sealer ollama[8664]: llama_model_loader: - type q6_K: 49 tensors
Aug 21 14:40:25 Sealer ollama[8664]: print_info: file format = GGUF V3 (latest)
Aug 21 14:40:25 Sealer ollama[8664]: print_info: file type = Q4_K - Medium
Aug 21 14:40:25 Sealer ollama[8664]: print_info: file size = 8.37 GiB (4.87 BPW)
Aug 21 14:40:25 Sealer ollama[8664]: load: 0 unused tokens
Aug 21 14:40:25 Sealer ollama[8664]: load: control-looking token: 128247 '</s>' was not control-type; this is probably a bug in the model. its type will be overridden
Aug 21 14:40:25 Sealer ollama[8664]: load: printing all EOG tokens:
Aug 21 14:40:25 Sealer ollama[8664]: load: - 128247 ('</s>')
Aug 21 14:40:25 Sealer ollama[8664]: load: - 151643 ('<｜end▁of▁sentence｜>')
Aug 21 14:40:25 Sealer ollama[8664]: load: - 151662 ('<|fim_pad|>')
Aug 21 14:40:25 Sealer ollama[8664]: load: - 151663 ('<|repo_name|>')
Aug 21 14:40:25 Sealer ollama[8664]: load: - 151664 ('<|file_sep|>')
Aug 21 14:40:25 Sealer ollama[8664]: load: special tokens cache size = 23
Aug 21 14:40:25 Sealer ollama[8664]: load: token to piece cache size = 0.9310 MB
Aug 21 14:40:25 Sealer ollama[8664]: print_info: arch = qwen2
Aug 21 14:40:25 Sealer ollama[8664]: print_info: vocab_only = 0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: no_alloc = 0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_ctx_train = 131072
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_embd_inp = 5120
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_embd = 5120
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_embd_out = 5120
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_layer = 48
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_layer_all = 48
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_head = 40
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_head_kv = 8
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_rot = 128
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_swa = 0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: is_swa_any = 0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_embd_head_k = 128
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_embd_head_v = 128
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_gqa = 5
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_embd_k_gqa = 1024
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_embd_v_gqa = 1024
Aug 21 14:40:25 Sealer ollama[8664]: print_info: f_norm_eps = 0.0e+00
Aug 21 14:40:25 Sealer ollama[8664]: print_info: f_norm_rms_eps = 1.0e-05
Aug 21 14:40:25 Sealer ollama[8664]: print_info: f_clamp_kqv = 0.0e+00
Aug 21 14:40:25 Sealer ollama[8664]: print_info: f_max_alibi_bias = 0.0e+00
Aug 21 14:40:25 Sealer ollama[8664]: print_info: f_logit_scale = 0.0e+00
Aug 21 14:40:25 Sealer ollama[8664]: print_info: f_attn_scale = 0.0e+00
Aug 21 14:40:25 Sealer ollama[8664]: print_info: f_attn_value_scale = 0.0000
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_ff = 13824
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_expert = 0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_expert_used = 0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_expert_groups = 0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_group_used = 0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: causal attn = 1
Aug 21 14:40:25 Sealer ollama[8664]: print_info: pooling type = -1
Aug 21 14:40:25 Sealer ollama[8664]: print_info: rope type = 2
Aug 21 14:40:25 Sealer ollama[8664]: print_info: rope scaling = linear
Aug 21 14:40:25 Sealer ollama[8664]: print_info: freq_base_train = 1000000.0
Aug 21 14:40:25 Sealer ollama[8664]: print_info: freq_scale_train = 1
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_ctx_orig_yarn = 131072
Aug 21 14:40:25 Sealer ollama[8664]: print_info: rope_yarn_log_mul = 0.0000
Aug 21 14:40:25 Sealer ollama[8664]: print_info: rope_finetuned = unknown
Aug 21 14:40:25 Sealer ollama[8664]: print_info: model type = 14B
Aug 21 14:40:25 Sealer ollama[8664]: print_info: model params = 14.77 B
Aug 21 14:40:25 Sealer ollama[8664]: print_info: general.name = DeepSeek R1 Distill Qwen 14B
