Ollama环境变量列表

Ollama提供了一个管理大语言模型的便捷接口，其部分配置参数是通过环境变量指定的。

请注意，修改环境变量后需要重新启动Ollama以使改动生效。

// Global Options
"OLLAMA_DEBUG":             "Show additional debug information (e.g. OLLAMA_DEBUG=1)"
"OLLAMA_FLASH_ATTENTION":   "Enabled flash attention"
"OLLAMA_GPU_OVERHEAD":      "Reserve a portion of VRAM per GPU (bytes)"
"OLLAMA_HOST":              "IP Address for the ollama server (default 127.0.0.1:11434)"
"OLLAMA_KEEP_ALIVE":        "The duration that models stay loaded in memory (default \"5m\")"
"OLLAMA_KV_CACHE_TYPE":     "The quantization type for the K/V cache (default is \"f16\")"
"OLLAMA_LLM_LIBRARY":       "Set LLM library to bypass autodetection"
"OLLAMA_LOAD_TIMEOUT":      "How long to allow model loads to stall before giving up (default \"5m\")"
"OLLAMA_MAX_LOADED_MODELS": "Maximum number of loaded models per GPU"
"OLLAMA_MAX_QUEUE":         "Maximum number of queued requests"
"OLLAMA_MODELS":            "The path to the models directory"
"OLLAMA_NOHISTORY":         "Do not preserve readline history"
"OLLAMA_NOPRUNE":           "Do not prune model blobs on startup"
"OLLAMA_NUM_PARALLEL":      "Maximum number of parallel requests"
"OLLAMA_ORIGINS":           "A comma separated list of allowed origins"
"OLLAMA_SCHED_SPREAD":      "Always schedule model across all GPUs"
"OLLAMA_TMPDIR":            "Location for temporary files"
"OLLAMA_MULTIUSER_CACHE":   "Optimize prompt caching for multi-user scenarios"
"OLLAMA_CONTEXT_LENGTH":    "Context length to use unless otherwise specified (default: 4096)"
"OLLAMA_NEW_ENGINE":        "Enable the new Ollama engine"

// Proxy Options
"HTTP_PROXY":  "HTTP proxy"
"HTTPS_PROXY": "HTTPS proxy"
"NO_PROXY":    "No proxy"

// GPU Detection Options
"CUDA_VISIBLE_DEVICES":        "Set which NVIDIA devices are visible"
"HIP_VISIBLE_DEVICES":        "Set which AMD devices are visible by numeric ID"
"ROCR_VISIBLE_DEVICES":        "Set which AMD devices are visible by UUID or numeric ID"
"GPU_DEVICE_ORDINAL":        "Set which AMD devices are visible by numeric ID"
"HSA_OVERRIDE_GFX_VERSION":    "Override the gfx used for all detected AMD GPUs"
"OLLAMA_INTEL_GPU":            "Enable experimental Intel GPU detection"

参考资料：

https://github.com/ollama/ollama/blob/main/envconfig/config.go

https://github.com/ollama/ollama/blob/main/docs/faq.md

https://github.com/ollama/ollama/issues/2941

https://www.whuanle.cn/archives/21668

https://zhuanlan.zhihu.com/p/7766226260

https://www.lvtao.net/tool/ollama-parallel-max-models.html

https://zhuanlan.zhihu.com/p/719540154