[GH-ISSUE #1536] Ability to keep a model in memory for longer #62874

New Issue

GiteaMirror · 2026-05-03T10:35:06-05:00

GiteaMirror commented

2026-05-03 10:35:06 -05:00

Originally created by @helloimcx on GitHub (Dec 15, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/1536

is there a way to keep the model in memory or gpu memory ?

Originally created by @helloimcx on GitHub (Dec 15, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/1536 is there a way to keep the model in memory or gpu memory ?

GiteaMirror added the feature request label 2026-05-03 10:35:06 -05:00

GiteaMirror closed this issue

2026-05-03 10:35:07 -05:00

GiteaMirror commented

2026-05-03 10:35:08 -05:00

@nelsongomes commented on GitHub (Dec 15, 2023):

This a very important feature and models should be kept in memory by default. In reality, it makes sense even to keep multiple instances of same model if memory is available and the loaded models are already in use. This way Ollama can be cost effective and performant @jmorganca.

This is needed to make Ollama a usable server, just came out of a meeting and this was the main reason not to choose it, it needs to cost effective and performant.

@nelsongomes commented on GitHub (Dec 15, 2023): This a very important feature and models should be kept in memory by default. In reality, it makes sense even to keep multiple instances of same model if memory is available and the loaded models are already in use. This way Ollama can be cost effective and performant @jmorganca. This is needed to make Ollama a usable server, just came out of a meeting and this was the main reason not to choose it, it needs to cost effective and performant.

GiteaMirror commented

2026-05-03 10:35:08 -05:00

@rgaidot commented on GitHub (Dec 17, 2023):

Maybe you can do a curl every second (crontab or others) with the model (it's not great but it should work)

curl http://0.0.0.0:11434/api/chat -d '{
    "model": "mistral"
 }'

@rgaidot commented on GitHub (Dec 17, 2023): Maybe you can do a curl every second (crontab or others) with the model (it's not great but it should work) ``` curl http://0.0.0.0:11434/api/chat -d '{ "model": "mistral" }' ```

GiteaMirror commented

2026-05-03 10:35:09 -05:00

@mLpenguin commented on GitHub (Dec 17, 2023):

There is a potential solution to change the timeout from the hardcoded 5 minutes to an env variable but it is still waiting to be merged #1257

@mLpenguin commented on GitHub (Dec 17, 2023): There is a potential solution to change the timeout from the hardcoded 5 minutes to an env variable but it is still waiting to be merged #1257

GiteaMirror commented

2026-05-03 10:35:10 -05:00

@nelsongomes commented on GitHub (Dec 17, 2023):

Thanks for replying.

The issue is that GPUs are expensive and majority of models could have
several replicas inside of a GPU memory. This would allow several threads
running in parallel answering different requests.
Otherwise a 64gb memory GPU is almost wasted.

Ideally you should keep adding replicas until memory is either full or
there is the need to evict an instance to make room for a different model
to be loaded and keep all instances always loaded, no eviction if not
needed.

No one wants a server to load models on the fly if memory is enough and
make requests slower.

Do you get my point?

Thanks
Nelson Gomes

A domingo, 17/12/2023, 15:59, Régis Gaidot @.***>
escreveu:

Maybe you can do a curl every second with the model (it's not great but it
should work))

curl http://0.0.0.0:11434/api/chat -d '{
"model": "mistral"
}'

—
Reply to this email directly, view it on GitHub
https://github.com/jmorganca/ollama/issues/1536#issuecomment-1859209612,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AAITSRD2LWWUUX2JYFRZQZDYJ4JGFAVCNFSM6AAAAABAVZNPB6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNJZGIYDSNRRGI
.
You are receiving this because you commented.Message ID:
@.***>

