2 Releases published by 1 person

• v0.5.1

  published Dec 6, 2024

• v0.5.2-rc3 v0.5.2

  published Dec 11, 2024

25 Pull requests merged by 13 people

• llama: enable JSON schema key ordering for generating grammars

  #8055 merged Dec 12, 2024

• server: more support for mixed-case model names

  #8017 merged Dec 11, 2024

• ci: fix linux version

  #8054 merged Dec 11, 2024

• llama: preserve field order in user-defined JSON schemas

  #8002 merged Dec 11, 2024

• ci: fix artifact path prefix for missing windows payloads

  #8052 merged Dec 11, 2024

• win: builtin arm runner

  #8039 merged Dec 11, 2024

• ci: build dir changed

  #8037 merged Dec 11, 2024

• llama: update llama.cpp and ggml to 40c6d79f

  #7875 merged Dec 11, 2024

• go.mod: go 1.22.8 -> 1.23.4

  #8036 merged Dec 11, 2024

• Return err when NewHipLib() detect error.

  #8012 merged Dec 11, 2024

• Add AI Summary Helper to list of community integrations

  #7202 merged Dec 11, 2024

• Update README.md

  #7948 merged Dec 10, 2024

• Lowercase hostname for CORS.

  #5851 merged Dec 10, 2024

• readme: add aidful-ollama-model-delete to community integration

  #8024 merged Dec 10, 2024

• Remove unused runner CpuFeatures

  #8032 merged Dec 10, 2024

• Fix some typos in documentation, code, code comments etc.

  #7021 merged Dec 10, 2024

• build: fix typo in override variable

  #8031 merged Dec 10, 2024

• build: Make target improvements

  #7499 merged Dec 10, 2024

• Avoid underflow when FreeMemory < overhead

  #8014 merged Dec 10, 2024

• prompt: Don't trim whitespace from prompts

  #7980 merged Dec 9, 2024

• Update OpenAI docs to reflect tool use functionality

  #7960 merged Dec 8, 2024

• Fix wrong import python package.

  #7986 merged Dec 8, 2024

• bugfix: "null" value for format

  #7979 merged Dec 6, 2024

• readme: add llama3.3 to readme

  #7975 merged Dec 6, 2024

• Update readmes for structured outputs

  #7962 merged Dec 6, 2024

21 Pull requests opened by 20 people

• Add IntelliBar to list of community integrations

  #7950 opened Dec 5, 2024

• wip: next ollama runner build updates

  #7954 opened Dec 5, 2024

• merge llama/ggml into ml/backend/ggml

  #7957 opened Dec 5, 2024

• macapp: add error handling for symlink operations

  #7961 opened Dec 6, 2024

• openai: finish streaming tool calls as tool_calls

  #7963 opened Dec 6, 2024

• Fix message truncation logic and ensure at least one system message i…

  #7964 opened Dec 6, 2024

• server: print environment variables in sorted, one-line

  #7970 opened Dec 6, 2024

• ADD: OLLAMA_LLM_DEFAULT

  #7971 opened Dec 6, 2024

• Document that `--format` now supports passing JSON Schemas

  #7973 opened Dec 6, 2024

• Add K/V cache quantization config to Modelfile (Follow-Up to PR #6279)

  #7983 opened Dec 7, 2024

• api: change /delete endpoint to use POST method

  #8018 opened Dec 9, 2024

• Delete redundancy code when never happends.

  #8019 opened Dec 9, 2024

• readme + examples: add pgai to readme projects and add example for semantic search