Aug 21 14:40:25 Sealer ollama[8664]: print_info: vocab type = BPE
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_vocab = 152064
Aug 21 14:40:25 Sealer ollama[8664]: print_info: n_merges = 151387
Aug 21 14:40:25 Sealer ollama[8664]: print_info: BOS token = 151646 '<｜begin▁of▁sentence｜>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: EOS token = 151643 '<｜end▁of▁sentence｜>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: EOT token = 151643 '<｜end▁of▁sentence｜>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: PAD token = 151643 '<｜end▁of▁sentence｜>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: LF token = 198 'Ċ'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: FIM PRE token = 151659 '<|fim_prefix|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: FIM SUF token = 151661 '<|fim_suffix|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: FIM MID token = 151660 '<|fim_middle|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: FIM PAD token = 151662 '<|fim_pad|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: FIM REP token = 151663 '<|repo_name|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: FIM SEP token = 151664 '<|file_sep|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: EOG token = 128247 '</s>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: EOG token = 151643 '<｜end▁of▁sentence｜>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: EOG token = 151662 '<|fim_pad|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: EOG token = 151663 '<|repo_name|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: EOG token = 151664 '<|file_sep|>'
Aug 21 14:40:25 Sealer ollama[8664]: print_info: max token length = 256
Aug 21 14:40:25 Sealer ollama[8664]: load_tensors: loading model tensors, this can take a while... (load_mode = none)
Aug 21 14:40:25 Sealer ollama[8664]: load_tensors: CPU model buffer size = 2457.29 MiB
Aug 21 14:40:25 Sealer ollama[8664]: load_tensors: CPU_REPACK model buffer size = 6108.75 MiB
Aug 21 14:40:45 Sealer ollama[8664]: cmn common_init_: added </s> logit bias = -inf
Aug 21 14:40:45 Sealer ollama[8664]: cmn common_init_: added <｜end▁of▁sentence｜> logit bias = -inf
Aug 21 14:40:45 Sealer ollama[8664]: cmn common_init_: added <|fim_pad|> logit bias = -inf
Aug 21 14:40:45 Sealer ollama[8664]: cmn common_init_: added <|repo_name|> logit bias = -inf
Aug 21 14:40:45 Sealer ollama[8664]: cmn common_init_: added <|file_sep|> logit bias = -inf
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: constructing llama_context
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_seq_max = 1
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_ctx = 4096
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_ctx_seq = 4096
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_batch = 512
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_ubatch = 512
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: causal_attn = 1
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: flash_attn = auto
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: kv_unified = false
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: freq_base = 1000000.0
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: freq_scale = 1
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_rs_seq = 0
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_outputs_max = 1
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_outputs_max_per_seq = 1
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: n_ctx_seq (4096) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
Aug 21 14:40:45 Sealer ollama[8664]: llama_context: CPU output buffer size = 0.58 MiB
Aug 21 14:40:45 Sealer ollama[8664]: llama_kv_cache: CPU KV buffer size = 768.00 MiB
Aug 21 14:40:45 Sealer ollama[8664]: llama_kv_cache: size = 768.00 MiB ( 4096 cells, 48 layers, 1/1 seqs), K (f16): 384.00 MiB, V (f16): 384.00 MiB
Aug 21 14:40:45 Sealer ollama[8664]: llama_kv_cache: attn_rot_k = 0, n_embd_head_k_all = 128
Aug 21 14:40:45 Sealer ollama[8664]: llama_kv_cache: attn_rot_v = 0, n_embd_head_k_all = 128
Aug 21 14:40:45 Sealer ollama[8664]: sched_reserve: reserving ...
