mirror of
https://github.com/ollama/ollama.git
synced 2026-05-21 05:01:48 -05:00
Closed
opened 2026-04-29 04:54:59 -05:00 by GiteaMirror
·
14 comments
No Branch/Tag Specified
main
parth-migrate-pi
parth-launch-enriched-model-inventory
dhiltgen/llama-runner
codex/make-integration-hidden-and-lunchable
hoyyeva/migrate-pi
hoyyeva/opencode-thinking
hoyyeva/anthropic-local-image-path
parth-launch-codex-app
hoyyeva/anthropic-reference-images-path
parth-anthropic-reference-images-path
brucemacd/download-before-remove
hoyyeva/editor-config-repair
parth-mlx-decode-checkpoints
hoyyeva/qwen
parth/hide-claude-desktop-till-release
parth-add-claude-code-autoinstall
release_v0.22.0
pdevine/manifest-list
codex/fix-codex-model-metadata-warning
pdevine/addressable-manifest
brucemacd/launch-fetch-reccomended
jmorganca/llama-compat
launch-copilot-cli
release_v0.20.7
parth-auto-save-backup
parth-test
jmorganca/gemma4-audio-replacements
fix-manifest-digest-on-pull
hoyyeva/vscode-improve
brucemacd/install-server-wait
parth/update-claude-docs
brucemac/start-ap-install
pdevine/mlx-update
pdevine/qwen35_vision
drifkin/api-show-fallback
mintlify/image-generation-1773352582
hoyyeva/server-context-length-local-config
jmorganca/faster-reptition-penalties
jmorganca/convert-nemotron
parth-pi-thinking
pdevine/sampling-penalties
jmorganca/fix-create-quantization-memory
dongchen/resumable_transfer_fix
pdevine/sampling-cache-error
jessegross/mlx-usage
hoyyeva/openclaw-config
hoyyeva/app-html
pdevine/qwen3next
brucemacd/sign-sh-install
brucemacd/tui-update
brucemacd/usage-api
jmorganca/launch-empty
fix-app-dist-embed
mxyng/mlx-compile
mxyng/mlx-quant
mxyng/mlx-glm4.7
mxyng/mlx
brucemacd/simplify-model-picker
jmorganca/qwen3-concurrent
fix-glm-4.7-flash-mla-config
drifkin/qwen3-coder-opening-tag
brucemacd/usage-cli
fix-cuda12-fattn-shmem
ollama-imagegen-docs
parth/fix-multiline-inputs
brucemacd/config-docs
mxyng/model-files
mxyng/simple-execute
fix-imagegen-ollama-models
mxyng/async-upload
jmorganca/lazy-no-dtype-changes
imagegen-auto-detect-create
parth/decrease-concurrent-download-hf
fix-mlx-quantize-init
jmorganca/x-cleanup
usage
imagegen-readme
jmorganca/glm-image
mlx-gpu-cd
jmorganca/imagegen-modelfile
parth/agent-skills
parth/agent-allowlist
parth/signed-in-offline
parth/agents
parth/fix-context-chopping
improve-cloud-flow
parth/add-models-websearch
parth/prompt-renderer-mcp
jmorganca/native-settings
jmorganca/download-stream-hash
jmorganca/client2-rebased
brucemacd/oai-chat-req-multipart
jessegross/multi_chunk_reserve
grace/additional-omit-empty
grace/mistral-3-large
mxyng/tokenizer2
mxyng/tokenizer
