1 Release published by 1 person

• v0.5.5

  published Jan 8, 2025

9 Pull requests merged by 6 people

• make the modelfile path relative for ollama create

  #8380 merged Jan 11, 2025

• chore: upgrade to gods v2

  #8196 merged Jan 10, 2025

• Mxyng/next llama

  #8194 merged Jan 10, 2025

• merge llama/ggml into ml/backend/ggml

  #7957 merged Jan 10, 2025

• better client error for /api/create

  #8351 merged Jan 9, 2025

• OpenAI: accept additional headers to fix CORS error #8342

  #8343 merged Jan 8, 2025

• llama: update vendored code to commit 46e3556

  #8308 merged Jan 8, 2025

• readme: add phi4 model

  #8350 merged Jan 8, 2025

• Add CUSTOM_CPU_FLAGS to Dockerfile.

  #8284 merged Jan 6, 2025

8 Pull requests opened by 8 people

• Add Safetensor Conversion for Granite Models

  #8319 opened Jan 6, 2025

• readme: add Reins to community integrations

  #8334 opened Jan 7, 2025

• cli/run: add --system and --parameter CLI options

  #8340 opened Jan 7, 2025

• Add LangChain for .NET to libraries list

  #8352 opened Jan 8, 2025

• feat(install): use curl -C to continue download

  #8359 opened Jan 9, 2025

• Explicit mention `ollama serve` will start a server, friendly for new users

  #8381 opened Jan 10, 2025

• fix: add rootless container support

  #8383 opened Jan 11, 2025

• add new create api doc

  #8388 opened Jan 12, 2025

79 Issues closed by 28 people

• Run phi4 failed

  #8370 closed Jan 13, 2025

• Llava embeddings: string into Go value of type api.EmbeddingRequest

  #5861 closed Jan 13, 2025

• Error: max retries exceeded for all ollama model pulls (read: connection reset by peer)

  #8167 closed Jan 13, 2025

• More API compatibility

  #8251 closed Jan 13, 2025

• Ollama v0.5.4 not response with stream mode when submit tool option

  #8278 closed Jan 13, 2025

• Ollama cannot load model after several hours on some GPUs

  #8276 closed Jan 13, 2025

• Ollama models give low inference with Continue extension on VS Code Community Edition.

  #8272 closed Jan 13, 2025

• Enhanced System Observability for Multi-Server Environments (Unified Endpoints?)

  #8247 closed Jan 13, 2025

• Significant Performance Differences of ARM64 Installation Package Across Different CPU Architectures

  #8246 closed Jan 13, 2025

• Error: open config.json: file does not exist

  #8245 closed Jan 13, 2025

• ollama._types.ResponseError: llama runner process has terminated: signal: broken pipe

  #8216 closed Jan 13, 2025

• Ollama hangs when running llama3.2 and llama3.2:1b

  #8200 closed Jan 13, 2025

• Check Available Memory Before Downloading

  #8192 closed Jan 13, 2025

• mllama doesn't support parallel requests yet - llama3.2-vision:11b for Standard_NC24ads_A100_v4

  #8186 closed Jan 13, 2025

• Falcon3 10B in 1.58bit format

  #8184 closed Jan 13, 2025

• How do I specify specific GPUs when running a model?

  #8183 closed Jan 13, 2025

• {"error":"POST predict: Post \"http://127.0.0.1:33603/completion\": EOF"}

  #8182 closed Jan 13, 2025

• LLAMA 3:70B is crashing inside K8s pods

  #8179 closed Jan 13, 2025

• c4ai-command-r7b-12-2024

  #8104 closed Jan 13, 2025

• ollama connection error on windows

  #8075 closed Jan 13, 2025

• Ollama behind proxy can't pull new models anymore

  #8050 closed Jan 13, 2025

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

  #8025 closed Jan 13, 2025

• Ollama is very slow after running for a while

  #8023 closed Jan 13, 2025

• Web browsing

  #8015 closed Jan 13, 2025

• Performance decline

  #7919 closed Jan 13, 2025

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

  #7896 closed Jan 13, 2025

• Proper way to train model on my data and load into Ollama?