@nelsongomes commented on GitHub (Dec 17, 2023): Thanks for replying. The issue is that GPUs are expensive and majority of models could have several replicas inside of a GPU memory. This would allow several threads running in parallel answering different requests. Otherwise a 64gb memory GPU is almost wasted. Ideally you should keep adding replicas until memory is either full or there is the need to evict an instance to make room for a different model to be loaded and keep all instances always loaded, no eviction if not needed. No one wants a server to load models on the fly if memory is enough and make requests slower. Do you get my point? Thanks Nelson Gomes A domingo, 17/12/2023, 15:59, Régis Gaidot ***@***.***> escreveu: > Maybe you can do a curl every second with the model (it's not great but it > should work)) > > curl http://0.0.0.0:11434/api/chat -d '{ > "model": "mistral" > }' > > — > Reply to this email directly, view it on GitHub > <https://github.com/jmorganca/ollama/issues/1536#issuecomment-1859209612>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAITSRD2LWWUUX2JYFRZQZDYJ4JGFAVCNFSM6AAAAABAVZNPB6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNJZGIYDSNRRGI> > . > You are receiving this because you commented.Message ID: > ***@***.***> >

GiteaMirror commented

2026-05-03 10:35:11 -05:00

@helloimcx commented on GitHub (Dec 18, 2023):

There is a potential solution to change the timeout from the hardcoded 5 minutes to an env variable but it is still waiting to be merged #1257

thanks, i changed defaultSessionDuration in routes.go with a longer period, it worked.

@helloimcx commented on GitHub (Dec 18, 2023): > There is a potential solution to change the timeout from the hardcoded 5 minutes to an env variable but it is still waiting to be merged #1257 thanks, i changed ` defaultSessionDuration` in routes.go with a longer period, it worked.

GiteaMirror commented

2026-05-03 10:35:12 -05:00

@nelsongomes commented on GitHub (Dec 18, 2023):

Ideally, we could have a fixed config file being passed to server:
{
"mistral": { "instances": 3, "target": "gpu" },
"llama2": { "instances": 1, "target": "cpu"}
}
that would allow to set up the intended usage we wanna give to the server.

@nelsongomes commented on GitHub (Dec 18, 2023): Ideally, we could have a fixed config file being passed to server: { "mistral": { "instances": 3, "target": "gpu" }, "llama2": { "instances": 1, "target": "cpu"} } that would allow to set up the intended usage we wanna give to the server.

GiteaMirror commented

2026-05-03 10:35:12 -05:00

@helloimcx commented on GitHub (Dec 19, 2023):

Ideally, we could have a fixed config file being passed to server: { "mistral": { "instances": 3, "target": "gpu" }, "llama2": { "instances": 1, "target": "cpu"} } that would allow to set up the intended usage we wanna give to the server.

yea, that would be great!

@helloimcx commented on GitHub (Dec 19, 2023): > Ideally, we could have a fixed config file being passed to server: { "mistral": { "instances": 3, "target": "gpu" }, "llama2": { "instances": 1, "target": "cpu"} } that would allow to set up the intended usage we wanna give to the server. yea, that would be great!

GiteaMirror commented

2026-05-03 10:35:12 -05:00

@nelsongomes commented on GitHub (Dec 26, 2023):

Was I able to convince you guys that memory strategy needs some revisiting?

@nelsongomes commented on GitHub (Dec 26, 2023): Was I able to convince you guys that memory strategy needs some revisiting?

GiteaMirror commented

2026-05-03 10:35:12 -05:00

@nelsongomes commented on GitHub (Dec 26, 2023):

Another option for memory management would be to add commands to manage permanently loaded models and their count... this way we could tell the server what we wanted in memory.

@nelsongomes commented on GitHub (Dec 26, 2023): Another option for memory management would be to add commands to manage permanently loaded models and their count... this way we could tell the server what we wanted in memory.

GiteaMirror commented

2026-05-03 10:35:13 -05:00

@jooneyp commented on GitHub (Dec 26, 2023):

@nelsongomes +1, I'm the another one thinking of make user decide the manage strategy of the memory. Need to make env variable or even dynamic option for each model. I have two systems with i7 6700k & 4080 + 3070 llm machine and amd 5600 & 3070 gaming machine but my gaming machine is much faster loading speed of models.

@jooneyp commented on GitHub (Dec 26, 2023): @nelsongomes +1, I'm the another one thinking of make user decide the manage strategy of the memory. Need to make env variable or even dynamic option for each model. I have two systems with i7 6700k & 4080 + 3070 llm machine and amd 5600 & 3070 gaming machine but my gaming machine is much faster loading speed of models.