  #8028 opened Dec 10, 2024

• Prevent model thrashing from unset num_ctx

  #8029 opened Dec 10, 2024

• readme: include IBM Granite models

  #8030 opened Dec 10, 2024

• cmd: Add --base2 option to ps to show model sizes in KiB/MiB/GiB

  #8034 opened Dec 11, 2024

• Implement OLLAMA_MAX_KEEP_ALIVE environment variable

  #8049 opened Dec 11, 2024

• feat: add option to specify runner name and path in env

  #8051 opened Dec 11, 2024

• tmp

  #8059 opened Dec 12, 2024

• docs: remove repetitive words

  #8060 opened Dec 12, 2024

• Refactor fixBlobs to use WalkDir for efficiency instead of Walk

  #8061 opened Dec 12, 2024

43 Issues closed by 18 people

• dial tcp: lookup registry.ollama.ai on 127.0.0.53:53: server misbehaving

  #8065 closed Dec 12, 2024

• underscore in the Ollama username and caps in model name while Pushing results into following errors

  #7944 closed Dec 11, 2024

• Idea: Model Pre-Pulling on Startup

  #6815 closed Dec 11, 2024

• Retain existing systemd control file when upgrading ollama versions

  #8048 closed Dec 11, 2024

• ollama CORS check is case-sensitive

  #5838 closed Dec 10, 2024

• I built from source a while back and everything worked ok, the current code is dumping core, bus error.

  #7638 closed Dec 10, 2024

• ollama_llama_server not signed

  #8027 closed Dec 10, 2024

• Suggestion - Custom Model Actions

  #2597 closed Dec 10, 2024

• Issues getting rocm support to compile on Gentoo

  #6857 closed Dec 10, 2024

• Older CUDA compute capability 3.5 and 3.7 support

  #1756 closed Dec 10, 2024

• Missing ROCm Library Files In ollama-linux-amd64-rocm.tgz

  #7817 closed Dec 10, 2024

• Build fails and wipes parent folder if the path to the source contains a directory name with a space

  #7577 closed Dec 10, 2024

• Adding avx2+avx512 to cuda runner in new ollama code

  #7457 closed Dec 10, 2024

• ollama doesn't seem to use my GPU after update

  #7622 closed Dec 10, 2024

• Support GPU runners on CPUs without AVX

  #2187 closed Dec 10, 2024

• Support GPU runners with AVX2

  #2281 closed Dec 10, 2024

• Support additional AVX instruction sets

  #2205 closed Dec 10, 2024

• makefiles should verify compiler before trying to build GPU target

  #7452 closed Dec 10, 2024

• cuda runner fails to build correctly without CUDA_PATH set

  #7491 closed Dec 10, 2024

• Packaging ollama: make including ROCm libraries in the dist optional

  #7483 closed Dec 10, 2024

• Underflow error when using GPU memory overhead

  #8011 closed Dec 10, 2024

• Please add LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct model

  #6269 closed Dec 10, 2024

• EXAONE-3.5 2.4B, 7.8B, and 32B

  #8007 closed Dec 10, 2024

• Incorrect configuration in EXAONE 3.5

  #8021 closed Dec 10, 2024

• go get version mismatch

  #8013 closed Dec 10, 2024

• Empty output from chat-endpoint / non-empty endpoint for non-chat endpoint

  #7795 closed Dec 9, 2024

• Allow for forcing an order of properties in structured JSON response

  #8003 closed Dec 9, 2024

• Not working properly after upgrading to the latest macOS version.

  #7995 closed Dec 9, 2024

• Add option to disable auto-completion

  #8001 closed Dec 9, 2024

• {"error":"json: cannot unmarshal object into Go struct field GenerateRequest.format of type string"}

  #7991 closed Dec 8, 2024

• Incorrect version for v0.5.1 - ollama version is 0.4.4

  #7990 closed Dec 7, 2024

• better diagnosis / error messages when ctx is too small

  #7987 closed Dec 7, 2024

• ggml_cuda_cpy_fn: unsupported type combination (q4_0 to f32) in pre-release version

  #7966 closed Dec 7, 2024

• Format JSON is used when providing `null` to `format` field

  #7977 closed Dec 6, 2024

• PaliGemma 2

  #7968 closed Dec 6, 2024

• Please support the Whisper models. It's a real hassle to run them locally.

  #7976 closed Dec 6, 2024

• Extra command line option on ollama list

  #7943 closed Dec 6, 2024

• How to update ollama desktop on windows?