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: Flash Attention enabled
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: resolving fused Gated Delta Net support:
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: fused Gated Delta Net (autoregressive) enabled
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: fused Gated Delta Net (chunked) enabled
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: resolving fused Lightning Indexer support:
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: Lightning Indexer enabled
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: resolving fused DeepSeek V4 HC support:
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: fused DeepSeek V4 HC pre enabled
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: fused DeepSeek V4 HC comb enabled
Aug 21 14:40:45 Sealer ollama[8664]: resolve_fused_ops: fused DeepSeek V4 HC post enabled
Aug 21 14:40:45 Sealer ollama[8664]: sched_reserve: CPU compute buffer size = 125.01 MiB
Aug 21 14:40:45 Sealer ollama[8664]: sched_reserve: graph nodes = 1638
Aug 21 14:40:45 Sealer ollama[8664]: sched_reserve: graph splits = 1
Aug 21 14:40:45 Sealer ollama[8664]: sched_reserve: reserve took 25.42 ms, sched copies = 1
Aug 21 14:40:45 Sealer ollama[8664]: cmn init: llama threadpool init, n_threads = 2
Aug 21 14:40:45 Sealer ollama[8664]: cmn common_init_: warming up the model with an empty run - please wait ... (--no-warmup to disable)
Aug 21 14:40:49 Sealer ollama[8664]: srv load_model: initializing, n_slots = 1, n_ctx_slot = 4096, kv_unified = 'false'
Aug 21 14:40:49 Sealer ollama[8664]: spec common_specu: no implementations specified for speculative decoding
Aug 21 14:40:49 Sealer ollama[8664]: slot load_model: id 0 | task -1 | new slot, n_ctx = 4096
Aug 21 14:40:49 Sealer ollama[8664]: srv load_model: prompt cache is enabled, size limit: 8192 MiB
Aug 21 14:40:49 Sealer ollama[8664]: srv load_model: use `--cache-ram 0` to disable the prompt cache
Aug 21 14:40:49 Sealer ollama[8664]: srv load_model: for more info see https://github.com/ggml-org/llama.cpp/pull/16391
Aug 21 14:40:49 Sealer ollama[8664]: srv load_model: context checkpoints enabled, max = 32, min spacing = 8192
Aug 21 14:40:49 Sealer ollama[8664]: srv init: idle slots will be saved to prompt cache upon starting a new task
Aug 21 14:40:49 Sealer ollama[8664]: srv init: init: chat template, example_format: 'You are a helpful assistant<｜User｜>Hello<｜Assistant｜>Hi there<｜end▁of▁sentence｜><｜User｜>How are you?<｜Assistant｜>'
Aug 21 14:40:49 Sealer ollama[8664]: srv init: init: chat template, thinking = 1
Aug 21 14:40:49 Sealer ollama[8664]: srv llama_server: model loaded
Aug 21 14:40:49 Sealer ollama[8664]: srv llama_server: listening on http://127.0.0.1:33837
Aug 21 14:40:49 Sealer ollama[8664]: srv update_slots: all slots are idle
Aug 21 14:40:49 Sealer ollama[8664]: time=2026-08-21T14:40:49.578+08:00 level=INFO source=llama_server.go:1362 msg="llama-server started in 24.94 seconds"
Aug 21 14:40:49 Sealer ollama[8664]: time=2026-08-21T14:40:49.819+08:00 level=INFO source=images.go:381 msg="template selection" model=registry.ollama.ai/library/deepseek-r1:14b selected=gguf_chat_template renderer="" parser="" go_template="[completion thinking]" chat_template="[tools thinking completion]" harmony=null renderer_parser=null