GiteaMirror commented

2026-05-03 10:35:14 -05:00

@rgaidot commented on GitHub (Dec 29, 2023):

In the meantime, have you tested this 'solution' (use_mlock) https://github.com/jmorganca/ollama/issues/1672 ?

@rgaidot commented on GitHub (Dec 29, 2023): In the meantime, have you tested this 'solution' (use_mlock) https://github.com/jmorganca/ollama/issues/1672 ?

GiteaMirror commented

2026-05-03 10:35:15 -05:00

@easp commented on GitHub (Jan 2, 2024):

@nelsongomes, keeping multiple instances of a single model in memory doesn't make any sense, at all. It's a waste of memory and memory bandwidth. Concurrency is better handled by batching requests and amortizing the cost of memory access across the whole batch.

If they are going to add features to support concurrency they should focus on batching, rather than a kludge.

@easp commented on GitHub (Jan 2, 2024): @nelsongomes, keeping multiple instances of a single model in memory doesn't make any sense, at all. It's a waste of memory and memory bandwidth. Concurrency is better handled by batching requests and amortizing the cost of memory access across the whole batch. If they are going to add features to support concurrency they should focus on batching, rather than a kludge.

GiteaMirror commented

2026-05-03 10:35:16 -05:00

@nelsongomes commented on GitHub (Jan 3, 2024):

That is a valid point. Does Ollama support concurrency? Currently running
it on my MAC it doesn't but it's obvious because my MAC does not have an
usable GPU.

Erik S @.***> escreveu no dia terça, 2/01/2024 à(s)
21:38:

@nelsongomes https://github.com/nelsongomes, keeping multiple instances
of a single model in memory doesn't make any sense, at all. It's a waste of
memory and memory bandwidth. Concurrency is better handled by batching
requests and amortizing the cost of memory access across the whole batch.

If they are going to add features to support concurrency they should focus
on batching, rather than a kludge.

—
Reply to this email directly, view it on GitHub
https://github.com/jmorganca/ollama/issues/1536#issuecomment-1874598332,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AAITSRHVRNLEE65ED6BGBBLYMR45JAVCNFSM6AAAAABAVZNPB6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNZUGU4TQMZTGI
.
You are receiving this because you were mentioned.Message ID:
@.***>

@nelsongomes commented on GitHub (Jan 3, 2024): That is a valid point. Does Ollama support concurrency? Currently running it on my MAC it doesn't but it's obvious because my MAC does not have an usable GPU. Erik S ***@***.***> escreveu no dia terça, 2/01/2024 à(s) 21:38: > @nelsongomes <https://github.com/nelsongomes>, keeping multiple instances > of a single model in memory doesn't make any sense, at all. It's a waste of > memory and memory bandwidth. Concurrency is better handled by batching > requests and amortizing the cost of memory access across the whole batch. > > If they are going to add features to support concurrency they should focus > on batching, rather than a kludge. > > — > Reply to this email directly, view it on GitHub > <https://github.com/jmorganca/ollama/issues/1536#issuecomment-1874598332>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAITSRHVRNLEE65ED6BGBBLYMR45JAVCNFSM6AAAAABAVZNPB6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNZUGU4TQMZTGI> > . > You are receiving this because you were mentioned.Message ID: > ***@***.***> >

GiteaMirror commented

2026-05-03 10:35:17 -05:00

@nelsongomes commented on GitHub (Jan 4, 2024):

@easp I've done more reading about it and if I understood correctly batching is the way to go, but if a request does not enter the batch in time it needs to wait for next batch which may take seconds to start. Could it be that if we have several models loaded into GPU memory we could have several batches running continuously? Making full use of memory and concurrency?

@nelsongomes commented on GitHub (Jan 4, 2024): @easp I've done more reading about it and if I understood correctly batching is the way to go, but if a request does not enter the batch in time it needs to wait for next batch which may take seconds to start. Could it be that if we have several models loaded into GPU memory we could have several batches running continuously? Making full use of memory and concurrency?

GiteaMirror commented

2026-05-03 10:35:17 -05:00

@easp commented on GitHub (Jan 4, 2024):

For a single completion request, tokens are generated serially, which each token generated depending on the initial context + prompt and preceding tokens generated for the response. Generating each token involves computing, end to end, over the entirety of the model weights. With batch generation, the process is the same, but next tokens for a set of individual completions threads are calculated in a single traversal of the model weights.