jessegross/flash
hoyyeva/windows-nacked-app
mxyng/cleanup-attention
grace/deepseek-parser
hoyyeva/remember-unsent-prompt
parth/add-lfs-pointer-error-conversion
parth/olmo2-test2
hoyyeva/ollama-launchagent-plist
nicole/olmo-model
parth/olmo-test
mxyng/remove-embedded
parth/render-template
jmorganca/intellect-3
parth/remove-prealloc-linter
jmorganca/cmd-eval
nicole/nomic-embed-text-fix
mxyng/lint-2
hoyyeva/add-gemini-3-pro-preview
hoyyeva/load-model-list
mxyng/expand-path
mxyng/environ-2
hoyyeva/deeplink-json-encoding
parth/improve-tool-calling-tests
hoyyeva/conversation
hoyyeva/assistant-edit-response
hoyyeva/thinking
origin/brucemacd/invalid-char-i-err
parth/improve-tool-calling
jmorganca/required-omitempty
grace/qwen3-vl-tests
mxyng/iter-client
parth/docs-readme
nicole/embed-test
pdevine/integration-benchstat
parth/remove-generate-cmd
parth/add-toolcall-id
mxyng/server-tests
jmorganca/glm-4.6
jmorganca/gin-h-compat
drifkin/stable-tool-args
pdevine/qwen3-more-thinking
parth/add-websearch-client
nicole/websearch_local
jmorganca/qwen3-coder-updates
grace/deepseek-v3-migration-tests
mxyng/fix-create
jmorganca/cloud-errors
pdevine/parser-tidy
revert-12233-parth/simplify-entrypoints-runner
parth/enable-so-gpt-oss
brucemacd/qwen3vl
jmorganca/readme-simplify
parth/gpt-oss-structured-outputs
revert-12039-jmorganca/tools-braces
mxyng/embeddings
mxyng/gguf
mxyng/benchmark
mxyng/types-null
parth/move-parsing
mxyng/gemma2
jmorganca/docs
mxyng/16-bit
mxyng/create-stdin
pdevine/authorizedkeys
mxyng/quant
parth/opt-in-error-context-window
brucemacd/cache-models
brucemacd/runner-completion
jmorganca/llama-update-6
brucemacd/benchmark-list
brucemacd/partial-read-caps
parth/deepseek-r1-tools
mxyng/omit-array
parth/tool-prefix-temp
brucemacd/runner-test
jmorganca/qwen25vl
brucemacd/model-forward-test-ext
parth/python-function-parsing
jmorganca/cuda-compression-none
drifkin/num-parallel
drifkin/chat-truncation-fix
jmorganca/sync
parth/python-tools-calling
drifkin/array-head-count
brucemacd/create-no-loop
parth/server-enable-content-stream-with-tools
qwen25omni
mxyng/v3
brucemacd/ropeconfig
jmorganca/silence-tokenizer
parth/sample-so-test
parth/sampling-structured-outputs
brucemacd/doc-go-engine
parth/constrained-sampling-json
jmorganca/mistral-wip
brucemacd/mistral-small-convert
parth/sample-unmarshal-json-for-params
brucemacd/jomorganca/mistral
pdevine/bfloat16
jmorganca/mistral
brucemacd/mistral
pdevine/logging
parth/sample-correctness-fix
parth/sample-fix-sorting
jmorgan/sample-fix-sorting-extras
jmorganca/temp-0-images
brucemacd/parallel-embed-models
brucemacd/shim-grammar
jmorganca/fix-gguf-error