  #7787 closed Dec 6, 2024

• The parameter 'keep_alive' is invalid when cpu (100%)

  #7645 closed Dec 6, 2024

• this model is not supported by your version of Ollama. You may need to upgrade

  #7946 closed Dec 6, 2024

• Inconsistency between Ollama REST API and CLI Model List causing model accessibility issues

  #7955 closed Dec 5, 2024

• FROM ./vicuna-33b.Q4_0.gguf

  #7959 closed Dec 5, 2024

• Ollama not using GPU (windows)

  #7953 closed Dec 5, 2024

51 Issues opened by 49 people

• have a ques

  #8067 opened Dec 12, 2024

• ollama 0.5.1 is detecting my NVIDIA Tesla M40, but they are not used.

  #8066 opened Dec 12, 2024

• How can I specify the GPU for running the LLM?

  #8064 opened Dec 12, 2024

• Ollama Not Respecting Structured Outputs

  #8063 opened Dec 12, 2024

• llama3.1 tool calling issue with role 'system'

  #8062 opened Dec 12, 2024

• CUSTOM_CPU_FLAGS="" / non avx2 build

  #8058 opened Dec 12, 2024

• where can I find ollama model path?

  #8057 opened Dec 12, 2024

• Create Endpoint Appears Broken

  #8056 opened Dec 12, 2024

• Documentation enhancement Idea - AWS Fargate Infra Implementation

  #8053 opened Dec 11, 2024

• Ollama behind proxy can't pull new models anymore

  #8050 opened Dec 11, 2024

• Can add Stable Diffusion 3.5 model？

  #8047 opened Dec 11, 2024

• Toggle theme

  #8046 opened Dec 11, 2024

• Ollama run hf.co - Error 401: Invalid username or password

  #8045 opened Dec 11, 2024

• I can't use llama3.2 after download.Error: llama runner process has terminated: exit status 0xc0000409

  #8044 opened Dec 11, 2024

• Running in WSL2 seems to be a little bit slow.

  #8043 opened Dec 11, 2024

• Add API endpoint for Ollama server version and feature information

  #8040 opened Dec 11, 2024

• undefined reference to `ggml_backend_cuda_reg'

  #8038 opened Dec 11, 2024

• NOT ABLE TO INSTALL "llama 3.2 model"

  #8035 opened Dec 11, 2024

• nvcc compilation problem -- error: user-defined literal operator not found

  #8033 opened Dec 10, 2024

• OpenAI Chat Completion Client For Multimodal

  #8026 opened Dec 10, 2024

• Ollama run very very slow in ARM cpu (KunPeng 920 CPU)

  #8025 opened Dec 10, 2024

• Ollama is very slow after running for a while

  #8023 opened Dec 10, 2024

• Error reported when importing a multimodal large model of type hugginface (llava-mistral-7b)

  #8022 opened Dec 10, 2024

• Nancy finds security vulnerabilities

  #8020 opened Dec 10, 2024

• access to the Internet

  #8015 opened Dec 9, 2024

• Llama 3.3 still has metadata from Llama 3.1

  #8010 opened Dec 9, 2024

• make is somehow visible that the context size in models is not used by default

  #8009 opened Dec 9, 2024

• Return prompt cache utilization on completion responses

  #8008 opened Dec 9, 2024

• ollama0.5.1, run llama3.3:70b-instruct-q8_0 failed. client connection closed before server finished loading, aborting load

  #8006 opened Dec 9, 2024

• Why is OllamaSetup.exe so large

  #8005 opened Dec 9, 2024

• QwQ 32B Preview: Q4_K_M better than Q8_0 at coding

  #8004 opened Dec 9, 2024

• Structured JSON does not handle arrays at the top level properly

  #8000 opened Dec 8, 2024

• Support loading models from multiple locations

  #7997 opened Dec 8, 2024

• Less available memory than expected

  #7996 opened Dec 8, 2024

• Structured generation cannot handle self referencing (recursion)

  #7993 opened Dec 8, 2024

• GPU no being used.