Batches are continuous, so a new request only waits for the next set of "in-flight" tokens to complete before joining the batch for the next token(s). New completions can be started with each new batch, and in-progress completions may conclude with each new batch, even as others continue. So, at, say, 25 batch generation rate, a new request doesn't wait more than 80ms for its first token to be generated.

So, it only makes sense to have a single copy of the model loaded. It does make sense to have multiple models loaded, if resources permit, to allow quick switching between models.

As I said though, Ollama doesn't support this, at least not yet. It is supported by llama.cpp, which Ollama uses to "run" models, but I'd expect that it would require some work in the Ollama server as well to support and so far Ollama seems to be pretty focused on single-user scenarios.

@easp commented on GitHub (Jan 4, 2024): For a single completion request, tokens are generated serially, which each token generated depending on the initial context + prompt and preceding tokens generated for the response. Generating each token involves computing, end to end, over the entirety of the model weights. With batch generation, the process is the same, but next tokens for a set of individual completions threads are calculated in a single traversal of the model weights. Batches are continuous, so a new request only waits for the next set of "in-flight" tokens to complete before joining the batch for the next token(s). New completions can be started with each new batch, and in-progress completions may conclude with each new batch, even as others continue. So, at, say, 25 batch generation rate, a new request doesn't wait more than 80ms for its first token to be generated. So, it only makes sense to have a single copy of the model loaded. It does make sense to have multiple models loaded, if resources permit, to allow quick switching between models. As I said though, Ollama doesn't support this, at least not yet. It is supported by llama.cpp, which Ollama uses to "run" models, but I'd expect that it would require some work in the Ollama server as well to support and so far Ollama seems to be pretty focused on single-user scenarios.

GiteaMirror commented

2026-05-03 10:35:18 -05:00

@jmorganca commented on GitHub (Jan 4, 2024):

Thanks for creating this issue folks – sorry this isn't easier yet. Will investigate how to better support keeping models loaded for longer

@jmorganca commented on GitHub (Jan 4, 2024): Thanks for creating this issue folks – sorry this isn't easier yet. Will investigate how to better support keeping models loaded for longer

GiteaMirror commented

2026-05-03 10:35:18 -05:00

@jmorganca commented on GitHub (Jan 25, 2024):

Keep an eye on this PR which will enable keeping a model in memory indefinitely (or for a custom amount of time): https://github.com/ollama/ollama/pull/2146

@jmorganca commented on GitHub (Jan 25, 2024): Keep an eye on this PR which will enable keeping a model in memory indefinitely (or for a custom amount of time): https://github.com/ollama/ollama/pull/2146

GiteaMirror commented

2026-05-03 10:35:19 -05:00

@pdevine commented on GitHub (Jan 26, 2024):

This is merged now, so I'm going to go ahead and close the issue. You'll be able to use this in 0.1.23.

@pdevine commented on GitHub (Jan 26, 2024): This is merged now, so I'm going to go ahead and close the issue. You'll be able to use this in `0.1.23`.

GiteaMirror commented

2026-05-03 10:35:19 -05:00

@nelsongomes commented on GitHub (Jan 28, 2024):

Thank you!

Patrick Devine @.***> escreveu (sexta, 26/01/2024 à(s)
23:53):

Closed #1536 https://github.com/ollama/ollama/issues/1536 as completed.

—
Reply to this email directly, view it on GitHub
https://github.com/ollama/ollama/issues/1536#event-11614331680, or
unsubscribe
https://github.com/notifications/unsubscribe-auth/AAITSRDD5VW765AHSMCLG2TYQQ6YRAVCNFSM6AAAAABAVZNPB6VHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMJRGYYTIMZTGE3DQMA
.
You are receiving this because you were mentioned.Message ID:
@.***>

@nelsongomes commented on GitHub (Jan 28, 2024): Thank you! Patrick Devine ***@***.***> escreveu (sexta, 26/01/2024 à(s) 23:53): > Closed #1536 <https://github.com/ollama/ollama/issues/1536> as completed. > > — > Reply to this email directly, view it on GitHub > <https://github.com/ollama/ollama/issues/1536#event-11614331680>, or > unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAITSRDD5VW765AHSMCLG2TYQQ6YRAVCNFSM6AAAAABAVZNPB6VHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMJRGYYTIMZTGE3DQMA> > . > You are receiving this because you were mentioned.Message ID: > ***@***.***> >