  #7755 closed Jan 13, 2025

• Nvidia fallback memory

  #7584 closed Jan 13, 2025

• Despite advertised, granite3-dense does not seem to support tools.

  #7544 closed Jan 13, 2025

• mac Errors when running

  #7495 closed Jan 13, 2025

• Error occurred: Error code: 400 - {'error': {'message': 'unexpected EOF', 'type': 'invalid_request_error', 'param': None, 'code': None}}

  #7277 closed Jan 13, 2025

• Add Tab-Enabled Autocomplete for Local Model Parameters in Ollama CLI

  #7239 closed Jan 13, 2025

• OLMo-2-1124-13B & 7B

  #7863 closed Jan 12, 2025

• Ollama not completing chat request

  #8387 closed Jan 12, 2025

• API is not giving up memory after responce is complete

  #8390 closed Jan 12, 2025

• Hangs after 20-30 mins, a perdiocal restart of the ollama service is required

  #6380 closed Jan 12, 2025

• How can we make model calls faster

  #4493 closed Jan 12, 2025

• Setting up HTTP Server Timeouts / Connection Management

  #3225 closed Jan 12, 2025

• Connection Error with OllamaFunctions in Langchain

  #2783 closed Jan 12, 2025

• Error: llama runner process has terminated: exit status 2

  #8382 closed Jan 11, 2025

• Unable to run llama on IPv6 Single Stack env

  #8346 closed Jan 11, 2025

• GPU runs at maximum load with 2 models

  #8285 closed Jan 11, 2025

• FROM path resolution uses working directory instead of Modelfile location

  #8353 closed Jan 11, 2025

• Allow context to be set from the command line.

  #8356 closed Jan 11, 2025

• Faiss document dependency error

  #8357 closed Jan 11, 2025

• Add tools support to dolphin3

  #8364 closed Jan 10, 2025

• deepseek v3

  #8366 closed Jan 10, 2025

• Ollama run don't work with recent llama version (due to a upgrade trouble ?)

  #8375 closed Jan 10, 2025

• ollama rm xxx failed to delete file /usr/share/ollama/.ollama/models/blobs/sha256-xxx

  #8373 closed Jan 10, 2025

• different between Modelfile PARAMETER and API

  #8374 closed Jan 10, 2025

• Single json expected when streaming set to false

  #8367 closed Jan 10, 2025

• type:file_url not supported?

  #8369 closed Jan 10, 2025

• llama3.1-8B doesn't utilize my gpu

  #8361 closed Jan 10, 2025

• Considerably slower response via API than terminal

  #8295 closed Jan 9, 2025

• can't pull model from hf

  #8360 closed Jan 9, 2025

• Fine-tuned Qwen2.5-Instruct isn't supported as expectation

  #8132 closed Jan 9, 2025

• Add Llama-3.1-Nemotron-70B-Instruct-HF

  #8358 closed Jan 9, 2025

• API Hangs

  #8348 closed Jan 9, 2025

• Ollama structured outputs not working on Windows

  #8338 closed Jan 8, 2025

• Phi-4 support

  #8347 closed Jan 8, 2025

• Error: listen tcp 127.0.0.1:11434: bind: Only one usage of each socket address

  #8261 closed Jan 8, 2025

• How to use the locally downloaded model in Ollama？

  #8300 closed Jan 8, 2025

• Magnet download

  #8275 closed Jan 8, 2025

• Function/Tool Call doesn't always work, is it memory dependent?

  #8344 closed Jan 8, 2025

• Cannot get a tool call and a message in the same response

  #8337 closed Jan 8, 2025

• Llama3.3:70b-intstruct-q5_K_M got EOF on MacMini M4 pro 64GB RAM

  #8331 closed Jan 8, 2025

• Is there a better model that can accurately recognize image information?下载了好几个多模态的模型，图片识别效果都不好

  #6949 closed Jan 8, 2025

• On multi-GPU inference speed limited by performance of single CPU core

  #6913 closed Jan 7, 2025

• LLava:13B Model Outputting ############### After Period of Inactivity

  #6825 closed Jan 7, 2025

• reader-lm - heavy hallucinations?