  #7989 opened Dec 7, 2024

• SVG Icon

  #7988 opened Dec 7, 2024

• Change /delete endpoint to use POST request

  #7985 opened Dec 7, 2024

• llama3.3:70b-instruct-q8_0 generates garbage

  #7984 opened Dec 7, 2024

• exit status 2

  #7982 opened Dec 7, 2024

• Internet Access To The Model

  #7981 opened Dec 7, 2024

• Structured Output is not OpenAI compliant

  #7978 opened Dec 6, 2024

• Pleias

  #7972 opened Dec 6, 2024

• Administrative / silent install is borked

  #7969 opened Dec 6, 2024

• Add stop word <|endoftext|> to qwq models

  #7967 opened Dec 6, 2024

• It seems that the new KV cache quantization feature is incorrectly allocating resources.

  #7965 opened Dec 6, 2024

• Model request: HunyuanVideo text-to-video

  #7958 opened Dec 5, 2024

• Low GPU usage on second GPU

  #7956 opened Dec 5, 2024

• Problems (with nvidia-smi) after upgrading to 0.4.7 (from 0.3 series)

  #7952 opened Dec 5, 2024

• Please can you add logits extraction functionalities ?

  #7951 opened Dec 5, 2024

• panic: failed to decode batch: could not find a kv cache slot goroutine 22 [running]: main.(*Server).run(0xc0000c2120, {0x556536b63ba0, 0xc00008a0a0}) github.com/ollama/ollama/llama/runner/runner.go:344 +0x23e created by main.main in goroutine 1 github.com/ollama/ollama/llama/runner/runner.go:978 +0xcc7

  #7949 opened Dec 5, 2024

87 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Update the /api/create endpoint to use JSON

  #7935 commented on Dec 10, 2024 • 14 new comments

• openai: support include_usage stream option to return final usage chunk

  #6784 commented on Dec 12, 2024 • 4 new comments

• Adding Yacana Multi-Agent Framework to the List of Community Integrations

  #7259 commented on Dec 10, 2024 • 2 new comments

• update default registry domain

  #2483 commented on Dec 10, 2024 • 2 new comments

• Add Vulkan support to ollama

  #5059 commented on Dec 11, 2024 • 1 new comment

• wip: next ollama runner

  #7913 commented on Dec 11, 2024 • 0 new comments

• Don't automatically start on startup / have an option to disable this

  #162 commented on Dec 12, 2024 • 0 new comments

• Ollama should prevent sleep when working.

  #4072 commented on Dec 11, 2024 • 0 new comments

• Enable using llama.cpp's --model-draft <model> feature for speculative decoding

  #5800 commented on Dec 11, 2024 • 0 new comments

• Unsupported Architecture for Vision Model Conversion to GGUF in Ollama

  #7912 commented on Dec 11, 2024 • 0 new comments

• Installing bolt.new and qwen2.5-coder:7b locally (error cudaMalloc failed: out of memory)

  #7896 commented on Dec 11, 2024 • 0 new comments

• AirLLM integration?

  #6294 commented on Dec 11, 2024 • 0 new comments

• tool_choice parameter

  #7778 commented on Dec 11, 2024 • 0 new comments

• MLX backend

  #1730 commented on Dec 11, 2024 • 0 new comments

• Support Radeon RX 5700 XT (gfx1010)

  #2503 commented on Dec 11, 2024 • 0 new comments

• 用qwen2微调训练出的hf转为gguf后，用ollama加载，要指定TEMPLATE才不会胡乱输出

  #6823 commented on Dec 11, 2024 • 0 new comments

• Ollama ps says 22 GB, but nvidia-smi says 16GB with flash attention enabled

  #6160 commented on Dec 11, 2024 • 0 new comments

• Error: unknown error was encountered while running the model GGML_ASSERT(i01 >= 0 && i01 < ne01) failed

  #7441 commented on Dec 10, 2024 • 0 new comments

• add code to enable ollama cli cmd logging , or disable the new ' if not tty exit ' code PLZZ

  #7925 commented on Dec 10, 2024 • 0 new comments

• Instant closure when using shell input with piped output.