GiteaMirror commented

2026-05-03 10:35:20 -05:00

@sirkuttin commented on GitHub (Apr 22, 2024):

62be2050dd/cmd/cmd.go (L1005)

@sirkuttin commented on GitHub (Apr 22, 2024): https://github.com/ollama/ollama/blob/62be2050dd83197864d771fe6891fc47486ee6a1/cmd/cmd.go#L1005

GiteaMirror commented

2026-05-03 10:35:20 -05:00

@loke-x commented on GitHub (Dec 29, 2024):

62be2050dd/cmd/cmd.go (L1005)

it will be great if there is lot more control given via command line

like for stropping server
increasing model in memory alive state and many more

@loke-x commented on GitHub (Dec 29, 2024): > https://github.com/ollama/ollama/blob/62be2050dd83197864d771fe6891fc47486ee6a1/cmd/cmd.go#L1005 it will be great if there is lot more control given via command line like for stropping server increasing model in memory alive state and many more

Sign in to join this conversation.

Branches Tags

main

dhiltgen/ci

parth-launch-plan-gating

hoyyeva/anthropic-reference-images-path

parth-anthropic-reference-images-path

brucemacd/download-before-remove

hoyyeva/editor-config-repair

parth-mlx-decode-checkpoints

parth-launch-codex-app

hoyyeva/fix-codex-model-metadata-warning

hoyyeva/qwen

parth/hide-claude-desktop-till-release

hoyyeva/opencode-image-modality

parth-add-claude-code-autoinstall

release_v0.22.0

pdevine/manifest-list

codex/fix-codex-model-metadata-warning

pdevine/addressable-manifest

brucemacd/launch-fetch-reccomended

jmorganca/llama-compat

launch-copilot-cli

hoyyeva/opencode-thinking

release_v0.20.7

parth-auto-save-backup

parth-test

jmorganca/gemma4-audio-replacements

fix-manifest-digest-on-pull

hoyyeva/vscode-improve

brucemacd/install-server-wait

parth/update-claude-docs

brucemac/start-ap-install

pdevine/mlx-update

pdevine/qwen35_vision

drifkin/api-show-fallback

mintlify/image-generation-1773352582

hoyyeva/server-context-length-local-config

jmorganca/faster-reptition-penalties

jmorganca/convert-nemotron

parth-pi-thinking

pdevine/sampling-penalties

jmorganca/fix-create-quantization-memory

dongchen/resumable_transfer_fix

pdevine/sampling-cache-error

jessegross/mlx-usage

hoyyeva/openclaw-config

hoyyeva/app-html

pdevine/qwen3next

brucemacd/sign-sh-install

brucemacd/tui-update

brucemacd/usage-api

jmorganca/launch-empty

fix-app-dist-embed

mxyng/mlx-compile

mxyng/mlx-quant

mxyng/mlx-glm4.7

mxyng/mlx

brucemacd/simplify-model-picker

jmorganca/qwen3-concurrent

fix-glm-4.7-flash-mla-config

drifkin/qwen3-coder-opening-tag

brucemacd/usage-cli

fix-cuda12-fattn-shmem

ollama-imagegen-docs

parth/fix-multiline-inputs

brucemacd/config-docs

mxyng/model-files

mxyng/simple-execute

fix-imagegen-ollama-models

mxyng/async-upload

jmorganca/lazy-no-dtype-changes

imagegen-auto-detect-create

parth/decrease-concurrent-download-hf

fix-mlx-quantize-init

jmorganca/x-cleanup

usage

imagegen-readme

jmorganca/glm-image

mlx-gpu-cd

jmorganca/imagegen-modelfile

parth/agent-skills

parth/agent-allowlist

parth/signed-in-offline

parth/agents

parth/fix-context-chopping

improve-cloud-flow

parth/add-models-websearch

parth/prompt-renderer-mcp

jmorganca/native-settings

jmorganca/download-stream-hash

jmorganca/client2-rebased

brucemacd/oai-chat-req-multipart

jessegross/multi_chunk_reserve

grace/additional-omit-empty

grace/mistral-3-large