bmizerany/nameswork
jmorganca/faster-releases
bmizerany/validatenames
brucemacd/err-no-vocab
brucemacd/rope-config
brucemacd/err-hint
brucemacd/qwen2_5
brucemacd/logprobs
brucemacd/new_runner_graph_bench
progress-flicker
brucemacd/forward-test
brucemacd/go_qwen2
pdevine/gemma2
jmorganca/add-missing-symlink-eval
mxyng/next-debug
parth/set-context-size-openai
brucemacd/next-bpe-bench
brucemacd/next-bpe-test
brucemacd/new_runner_e2e
brucemacd/new_runner_qwen2
pdevine/convert-cohere2
brucemacd/convert-cli
parth/log-probs
mxyng/next-mlx
mxyng/cmd-history
parth/templating
parth/tokenize-detokenize
brucemacd/check-key-register
bmizerany/grammar
jmorganca/vendor-081b29bd
mxyng/func-checks
jmorganca/fix-null-format
parth/fix-default-to-warn-json
jmorganca/qwen2vl
jmorganca/no-concat
parth/cmd-cleanup-SO
brucemacd/check-key-register-structured-err
parth/openai-stream-usage
parth/fix-referencing-so
stream-tools-stop
jmorganca/degin-1
brucemacd/install-path-clean
brucemacd/push-name-validation
brucemacd/browser-key-register
jmorganca/openai-fix-first-message
jmorganca/fix-proxy
jessegross/sample
parth/disallow-streaming-tools
dhiltgen/remove_submodule
jmorganca/ga
jmorganca/mllama
pdevine/newlines
pdevine/geems-2b
jmorganca/llama-bump
mxyng/modelname-7
mxyng/gin-slog
mxyng/modelname-6
jyan/convert-prog
jyan/quant5
paligemma-support
pdevine/import-docs
jmorganca/openai-context
jyan/paligemma
jyan/p2
jyan/palitest
bmizerany/embedspeedup
jmorganca/llama-vit
brucemacd/allow-ollama
royh/ep-methods
royh/whisper
mxyng/api-models
mxyng/fix-memory
jyan/q4_4/8
jyan/ollama-v
royh/stream-tools
roy-embed-parallel
bmizerany/hrm
revert-5963-revert-5924-mxyng/llama3.1-rope
royh/embed-viz
jyan/local2
jyan/auth
jyan/local
jyan/parse-temp
jmorganca/template-mistral
jyan/reord-g
royh-openai-suffixdocs
royh-imgembed
royh-embed-parallel
jyan/quant4
royh-precision
jyan/progress
pdevine/fix-template
jyan/quant3
pdevine/ggla
mxyng/update-registry-domain
jmorganca/ggml-static
mxyng/create-context
jyan/v0.146
mxyng/layers-from-files
build_dist
bmizerany/noseek
royh-ls
royh-name
timeout
mxyng/server-timestamp
bmizerany/nosillyggufslurps
royh-params
jmorganca/llama-cpp-7c26775
royh-openai-delete
royh-show-rigid
jmorganca/enable-fa
jmorganca/no-error-template
jyan/format
royh-testdelete
bmizerany/fastverify
language_support
pdevine/ps-glitches
brucemacd/tokenize
bruce/iq-quants
bmizerany/filepathwithcoloninhost
mxyng/split-bin
bmizerany/client-registry
jmorganca/if-none-match
native
jmorganca/native
jmorganca/batch-embeddings
jmorganca/initcmake
jmorganca/mm
pdevine/showggmlinfo
modenameenforcealphanum