  #6875 closed Jan 7, 2025

• Rerank models.... WHERE ARE THEY???????????

  #8336 closed Jan 7, 2025

• Missing logprob

  #1259 closed Jan 7, 2025

• Feature Request: Support logprobs before GTA 6 comes out

  #6880 closed Jan 7, 2025

• Invalid JSON Schema when using pydantic pattern

  #8325 closed Jan 7, 2025

• llama3.2-vision // llava - not receiving images via chat completions API

  #8322 closed Jan 7, 2025

• qwq instance crash

  #8309 closed Jan 7, 2025

• yi-coder: Suffix not supported

  #8320 closed Jan 6, 2025

• Standard Linux install includes CUDA libraries even if unused

  #8234 closed Jan 6, 2025

• context limit from user settings not actually applied

  #8318 closed Jan 6, 2025

40 Issues opened by 38 people

• Image pulling behind proxy index out of range

  #8400 opened Jan 13, 2025

• unable to use nvidia GPU & how to fix

  #8399 opened Jan 13, 2025

• Cancelling model loading for one requests cancels it for others

  #8398 opened Jan 13, 2025

• [UNK_BYTE_…] Output with gemma-2b-it in Ollama

  #8397 opened Jan 12, 2025

• Error: could not connect to ollama app, is it running?

  #8396 opened Jan 12, 2025

• Empty response via API

  #8395 opened Jan 12, 2025

• The same model could load all onto the GPU last year. Today, after upgrading to ollama, I found that it cannot be loaded onto the GPU all at once.

  #8394 opened Jan 12, 2025

• Unable to enable GPU for models

  #8393 opened Jan 12, 2025

• Empty 'assistant' message

  #8392 opened Jan 12, 2025

• Add MoonDream 2 rev:2025-1-9 support

  #8391 opened Jan 12, 2025

• Ollama install script relaces the systemd profile

  #8389 opened Jan 12, 2025

• Return in a response a flag if the input request was truncated

  #8386 opened Jan 11, 2025

• Cannot list or install models: Connection refused error on Windows 10

  #8385 opened Jan 11, 2025

• Unable to acess ollama model hosted on a raspberry pi 5 from an other device

  #8384 opened Jan 11, 2025

• Inquiry About Adding vikhyatk/moondream2 to Ollama

  #8379 opened Jan 10, 2025

• a lot of CUDA errors and CUDA error: out of memory and SIGSEGV: segmentation violation eventhough VRAM memory still available

  #8377 opened Jan 10, 2025

• Ollama version doesn't properly truncate tokens to 512 max for official snowflake-arctic-embed-l model

  #8376 opened Jan 10, 2025

• GPU using

  #8372 opened Jan 10, 2025

• ollama not working

  #8371 opened Jan 10, 2025

• Support Code Actions

  #8368 opened Jan 9, 2025

• When I use multiple GPUs, the utilization is very low.How can I configure it to maximize GPU utilization and reduce the reasoning time?

  #8365 opened Jan 9, 2025

• After fine-tuning llama3.1 and deploying it to ollama, the inference may not stop, or the output may be repeated, or the output may be gibberish.

  #8363 opened Jan 9, 2025

• please add model:QVQ-Preview 72B!

  #8362 opened Jan 9, 2025

• we need Ollama Video-LLaVA

  #8355 opened Jan 9, 2025

• Dynamic context size in OpenAI API compatibility.

  #8354 opened Jan 9, 2025

• can't use gpu after update

  #8349 opened Jan 8, 2025

• llama3.2-vision does not support tools

  #8345 opened Jan 8, 2025

• CORS error x-stainless-helper-method

  #8342 opened Jan 8, 2025

• [feature] start ollama automatically on startup

  #8341 opened Jan 7, 2025

• `GIN_MODE` not able to set to `release`

  #8339 opened Jan 7, 2025

• Make flash attention configurable via UI or enable by default

  #8335 opened Jan 7, 2025

• GOT-OCR and voice model support

  #8333 opened Jan 7, 2025

• Allow set the type of K/V cache separately

  #8332 opened Jan 7, 2025

• Using the Ollama 0.5.4 will cause the pull progress to decrease instead of increase.