mxyng/tokenizer2

mxyng/tokenizer

jessegross/flash

hoyyeva/windows-nacked-app

mxyng/cleanup-attention

grace/deepseek-parser

hoyyeva/remember-unsent-prompt

parth/add-lfs-pointer-error-conversion

parth/olmo2-test2

hoyyeva/ollama-launchagent-plist

nicole/olmo-model

parth/olmo-test

mxyng/remove-embedded

parth/render-template

jmorganca/intellect-3

parth/remove-prealloc-linter

jmorganca/cmd-eval

nicole/nomic-embed-text-fix

mxyng/lint-2

hoyyeva/add-gemini-3-pro-preview

hoyyeva/load-model-list

mxyng/expand-path

mxyng/environ-2

hoyyeva/deeplink-json-encoding

parth/improve-tool-calling-tests

hoyyeva/conversation

hoyyeva/assistant-edit-response

hoyyeva/thinking

origin/brucemacd/invalid-char-i-err

parth/improve-tool-calling

jmorganca/required-omitempty

grace/qwen3-vl-tests

mxyng/iter-client

parth/docs-readme

nicole/embed-test

pdevine/integration-benchstat

parth/remove-generate-cmd

parth/add-toolcall-id

mxyng/server-tests

jmorganca/glm-4.6

jmorganca/gin-h-compat

drifkin/stable-tool-args

pdevine/qwen3-more-thinking

parth/add-websearch-client

nicole/websearch_local

jmorganca/qwen3-coder-updates

grace/deepseek-v3-migration-tests

mxyng/fix-create

jmorganca/cloud-errors

pdevine/parser-tidy

revert-12233-parth/simplify-entrypoints-runner

parth/enable-so-gpt-oss

brucemacd/qwen3vl

jmorganca/readme-simplify

parth/gpt-oss-structured-outputs

revert-12039-jmorganca/tools-braces

mxyng/embeddings

mxyng/gguf

mxyng/benchmark

mxyng/types-null

parth/move-parsing

mxyng/gemma2

jmorganca/docs

mxyng/16-bit

mxyng/create-stdin

pdevine/authorizedkeys

mxyng/quant

parth/opt-in-error-context-window

brucemacd/cache-models

brucemacd/runner-completion

jmorganca/llama-update-6

brucemacd/benchmark-list

brucemacd/partial-read-caps

parth/deepseek-r1-tools

mxyng/omit-array

parth/tool-prefix-temp

brucemacd/runner-test

jmorganca/qwen25vl

brucemacd/model-forward-test-ext

parth/python-function-parsing

jmorganca/cuda-compression-none

drifkin/num-parallel

drifkin/chat-truncation-fix

jmorganca/sync

parth/python-tools-calling

drifkin/array-head-count

brucemacd/create-no-loop

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

qwen25omni

mxyng/v3

brucemacd/ropeconfig

jmorganca/silence-tokenizer

parth/sample-so-test

parth/sampling-structured-outputs

brucemacd/doc-go-engine

parth/constrained-sampling-json

jmorganca/mistral-wip

brucemacd/mistral-small-convert

parth/sample-unmarshal-json-for-params

brucemacd/jomorganca/mistral

pdevine/bfloat16

jmorganca/mistral

brucemacd/mistral

pdevine/logging

parth/sample-correctness-fix

parth/sample-fix-sorting

jmorgan/sample-fix-sorting-extras

jmorganca/temp-0-images

brucemacd/parallel-embed-models

brucemacd/shim-grammar

jmorganca/fix-gguf-error

bmizerany/nameswork

jmorganca/faster-releases

bmizerany/validatenames

brucemacd/err-no-vocab

brucemacd/rope-config

brucemacd/err-hint

brucemacd/qwen2_5

brucemacd/logprobs

brucemacd/new_runner_graph_bench

progress-flicker

brucemacd/forward-test

brucemacd/go_qwen2

pdevine/gemma2

jmorganca/add-missing-symlink-eval

mxyng/next-debug

parth/set-context-size-openai

brucemacd/next-bpe-bench

brucemacd/next-bpe-test

brucemacd/new_runner_e2e

brucemacd/new_runner_qwen2

pdevine/convert-cohere2

brucemacd/convert-cli

parth/log-probs