bmizerany/modenameenforcealphanum
jmorganca/done-reason
jmorganca/llama-cpp-8960fe8
ollama.com
bmizerany/filepathnobuild
bmizerany/types/model/defaultfix
rmdisplaylong
nogogen
bmizerany/x
modelfile-readme
bmizerany/replacecolon
jmorganca/limit
jmorganca/execstack
jmorganca/replace-assets
mxyng/tune-concurrency
jmorganca/testing
whitespace-detection
jmorganca/options
upgrade-all
scratch
cuda-search
mattw/airenamer
mattw/allmodelsonhuggingface
mattw/quantcontext
mattw/whatneedstorun
brucemacd/llama-mem-calc
mattw/faq-context
mattw/communitylinks
mattw/noprune
mattw/python-functioncalling
rename
mxyng/install
pulse
remove-first
editor
mattw/selfqueryingretrieval
cgo
mattw/howtoquant
api
matt/streamingapi
format-config
mxyng/extra-args
shell
update-nous-hermes
cp-model
upload-progress
fix-unknown-model
fix-model-names
delete-fix
insecure-registry
ls
deletemodels
progressbar
readme-updates
license-layers
skip-list
list-models
modelpath
matt/examplemodelfiles
distribution
go-opts
v0.30.0-rc21
v0.30.0-rc20
v0.30.0-rc19
v0.30.0-rc18
v0.25.0-rc0
v0.30.0-rc17
v0.30.0-rc16
v0.24.0-rc1
v0.24.0
v0.24.0-rc0
v0.23.4
v0.23.4-rc0
v0.30.0-rc15
v0.23.3
v0.23.3-rc1
v0.30.0-rc14
v0.23.3-rc0
v0.30.0-rc13
v0.30.0-rc12
v0.30.0-rc11
v0.30.0-rc10
v0.30.0-rc9
v0.30.0-rc8
v0.30.0-rc7
v0.30.0-rc6
v0.30.0-rc5
v0.23.2
v0.23.2-rc0
v0.30.0-rc4
v0.30.0-rc3
v0.30.0-rc2
v0.30.0-rc1
v0.30.0-rc0
v0.23.1
v0.23.1-rc0
v0.23.0
v0.23.0-rc0
v0.22.1
v0.22.1-rc1
v0.22.1-rc0
v0.22.0
v0.22.0-rc1
v0.21.3-rc0
v0.21.2-rc1
v0.21.2
v0.21.2-rc0
v0.21.1
v0.21.1-rc1
v0.21.1-rc0
v0.21.0
v0.21.0-rc1
v0.21.0-rc0
v0.20.8-rc0
v0.20.7
v0.20.7-rc1
v0.20.7-rc0
v0.20.6
v0.20.6-rc1
v0.20.6-rc0
v0.20.5
v0.20.5-rc2
v0.20.5-rc1
v0.20.5-rc0
v0.20.4
v0.20.4-rc2
v0.20.4-rc1
v0.20.4-rc0
v0.20.3
v0.20.3-rc0
v0.20.2
v0.20.1
v0.20.1-rc2
v0.20.1-rc1
v0.20.1-rc0
v0.20.0
v0.20.0-rc1
v0.20.0-rc0
v0.19.0
v0.19.0-rc2
v0.19.0-rc1
v0.19.0-rc0
v0.18.4-rc1
v0.18.4-rc0
v0.18.3
v0.18.3-rc2
v0.18.3-rc1
v0.18.3-rc0
v0.18.2
v0.18.2-rc1
v0.18.2-rc0
v0.18.1
v0.18.1-rc1
v0.18.1-rc0
v0.18.0
v0.18.0-rc2
v0.18.0-rc1
v0.18.0-rc0
v0.17.8-rc4
v0.17.8-rc3
v0.17.8-rc2
v0.17.8-rc1
v0.17.8-rc0
v0.17.7
v0.17.7-rc2
v0.17.7-rc1
v0.17.7-rc0
v0.17.6
v0.17.5
v0.17.4
v0.17.3
v0.17.2
v0.17.1
v0.17.1-rc2
v0.17.1-rc1
v0.17.1-rc0
v0.17.0
v0.17.0-rc2
v0.17.0-rc1
v0.17.0-rc0
v0.16.3
v0.16.3-rc2
v0.16.3-rc1
v0.16.3-rc0
v0.16.2
v0.16.2-rc0
v0.16.1
v0.16.0
v0.16.0-rc2
v0.16.0-rc0
v0.16.0-rc1
v0.15.6
v0.15.5
v0.15.5-rc5
v0.15.5-rc4
v0.15.5-rc3
v0.15.5-rc2
v0.15.5-rc1
v0.15.5-rc0
v0.15.4
v0.15.3
v0.15.2
v0.15.1
v0.15.1-rc1
v0.15.1-rc0
v0.15.0-rc6
v0.15.0
v0.15.0-rc5
v0.15.0-rc4
v0.15.0-rc3
v0.15.0-rc2
v0.15.0-rc1