  #8330 opened Jan 7, 2025

• dolphin3 template doesn't support tools

  #8329 opened Jan 7, 2025

• [Model request] alea-institute/kl3m-003-3.7b

  #8328 opened Jan 7, 2025

• I/O error on POST request for "http://localhost:11434/v1/chat/completions\

  #8327 opened Jan 7, 2025

• Error: pull model manifest: 400: The specified repository contains sharded GGUF. Ollama does not support this yet.

  #8326 opened Jan 7, 2025

• Add a CUDA+AVX2(VNNI) runner to the Docker image.

  #8324 opened Jan 6, 2025

• Manual linux install: runners/cuda_v11_avx/ollama_llama_server: undefined symbol: ggml_backend_cuda_reg

  #8323 opened Jan 6, 2025

69 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.

• Runner for OIlama engine

  #8301 commented on Jan 11, 2025 • 2 new comments

• examples: remove codified examples

  #8267 commented on Jan 13, 2025 • 2 new comments

• Remove redundant image_url format check in fromChatRequest

  #8232 commented on Jan 13, 2025 • 1 new comment

• openai: fix follow-on messages having "role": "assistant"

  #7722 commented on Jan 6, 2025 • 1 new comment

• Corrupt output on multiple GPU in Windows 11

  #8188 commented on Jan 13, 2025 • 0 new comments

• Deepseek (various) 236b crashes on run

  #7867 commented on Jan 13, 2025 • 0 new comments

• Please don’t clutter the user home directory

  #228 commented on Jan 13, 2025 • 0 new comments

• Error: could not connect to ollama app, is it running?

  #7524 commented on Jan 13, 2025 • 0 new comments

• Rate limit download speed on pulling new models

  #2006 commented on Jan 12, 2025 • 0 new comments

• Download slows to a crawl at 99%

  #1736 commented on Jan 12, 2025 • 0 new comments

• Support model allenai/OLMo-7B

  #2337 commented on Jan 12, 2025 • 0 new comments

• Does ollama support accelerated running on npu?

  #3004 commented on Jan 12, 2025 • 0 new comments

• Support partial loads of LLaMA 3.2 Vision 11b on 6G GPUs

  #7509 commented on Jan 12, 2025 • 0 new comments

• llama3.2-vision doesn't utilize my GPU.

  #8310 commented on Jan 12, 2025 • 0 new comments

• Qwen/Qwen2.5-Math

  #6889 commented on Jan 11, 2025 • 0 new comments

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

  #162 commented on Jan 10, 2025 • 0 new comments

• Support hot-plugging exernal GPU(s)

  #5411 commented on Jan 10, 2025 • 0 new comments

• The <toolcall> in nemotron-mini. Again.

  #8287 commented on Jan 10, 2025 • 0 new comments

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

  #5362 commented on Jan 9, 2025 • 0 new comments

• Way to stop all running models

  #6987 commented on Jan 9, 2025 • 0 new comments

• error looking up nvidia GPU memory - intermittent "cuda driver library failed to get device context 800"

  #6928 commented on Jan 13, 2025 • 0 new comments

• DeepSeek v3

  #8249 commented on Jan 13, 2025 • 0 new comments

• Add Vulkan support to ollama

  #5059 commented on Jan 12, 2025 • 0 new comments

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

  #6282 commented on Jan 6, 2025 • 0 new comments

• openai: increase context window when max_tokens is provided

  #6504 commented on Jan 9, 2025 • 0 new comments

• cmd: add "stop all" to stop all running models

  #7209 commented on Jan 9, 2025 • 0 new comments

• feat: Support Moore Threads GPU

  #7554 commented on Jan 10, 2025 • 0 new comments

• wip: next ollama runner

  #7913 commented on Jan 11, 2025 • 0 new comments

• wip: next ollama runner build updates

  #7954 commented on Jan 11, 2025 • 0 new comments

• Support Unix sockets for OLLAMA_HOST

  #8072 commented on Jan 12, 2025 • 0 new comments

• Update linux.md

  #8087 commented on Jan 8, 2025 • 0 new comments

• Examples(privategpt): update outdated langchain dependencies and add UnstructuredRSTLoader

  #8117 commented on Jan 13, 2025 • 0 new comments

• cuda: adjust variant based on detected runners

  #8121 commented on Jan 7, 2025 • 0 new comments

• fix: only add to history if different

  #8197 commented on Jan 10, 2025 • 0 new comments

• create a default, non-root user for the container image

  #8259 commented on Jan 13, 2025 • 0 new comments

• AirLLM integration?