Aug 21 14:40:49 Sealer ollama[8664]: time=2026-08-21T14:40:49.824+08:00 level=INFO source=sched.go:728 msg="loaded runners" count=1
Aug 21 14:40:49 Sealer ollama[8664]: time=2026-08-21T14:40:49.826+08:00 level=INFO source=llama_server.go:1295 msg="waiting for llama-server to start responding"
Aug 21 14:40:49 Sealer ollama[8664]: time=2026-08-21T14:40:49.832+08:00 level=INFO source=llama_server.go:1362 msg="llama-server started in 25.19 seconds"
Aug 21 14:40:49 Sealer ollama[8664]: srv server_strea: conv_id= (empty=1)
Aug 21 14:40:49 Sealer ollama[8664]: srv operator(): chat format: peg-native
Aug 21 14:40:49 Sealer ollama[8664]: slot get_availabl: id 0 | task -1 | - skipping, slot is empty
Aug 21 14:40:49 Sealer ollama[8664]: slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = -1
Aug 21 14:40:49 Sealer ollama[8664]: srv get_availabl: updating prompt cache
Aug 21 14:40:49 Sealer ollama[8664]: srv load: - looking for better prompt, base f_keep = -1.000, f_sim = 0.000
Aug 21 14:40:49 Sealer ollama[8664]: srv update: - cache state: 0 prompts, 0.000 MiB (limits: 8192.000 MiB, 4096 tokens, 8589934592 est)
Aug 21 14:40:49 Sealer ollama[8664]: srv get_availabl: prompt cache update took 0.01 ms
Aug 21 14:40:49 Sealer ollama[8664]: slot launch_slot_: id 0 | task -1 | sampler chain: logits -> ?penalties -> ?dry -> ?top-n-sigma -> top-k -> ?typical -> top-p -> ?min-p -> ?xtc -> temp-ext -> dist
Aug 21 14:40:49 Sealer ollama[8664]: slot launch_slot_: id 0 | task -1 | sampler params:
Aug 21 14:40:49 Sealer ollama[8664]: repeat_last_n = 64, repeat_penalty = 1.000, frequency_penalty = 0.000, presence_penalty = 0.000
Aug 21 14:40:49 Sealer ollama[8664]: dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = 64
Aug 21 14:40:49 Sealer ollama[8664]: top_k = 40, top_p = 0.900, min_p = 0.000, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, top_n_sigma = -1.000, temp = 0.800
Aug 21 14:40:49 Sealer ollama[8664]: mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000, adaptive_target = -1.000, adaptive_decay = 0.900
Aug 21 14:40:49 Sealer ollama[8664]: slot launch_slot_: id 0 | task 0 | processing task, is_child = 0
Aug 21 14:40:49 Sealer ollama[8664]: slot operator(): id 0 | task 0 | new prompt, n_ctx_slot = 4096, n_keep = 4, task.n_tokens = 19
Aug 21 14:40:49 Sealer ollama[8664]: slot operator(): id 0 | task 0 | cached n_tokens = 0, memory_seq_rm [0, end)
Aug 21 14:40:49 Sealer ollama[8664]: slot init_sampler: id 0 | task 0 | init sampler, took 0.01 ms, tokens: text = 19, total = 19
Aug 21 14:41:45 Sealer ollama[8664]: [GIN] 2026/08/21 - 14:41:45 | 200 | 1m20s | 127.0.0.1 | POST "/api/chat"
Aug 21 14:41:45 Sealer ollama[8664]: srv stop: cancel task, id_task = 0
Aug 21 14:41:46 Sealer ollama[8664]: slot release: id 0 | task 0 | stop processing: n_tokens = 77, truncated = 0
Aug 21 14:41:46 Sealer ollama[8664]: srv update_slots: all slots are idle
#For CPU only 安装open-webui （暂不选择该方法）
docker run -d -p 3000:8080 -v ollama:/root/.ollama -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:ollama
#宿主机直接安装ollama工具，选择合适AI model
## 访问ollama官网，选择model
https://ollama.com/library/deepseek-r1
### 下列deepseek-r1所有model
deepseek-r1:latest
5.2GB
deepseek-r1:1.5b
1.1GB
deepseek-r1:7b
4.7GB
deepseek-r1:8b
5.2GB
deepseek-r1:14b
9.0GB
deepseek-r1:32b
20GB
deepseek-r1:70b
43GB
deepseek-r1:671b