v0.15.0-rc0
v0.14.3
v0.14.3-rc3
v0.14.3-rc2
v0.14.3-rc1
v0.14.3-rc0
v0.14.2
v0.14.2-rc1
v0.14.2-rc0
v0.14.1
v0.14.0-rc11
v0.14.0
v0.14.0-rc10
v0.14.0-rc9
v0.14.0-rc8
v0.14.0-rc7
v0.14.0-rc6
v0.14.0-rc5
v0.14.0-rc4
v0.14.0-rc3
v0.14.0-rc2
v0.14.0-rc1
v0.14.0-rc0
v0.13.5
v0.13.5-rc1
v0.13.5-rc0
v0.13.4-rc2
v0.13.4
v0.13.4-rc1
v0.13.4-rc0
v0.13.3
v0.13.3-rc1
v0.13.3-rc0
v0.13.2
v0.13.2-rc2
v0.13.2-rc1
v0.13.2-rc0
v0.13.1
v0.13.1-rc2
v0.13.1-rc1
v0.13.1-rc0
v0.13.0
v0.13.0-rc0
v0.12.11
v0.12.11-rc1
v0.12.11-rc0
v0.12.10
v0.12.10-rc1
v0.12.10-rc0
v0.12.9-rc0
v0.12.9
v0.12.8
v0.12.8-rc0
v0.12.7
v0.12.7-rc1
v0.12.7-rc0
v0.12.7-citest0
v0.12.6
v0.12.6-rc1
v0.12.6-rc0
v0.12.5
v0.12.5-rc0
v0.12.4
v0.12.4-rc7
v0.12.4-rc6
v0.12.4-rc5
v0.12.4-rc4
v0.12.4-rc3
v0.12.4-rc2
v0.12.4-rc1
v0.12.4-rc0
v0.12.3
v0.12.2
v0.12.2-rc0
v0.12.1
v0.12.1-rc1
v0.12.1-rc2
v0.12.1-rc0
v0.12.0
v0.12.0-rc1
v0.12.0-rc0
v0.11.11
v0.11.11-rc3
v0.11.11-rc2
v0.11.11-rc1
v0.11.11-rc0
v0.11.10
v0.11.9
v0.11.9-rc0
v0.11.8
v0.11.8-rc0
v0.11.7-rc1
v0.11.7-rc0
v0.11.7
v0.11.6
v0.11.6-rc0
v0.11.5-rc4
v0.11.5-rc3
v0.11.5
v0.11.5-rc5
v0.11.5-rc2
v0.11.5-rc1
v0.11.5-rc0
v0.11.4
v0.11.4-rc0
v0.11.3
v0.11.3-rc0
v0.11.2
v0.11.1
v0.11.0-rc0
v0.11.0-rc1
v0.11.0-rc2
v0.11.0
v0.10.2-int1
v0.10.1
v0.10.0
v0.10.0-rc4
v0.10.0-rc3
v0.10.0-rc2
v0.10.0-rc1
v0.10.0-rc0
v0.9.7-rc1
v0.9.7-rc0
v0.9.6
v0.9.6-rc0
v0.9.6-ci0
v0.9.5
v0.9.4-rc5
v0.9.4-rc6
v0.9.4
v0.9.4-rc3
v0.9.4-rc4
v0.9.4-rc1
v0.9.4-rc2
v0.9.4-rc0
v0.9.3
v0.9.3-rc5
v0.9.4-citest0
v0.9.3-rc4
v0.9.3-rc3
v0.9.3-rc2
v0.9.3-rc1
v0.9.3-rc0
v0.9.2
v0.9.1
v0.9.1-rc1
v0.9.1-rc0
v0.9.1-ci1
v0.9.1-ci0
v0.9.0
v0.9.0-rc0
v0.8.0
v0.8.0-rc0
v0.7.1-rc2
v0.7.1
v0.7.1-rc1
v0.7.1-rc0
v0.7.0
v0.7.0-rc1
v0.7.0-rc0
v0.6.9-rc0
v0.6.8
v0.6.8-rc0
v0.6.7
v0.6.7-rc2
v0.6.7-rc1
v0.6.7-rc0
v0.6.6
v0.6.6-rc2
v0.6.6-rc1
v0.6.6-rc0
v0.6.5-rc1
v0.6.5
v0.6.5-rc0
v0.6.4-rc0
v0.6.4
v0.6.3-rc1
v0.6.3
v0.6.3-rc0
v0.6.2
v0.6.2-rc0
v0.6.1
v0.6.1-rc0
v0.6.0-rc0
v0.6.0
v0.5.14-rc0
v0.5.13
v0.5.13-rc6
v0.5.13-rc5
v0.5.13-rc4
v0.5.13-rc3
v0.5.13-rc2
v0.5.13-rc1
v0.5.13-rc0
v0.5.12
v0.5.12-rc1
v0.5.12-rc0
v0.5.11
v0.5.10
v0.5.9
v0.5.9-rc0
v0.5.8-rc13
v0.5.8
v0.5.8-rc12
v0.5.8-rc11
v0.5.8-rc10
v0.5.8-rc9
v0.5.8-rc8
v0.5.8-rc7
v0.5.8-rc6
v0.5.8-rc5
v0.5.8-rc4
v0.5.8-rc3
v0.5.8-rc2
v0.5.8-rc1
v0.5.8-rc0
v0.5.7
v0.5.6
v0.5.5
v0.5.5-rc0
v0.5.4
v0.5.3
v0.5.3-rc0
v0.5.2
v0.5.2-rc3
v0.5.2-rc2
v0.5.2-rc1
v0.5.2-rc0
v0.5.1
v0.5.0
v0.5.0-rc1
v0.4.8-rc0
v0.4.7
v0.4.6
v0.4.5
v0.4.4
v0.4.3
v0.4.3-rc0
v0.4.2
v0.4.2-rc1
v0.4.2-rc0
v0.4.1
v0.4.1-rc0
v0.4.0
v0.4.0-rc8
v0.4.0-rc7
v0.4.0-rc6
v0.4.0-rc5
v0.4.0-rc4
v0.4.0-rc3
v0.4.0-rc2
v0.4.0-rc1