  #7820 commented on Dec 10, 2024 • 0 new comments

• Feature Request: Add RSS feed to Blog

  #1669 commented on Dec 10, 2024 • 0 new comments

• Not using GPU

  #7947 commented on Dec 10, 2024 • 0 new comments

• Detecting macOS GPUs when using Podman with GPU passthrough

  #5954 commented on Dec 10, 2024 • 0 new comments

• JSON Schema conformity using Llama.cpp Grammar generation for Tool Calling

  #6002 commented on Dec 10, 2024 • 0 new comments

• The quality of the results returned by the embedding model become worse

  #5495 commented on Dec 8, 2024 • 0 new comments

• server: move /api/version to use http.Handler

  #7884 commented on Dec 6, 2024 • 0 new comments

• feat: Support Moore Threads GPU

  #7554 commented on Dec 12, 2024 • 0 new comments

• Implement tokenize and de-tokenize endpoints

  #7412 commented on Dec 10, 2024 • 0 new comments

• Migrate off centos 7 for intermediate build layers in container image builds

  #7265 commented on Dec 5, 2024 • 0 new comments

• FEAT: add rerank support

  #7219 commented on Dec 10, 2024 • 0 new comments

• example: add example notebook on llm tracing

  #6954 commented on Dec 5, 2024 • 0 new comments

• Expose Tokenize and Detokenize API

  #6586 commented on Dec 10, 2024 • 0 new comments

• AMD integrated graphic on linux kernel 6.9.9+, GTT memory, loading freeze fix

  #6282 commented on Dec 8, 2024 • 0 new comments

• Add API integration tests

  #5678 commented on Dec 12, 2024 • 0 new comments

• Enable AMD iGPU 780M in Linux, Create amd-igpu-780m.md

  #5426 commented on Dec 11, 2024 • 0 new comments

• llm/server.go: Fix ollama ps show 100%GPU even use CPU as runner

  #4906 commented on Dec 11, 2024 • 0 new comments

• Add support for running llama.cpp with SYCL for Intel GPUs

  #2458 commented on Dec 8, 2024 • 0 new comments

• added logprobs (`n_probs`)

  #1640 commented on Dec 11, 2024 • 0 new comments

• add Qwen2-VL

  #6564 commented on Dec 12, 2024 • 0 new comments

• What is "Error: unsupported content type: text/plain; charset=utf-8"?

  #5600 commented on Dec 12, 2024 • 0 new comments

• Reranking models

  #3368 commented on Dec 12, 2024 • 0 new comments

• Connection Error with OllamaFunctions in Langchain

  #2783 commented on Dec 12, 2024 • 0 new comments

• Realtime API like OpenAI (full fledged voice to voice integrations)

  #7514 commented on Dec 12, 2024 • 0 new comments

• MacOS Ollama not binding to 0.0.0.0

  #3581 commented on Dec 12, 2024 • 0 new comments

• Available memory calculation on AMD APU no longer takes GTT into account

  #5471 commented on Dec 8, 2024 • 0 new comments

• Integrated Intel GPU support

  #3113 commented on Dec 8, 2024 • 0 new comments

• GGML_ASSERT(ggml_nelements(a) == ne0*ne1*ne2) failed

  #7590 commented on Dec 8, 2024 • 0 new comments

• Add Tokenize and Detokenize Endpoints to Ollama Server

  #3582 commented on Dec 8, 2024 • 0 new comments

• Getting Unsupported architecture error When Importing Llama-vision.