404GB
deepseek-r1:1.5b-qwen-distill-q4_K_M
1.1GB
deepseek-r1:1.5b-qwen-distill-q8_0
1.9GB
deepseek-r1:1.5b-qwen-distill-fp16
3.6GB
deepseek-r1:7b-qwen-distill-q4_K_M
4.7GB
deepseek-r1:7b-qwen-distill-q8_0
8.1GB
deepseek-r1:7b-qwen-distill-fp16
15GB
deepseek-r1:8b-0528-qwen3-q4_K_M
5.2GB
deepseek-r1:8b-0528-qwen3-q8_0
8.9GB
deepseek-r1:8b-0528-qwen3-fp16
16GB
deepseek-r1:8b-llama-distill-q4_K_M
4.9GB
deepseek-r1:8b-llama-distill-q8_0
8.5GB
deepseek-r1:8b-llama-distill-fp16
16GB
deepseek-r1:14b-qwen-distill-q4_K_M
9.0GB
deepseek-r1:14b-qwen-distill-q8_0
16GB
deepseek-r1:14b-qwen-distill-fp16
30GB
deepseek-r1:32b-qwen-distill-q4_K_M
20GB
deepseek-r1:32b-qwen-distill-q8_0
35GB
deepseek-r1:32b-qwen-distill-fp16
66GB
deepseek-r1:70b-llama-distill-q4_K_M
43GB
deepseek-r1:70b-llama-distill-q8_0
75GB
deepseek-r1:70b-llama-distill-fp16
141GB
deepseek-r1:671b-0528-q4_K_M
404GB
deepseek-r1:671b-0528-q8_0
713GB
deepseek-r1:671b-0528-fp16
deepseek-r1:671b-q4_K_M
404GB
deepseek-r1:671b-q8_0
713GB
deepseek-r1:671b-fp16
[root@Sealer tmp]# egrep "deep|GB|TB" 1.txt
deepseek-r1:latest
5.2GB
deepseek-r1:1.5b
1.1GB
deepseek-r1:7b
4.7GB
deepseek-r1:8b
5.2GB
deepseek-r1:14b
9.0GB
deepseek-r1:32b
20GB
deepseek-r1:70b
43GB
deepseek-r1:671b
404GB
deepseek-r1:1.5b-qwen-distill-q4_K_M
1.1GB
deepseek-r1:1.5b-qwen-distill-q8_0
1.9GB
deepseek-r1:1.5b-qwen-distill-fp16
3.6GB
deepseek-r1:7b-qwen-distill-q4_K_M
4.7GB
deepseek-r1:7b-qwen-distill-q8_0
8.1GB
deepseek-r1:7b-qwen-distill-fp16
15GB
deepseek-r1:8b-0528-qwen3-q4_K_M
5.2GB
deepseek-r1:8b-0528-qwen3-q8_0
8.9GB
deepseek-r1:8b-0528-qwen3-fp16
16GB
deepseek-r1:8b-llama-distill-q4_K_M
4.9GB
deepseek-r1:8b-llama-distill-q8_0
8.5GB
deepseek-r1:8b-llama-distill-fp16
16GB
deepseek-r1:14b-qwen-distill-q4_K_M
9.0GB
deepseek-r1:14b-qwen-distill-q8_0
16GB
deepseek-r1:14b-qwen-distill-fp16
30GB
deepseek-r1:32b-qwen-distill-q4_K_M
20GB
deepseek-r1:32b-qwen-distill-q8_0
35GB
deepseek-r1:32b-qwen-distill-fp16
66GB
deepseek-r1:70b-llama-distill-q4_K_M
43GB
deepseek-r1:70b-llama-distill-q8_0
75GB
deepseek-r1:70b-llama-distill-fp16
141GB
deepseek-r1:671b-0528-q4_K_M
404GB
deepseek-r1:671b-0528-q8_0
713GB
deepseek-r1:671b-0528-fp16
1.3TB
deepseek-r1:671b-q4_K_M
404GB
deepseek-r1:671b-q8_0
713GB
deepseek-r1:671b-fp16
1.3TB
## 选择合适的model，安装deepseek-r1
### 下载deepseek-r1:7b模型
#### 深度求索-R1是一系列开放推理模型，其性能接近主流模型，例如03和 Gemini2.5 Pro。
[root@Sealer tmp]# ollama pull deepseek-r1:7b
pulling manifest
pulling 96c415656d37: 100% ▕███████████████████████████████████████████████████████████ ▏ 4.7 GB/4.7 GB 8.0 MB/s 0s
verifying sha256 digest
writing manifest
success
#查看本地模型列表，确认下载完成
[root@Sealer tmp]# ollama list
NAME ID SIZE MODIFIED
deepseek-r1:7b 755ced02ce7b 4.7 GB 11 seconds ago
#测试运行
ollama run deepseek-r1:7b
## 退出模型
Use Ctrl + d or /bye to exit.
#测试API接口
[root@Sealer tmp]# curl http://127.0.0.1:11434/api/tags
{"models":[{"name":"deepseek-r1:7b","model":"deepseek-r1:7b","modified_at":"2026-08-26T16:44:10.083132066+08:00","size":4683075440,"digest":"755ced02ce7befdb13b7ca74e1e4d08cddba4986afdb63a480f2c93d3140383f","details":{"parent_model":"","format":"gguf","family":"qwen2","families":["qwen2"],"parameter_size":"7.6B","quantization_level":"Q4_K_M","context_length":131072,"embedding_length":3584},"capabilities":["completion","thinking"]}]}[root@Sealer tmp]#
浙公网安备 33010602011771号