v0.4.0-rc0
v0.4.0-ci3
v0.3.14
v0.3.14-rc0
v0.3.13
v0.3.12
v0.3.12-rc5
v0.3.12-rc4
v0.3.12-rc3
v0.3.12-rc2
v0.3.12-rc1
v0.3.11
v0.3.11-rc4
v0.3.11-rc3
v0.3.11-rc2
v0.3.11-rc1
v0.3.10
v0.3.10-rc1
v0.3.9
v0.3.8
v0.3.7
v0.3.7-rc6
v0.3.7-rc5
v0.3.7-rc4
v0.3.7-rc3
v0.3.7-rc2
v0.3.7-rc1
v0.3.6
v0.3.5
v0.3.4
v0.3.3
v0.3.2
v0.3.1
v0.3.0
v0.2.8
v0.2.8-rc2
v0.2.8-rc1
v0.2.7
v0.2.6
v0.2.5
v0.2.4
v0.2.3
v0.2.2
v0.2.2-rc2
v0.2.2-rc1
v0.2.1
v0.2.0
v0.1.49-rc14
v0.1.49-rc13
v0.1.49-rc12
v0.1.49-rc11
v0.1.49-rc10
v0.1.49-rc9
v0.1.49-rc8
v0.1.49-rc7
v0.1.49-rc6
v0.1.49-rc4
v0.1.49-rc5
v0.1.49-rc3
v0.1.49-rc2
v0.1.49-rc1
v0.1.48
v0.1.47
v0.1.46
v0.1.45-rc5
v0.1.45
v0.1.45-rc4
v0.1.45-rc3
v0.1.45-rc2
v0.1.45-rc1
v0.1.44
v0.1.43
v0.1.42
v0.1.41
v0.1.40
v0.1.40-rc1
v0.1.39
v0.1.39-rc2
v0.1.39-rc1
v0.1.38
v0.1.37
v0.1.36
v0.1.35
v0.1.35-rc1
v0.1.34
v0.1.34-rc1
v0.1.33
v0.1.33-rc7
v0.1.33-rc6
v0.1.33-rc5
v0.1.33-rc4
v0.1.33-rc3
v0.1.33-rc2
v0.1.33-rc1
v0.1.32
v0.1.32-rc2
v0.1.32-rc1
v0.1.31
v0.1.30
v0.1.29
v0.1.28
v0.1.27
v0.1.26
v0.1.25
v0.1.24
v0.1.23
v0.1.22
v0.1.21
v0.1.20
v0.1.19
v0.1.18
v0.1.17
v0.1.16
v0.1.15
v0.1.14
v0.1.13
v0.1.12
v0.1.11
v0.1.10
v0.1.9
v0.1.8
v0.1.7
v0.1.6
v0.1.5
v0.1.4
v0.1.3
v0.1.2
v0.1.1
v0.1.0
v0.0.21
v0.0.20
v0.0.19
v0.0.18
v0.0.17
v0.0.16
v0.0.15
v0.0.14
v0.0.13
v0.0.12
v0.0.11
v0.0.10
v0.0.9
v0.0.8
v0.0.7
v0.0.6
v0.0.5
v0.0.4
v0.0.3
v0.0.2
v0.0.1
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
python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
Mirrored from GitHub Pull Request
No Label
feature request
Milestone
No items
No Milestone
Projects
Clear projects
No project
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: github-starred/ollama#53889
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
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?
Originally created by @jeepshop on GitHub (Jun 25, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11196
I know that Ollama will batch queries to a Model, but all that really seems to do is make users wait in a queue. Ollama seems to have no problems loading and running simultaneous models, but only if they are different. What I'd really like to see is to allow Ollama to load multiple instances of the same model to service requests.
For example, I have 2x V100s and run the Devstral model for code completion. And even though I have 2x cards, and multiple users using code completion, they all get stuck on one GPU in a queue while the other GPU is empty and idle. I've tried all combinations of settings, but only 1 user ever gets serviced at a time unless I make some users use a different model.