  #7581 commented on Dec 7, 2024 • 0 new comments

• Linux ollama 0.4.0, 0.4.2, 0.4.5, 0.5.0 custom compile for AMD ROCm fails missing ggml_rocm in go compile

  #7565 commented on Dec 7, 2024 • 0 new comments

• zfs ARC leads to incorrect system memory prediction and refusal to load models that could work

  #5700 commented on Dec 7, 2024 • 0 new comments

• Support for jinaai/jina-embeddings-v3 embedding model

  #6922 commented on Dec 7, 2024 • 0 new comments

• A way to update all downloaded models

  #1890 commented on Dec 6, 2024 • 0 new comments

• Add support for function call (response back) (message.role=tool)

  #7510 commented on Dec 6, 2024 • 0 new comments

• Support model allenai/OLMo-7B

  #2337 commented on Dec 6, 2024 • 0 new comments

• after some time idle / phone standby , getting to the termux ollama run cmd makes it restart the dl from 0

  #7344 commented on Dec 6, 2024 • 0 new comments

• model requires more system memory than is available when useMmap

  #7942 commented on Dec 6, 2024 • 0 new comments

• Adrenalin Edition 24.9.1/24.10.1 slow ollama performance

  #7107 commented on Dec 6, 2024 • 0 new comments

• Deepseek (various) 236b crashes on run

  #7867 commented on Dec 5, 2024 • 0 new comments

• Expose DRY and XTC parameters

  #7504 commented on Dec 5, 2024 • 0 new comments

• Teuken-7b

  #7848 commented on Dec 5, 2024 • 0 new comments

• ollama api（port：11434）Regarding parallel processing and CPU single core usage

  #7928 commented on Dec 5, 2024 • 0 new comments

• Support Qwen VL

  #2874 commented on Dec 5, 2024 • 0 new comments

• Support Mistral's new visual model: Pixtral-12b-240910

  #6748 commented on Dec 5, 2024 • 0 new comments

• Phi3 model starts responding crazy thing after thousand of calls.

  #7931 commented on Dec 5, 2024 • 0 new comments

• allow temperature to be set on command line ( w/out using a modelfile )

  #5362 commented on Dec 10, 2024 • 0 new comments

• Large host RAM allocation when using full gpu offloading

  #7711 commented on Dec 10, 2024 • 0 new comments

• add /metrics endpoint

  #3144 commented on Dec 10, 2024 • 0 new comments

• Does having the default quant type being Q4_0 (a legacy format) on the model hub still make sense?

  #5425 commented on Dec 10, 2024 • 0 new comments

• Multiple ollama_llama_server process are created and then not released

  #7927 commented on Dec 10, 2024 • 0 new comments

• Possibility to remove max retries exceeded when downloading models from a slow connection

  #3162 commented on Dec 10, 2024 • 0 new comments

• Getting Error with OpenAI compatibility

  #7132 commented on Dec 10, 2024 • 0 new comments

• Add support for Intel Arc GPUs

  #1590 commented on Dec 10, 2024 • 0 new comments

• GPU radeon not used

  #7729 commented on Dec 10, 2024 • 0 new comments

• MCP NEEDS ATTENTION!!!

  #7865 commented on Dec 10, 2024 • 0 new comments

• tool parsing issues with "'"

  #7910 commented on Dec 9, 2024 • 0 new comments

• Provide logits or logprobs in the API

  #2415 commented on Dec 9, 2024 • 0 new comments

• Missing logprob

  #1259 commented on Dec 9, 2024 • 0 new comments

• Feature Request: Support logprobs before GTA 6 comes out

  #6880 commented on Dec 9, 2024 • 0 new comments

• Integrated AMD GPU support

  #2637 commented on Dec 9, 2024 • 0 new comments

• signal arrived during cgo execution

  #7941 commented on Dec 9, 2024 • 0 new comments

• OLMo-2-1124-13B & 7B

  #7863 commented on Dec 9, 2024 • 0 new comments

• Pulling Multiple Models at Once

  #4351 commented on Dec 9, 2024 • 0 new comments

• Document how to list models in the registry

  #286 commented on Dec 8, 2024 • 0 new comments

• Add ability to listen on unix socket

  #739 commented on Dec 8, 2024 • 0 new comments

• Fails to build on macOS with "fatal error: {'string','cstdint'} file not found"

  #7392 commented on Dec 8, 2024 • 0 new comments

• Ollama push got `retrieving manifest Error: file does not exist`

  #3501 commented on Dec 8, 2024 • 0 new comments