mxyng/next-mlx

mxyng/cmd-history

parth/templating

parth/tokenize-detokenize

brucemacd/check-key-register

bmizerany/grammar

jmorganca/vendor-081b29bd

mxyng/func-checks

jmorganca/fix-null-format

parth/fix-default-to-warn-json

jmorganca/qwen2vl

jmorganca/no-concat

parth/cmd-cleanup-SO

brucemacd/check-key-register-structured-err

parth/openai-stream-usage

parth/fix-referencing-so

stream-tools-stop

jmorganca/degin-1

brucemacd/install-path-clean

brucemacd/push-name-validation

brucemacd/browser-key-register

jmorganca/openai-fix-first-message

jmorganca/fix-proxy

jessegross/sample

parth/disallow-streaming-tools

dhiltgen/remove_submodule

jmorganca/ga

jmorganca/mllama

pdevine/newlines

pdevine/geems-2b

jmorganca/llama-bump

mxyng/modelname-7

mxyng/gin-slog

mxyng/modelname-6

jyan/convert-prog

jyan/quant5

paligemma-support

pdevine/import-docs

jmorganca/openai-context

jyan/paligemma

jyan/p2

jyan/palitest

bmizerany/embedspeedup

jmorganca/llama-vit

brucemacd/allow-ollama

royh/ep-methods

royh/whisper

mxyng/api-models

mxyng/fix-memory

jyan/q4_4/8

jyan/ollama-v

royh/stream-tools

roy-embed-parallel

bmizerany/hrm

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

royh/embed-viz

jyan/local2

jyan/auth

jyan/local

jyan/parse-temp

jmorganca/template-mistral

jyan/reord-g

royh-openai-suffixdocs

royh-imgembed

royh-embed-parallel

jyan/quant4

royh-precision

jyan/progress

pdevine/fix-template

jyan/quant3

pdevine/ggla

mxyng/update-registry-domain

jmorganca/ggml-static

mxyng/create-context

jyan/v0.146

mxyng/layers-from-files

build_dist

bmizerany/noseek

royh-ls

royh-name

timeout

mxyng/server-timestamp

bmizerany/nosillyggufslurps

royh-params

jmorganca/llama-cpp-7c26775

royh-openai-delete

royh-show-rigid

jmorganca/enable-fa

jmorganca/no-error-template

jyan/format

royh-testdelete

bmizerany/fastverify

language_support

pdevine/ps-glitches

brucemacd/tokenize

bruce/iq-quants

bmizerany/filepathwithcoloninhost

mxyng/split-bin

bmizerany/client-registry

jmorganca/if-none-match

native

jmorganca/native

jmorganca/batch-embeddings

jmorganca/initcmake

jmorganca/mm

pdevine/showggmlinfo

modenameenforcealphanum

bmizerany/modenameenforcealphanum

jmorganca/done-reason

jmorganca/llama-cpp-8960fe8

ollama.com

bmizerany/filepathnobuild

bmizerany/types/model/defaultfix

rmdisplaylong

nogogen

bmizerany/x

modelfile-readme

bmizerany/replacecolon

jmorganca/limit

jmorganca/execstack

jmorganca/replace-assets

mxyng/tune-concurrency

jmorganca/testing

whitespace-detection

jmorganca/options

upgrade-all

scratch

cuda-search

mattw/airenamer

mattw/allmodelsonhuggingface

mattw/quantcontext

mattw/whatneedstorun

brucemacd/llama-mem-calc

mattw/faq-context

mattw/communitylinks

mattw/noprune

mattw/python-functioncalling

rename

mxyng/install

pulse

remove-first

editor

mattw/selfqueryingretrieval

cgo

mattw/howtoquant

api

matt/streamingapi

format-config

mxyng/extra-args

shell

update-nous-hermes

cp-model

upload-progress

fix-unknown-model

fix-model-names

delete-fix

insecure-registry

ls

deletemodels

progressbar

readme-updates

license-layers

skip-list

list-models

modelpath

matt/examplemodelfiles

distribution

go-opts

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/ollama#62874