@rick-github commented on GitHub (Jun 25, 2025):
Ollama doesn't support loading the same model more than once (#3902). To run simultaneous parallel queries, set
OLLAMA_NUM_PARALLEL.@jeepshop commented on GitHub (Jun 25, 2025):
Thanks for pointing out the other ticket, I searched for a while and
couldn't find a similar one.
Yes I know about OLLAMA_NUM_PARALLEL, and no it doesn't work. I can set it
to 1, 2, 4 and it still only uses one GPU and only answers 1 query at a
time.
On Wed, Jun 25, 2025 at 11:24 AM frob @.***> wrote:
@rick-github commented on GitHub (Jun 25, 2025):
How well it work depends on your hardware. For example, on a 4070:
If your hardware doesn't parallelize very well, you can try running a model per GPU and using a reverse proxy server to distribute queries similar to what's described here
@jeepshop commented on GitHub (Jun 25, 2025):
Is there documentation somewhere that lists which features of Ollama works
on what cards?
Seems the V100 just doesn't handle parallel (yes it's old) but it also has
32GB VRAM which makes it better at running larger models than a 4070. And a
$6000 A100 isn't in the cards right now :(
I did look into the proxy concept, but then I lost the ability to run
certain models that require both cards. And I can't just reconfigure the
computer every time an employee's needs shift a bit.
Having Ollama be able to spin up an identical model just makes sense to me,
and many others. Is it that much of a code change since it already can spin
up several 'different' models to let it spin up several 'identical' models?
On Wed, Jun 25, 2025 at 2:18 PM frob @.***> wrote:
@rick-github commented on GitHub (Jun 25, 2025):
In my experience the A100 has the same problem with parallelization.
It's not a requirement to bind a GPU to an ollama server. If two servers have access to both cards, then one is free to take the combined VRAM for loading a larger model, although the other server would have to be told to evict any running models. That function could be handled by the proxy.
@jeepshop commented on GitHub (Jun 26, 2025):
Yes and we've discussed the issues with doing so above; Namely that I lose
the ability to run models on both cards without reconfiguring a bunch of
instances throughout the day as users need's change.
On Thu, Jun 26, 2025 at 4:38 PM sammyvoncheese @.***>
wrote:
@rick-github commented on GitHub (Jun 27, 2025):
It turns out I was a bit unfair to the A100. It does scale, just not as well as the 4070:
If you share logs with
OLLAMA_DEBUG=1there might be some information logged about the V100s which could identify the bottleneck.@jeepshop commented on GitHub (Jun 30, 2025):
Here is a log after a fresh restart of Ollama, and loading Devstral. I don't know what I'm looking for, but it seems as if --parallel is set to 2.
Do I have a fundamental misunderstanding of how parallel works? I assumed it meant how many parallel queries (api services) it could run at the same time, directly relating to how many users could run, but is that really true? Or does something like AnythingLLM use multiple simultanious queries somehow negating the benifit to multiple users?
For example, in your attached Graph of an 2x A100 - Can it really serve 4-10 users simultaneously???
@jeepshop commented on GitHub (Jun 30, 2025):
Couldn't fit the whole log into one comment, here is the 2nd half:
@rick-github commented on GitHub (Jun 30, 2025):
If
OLLAMA_NUM_PARALLELis unset, the ollama server takes a look at available VRAM to set the number of parallel requests. If there's enough VRAM to hold two context buffers, it sets parallelism to 2, otherwise 1.It's the number of simultaneous requests that the ollama server will handle.
If
OLLAMA_NUM_PARALLEL=10, then the system will process up to 10 simultaneous queries. Because there's only a limited amount of GPU to be had, the tokens/sec rate for each individual request will be lower than if the system was processing 10 requests in serial, but the overall tokens/sec of the system will be higher.For a simple test of the parallelism of your system, try the following: set
OLLAMA_NUM_PARALLEL=10in the server environment, restart the ollama server, and run the following:This sends ten queries with a varying amount of parallelism, and systems that support that should see the overall time go down as parallelism increases. For example, on an RTX6000:
@jeepshop commented on GitHub (Jun 30, 2025):
First thank you for taking the time to help me process/understand. It
means a lot to me. Getting your script to not dump over to CPU took me a
couple hours, and a learned a lot in the process. My biggest problem
was because the model kept defaulting to 128k context. I changed the above
script to force context size and ran several iterations watching use from
nvtop.
@4k Context (Fits on single GPU)
1 39.333
2 19.789
3 19.235
4 17.930
5 16.144
6 16.918
7 17.359
8 18.518
9 16.616
10 13.493
@8k Context (Fits in single GPU)
1 33.063
2 20.042
3 19.327
4 18.050
5 16.261
6 17.102
7 17.990
8 18.614
9 16.317
10 15.833
@16k Context (Spans both GPUs ~50% VRAM)
1 41.768
2 20.028
3 19.442
4 18.124
5 16.164
6 16.799
7 17.152
8 18.826
9 16.180
10 13.242
NVTOP During 16k context run shows about 50% VRAM, and only 50% GPU use...
Since I'm getting about the same results as the smaller context runs that
used a single GPU, I feel that running a duplicate model on the second GPU
should double my performance, whereas the performance is nearly identical
between running 1 vs 2 GPUs above.
But I admit I still don't understand all the ramifications;
[image: image.png]
But amongst all of the above testing, I went back to AnythingLLM and spun
up multiple instances in Firefox - and lo-and-behold for the first time
ever I got two simultaneous chats working. Since it never worked before
with Parallel = 2, what are the expectations? Is it Users + 1, Users x 2?
On Mon, Jun 30, 2025 at 11:07 AM frob @.***> wrote:
@rick-github commented on GitHub (Jun 30, 2025):
The default context for the server can be set with
OLLAMA_CONTEXT_LENGTH.Models are a set of layers, and computation in each layer needs to be completed before computation starts in the next. If a model is composed of 40 layers, 20 get loaded in GPU1 and 20 in GPU2. Inference starts in the first layer in GPU1, goes to layer 20, then intermediate results get moved to GPU2, and inference continues with layer 1 of GPU (21st layer of the model), then to layer 20 of GPU2. That completes one token, and processing starts again at layer 1 in GPU1 for the next token.
This means that for a single completion, the average utilization of the GPUs is 50%. The test script tries to maximize parallel processing (same seed, same prompt, same token limit) so even when doing multiple queries, processing will be sequentially localized to one of the GPUs during processing. This means that over all queries, average utilization of the two GPUs will be 50%. In real life, the query processing would be more distributed, likely resulting in more than 50% utilization, although the actual amount will depend on hardware and query patterns.
The problem with running a second copy is that you waste a bunch of VRAM that could be used to either increase parallelism or increase context size (or both). Running a second copy is doable via the proxy method as described earlier, for which the downsides have already been discussed.
Parallel=2 is only true if
OLLAMA_NUM_PARALLELis unset and there is enough VRAM to hold two context buffers. It could be that in previous chats, context size was larger or some VRAM was being used that resulted in ollama using parallel=1. Every extra context buffer (ie, every increment ofOLLAMA_NUM_PARALLEL) results in another concurrent request processing slot, so increasing the number of simultaneous requests. However, if all of those requests slots are filled at the same time, the processing time will increase, so clients will see a lower tokens/second rate. So the capacity increase is not a linear function and depends on a number of factors - parallelism, types of queries, hardware, etc - see the earlier graphs.@jeepshop commented on GitHub (Jul 1, 2025):
Thank you, that makes a lot of since. I feel like I've read it before, but it never clicked.
You see it as a problem 'wasting VRAM' for parallelism. I see it as a worthy tradeoff wasting some VRAM to get a 30% performance gain.
That said, I might have found a work around - use two 'nearly' identical models; e.g. one that is 4_k_m and one that is 4_k_xl. It's not ideal because I have to balance configurations between developers, but it's better than dedicating GPUs behind a proxy.
@rick-github commented on GitHub (Jul 1, 2025):
You can create a copy with the exact same weights by copying the GGUF file of the model, hex editing an entry in the KV table to change a byte, and then using
ollama createto create the new model. Due to the edited KV entry the model will have a different sha256sum and ollama will happily load it alongside the original model. Then you could run a simple proxy that switches themodelparameter in an API request to choose one or the other model with a choice of selection algorithms: round robin, sticky, least loaded, etc. This can be done with litellm or a simple script in nginx or mitmproxy. This way the developers wouldn't have to deal with configuration, they just send a request toour-developer-model.