  #6294 commented on Jan 8, 2025 • 0 new comments

• MultiGPU ROCm

  #8206 commented on Jan 7, 2025 • 0 new comments

• add /metrics endpoint

  #3144 commented on Jan 7, 2025 • 0 new comments

• Warn users when input is longer than supported context

  #7043 commented on Jan 7, 2025 • 0 new comments

• version aware linux upgrade

  #8233 commented on Jan 7, 2025 • 0 new comments

• cudart initialization failure on ppc64le (AlmaLinux/RHEL8) with ollama-0.1.31-1.1

  #8151 commented on Jan 7, 2025 • 0 new comments

• Support to Intel NPU by Intel NPU Acceleration Library

  #5747 commented on Jan 7, 2025 • 0 new comments

• Integrated AMD GPU support

  #2637 commented on Jan 7, 2025 • 0 new comments

• Reverse download Progress During Model Pull Ollama pull phi3:medium

  #8280 commented on Jan 7, 2025 • 0 new comments

• Update context window management to avoid context shifts

  #3176 commented on Jan 7, 2025 • 0 new comments

• launchctl setenv OLLAMA_HOST "0.0.0.0" does not work in version 0.3.14

  #7331 commented on Jan 6, 2025 • 0 new comments

• ollama : /usr/lib64/libstdc++.so.6: version GLIBCXX_3.4.25 not found - Kylin Linux glibc++ version incompatible with official builds

  #8085 commented on Jan 6, 2025 • 0 new comments

• Ollama Not Respecting Structured Outputs

  #8063 commented on Jan 6, 2025 • 0 new comments

• Feature Request: Limit Ollama to use only GPU RAM (VRAM), not system RAM

  #8316 commented on Jan 6, 2025 • 0 new comments

• Error: llama runner process has terminated: error:/opt/rocm/lib/libhipblas.so.2: undefined symbol: rocblas_sgbmv_64

  #8108 commented on Jan 6, 2025 • 0 new comments

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

  #6922 commented on Jan 6, 2025 • 0 new comments

• `digest mismatch` on download

  #941 commented on Jan 6, 2025 • 0 new comments

• Allow reading from file while in `ollama run` prompt

  #2305 commented on Jan 9, 2025 • 0 new comments

• Internal error at url manifests/sha256:

  #5298 commented on Jan 9, 2025 • 0 new comments

• Unable to bind 0.0.0.0 port to all interfaces, except localhost

  #8304 commented on Jan 9, 2025 • 0 new comments

• Ollama: running Vite in production mode fails

  #4516 commented on Jan 9, 2025 • 0 new comments

• Pulling models from private OCI Registries

  #7244 commented on Jan 9, 2025 • 0 new comments

• Locally-hosted library

  #914 commented on Jan 9, 2025 • 0 new comments

• Some APIs in registry.ollama returns 404

  #5021 commented on Jan 9, 2025 • 0 new comments

• Ollama only using half of available CPU cores with NUMA multi-socket systems

  #2929 commented on Jan 9, 2025 • 0 new comments

• COULDN'T run qwen2.5-7b-instuct-q4_k on cpu; error wsarecv: An existing connection was forcibly closed by the remote host.

  #8317 commented on Jan 9, 2025 • 0 new comments

• mac app silently fails to install CLI link if /usr/local/bin/ missing

  #4882 commented on Jan 9, 2025 • 0 new comments

• Ollama is not using the 100% of RTX4000 VRAM (18 of 20GB)

  #3078 commented on Jan 8, 2025 • 0 new comments

• Enable auto-save functionality via CLI flag

  #8288 commented on Jan 8, 2025 • 0 new comments

• Runing ollama on Intel Ultra NPU or GPU

  #8281 commented on Jan 8, 2025 • 0 new comments

• Ollama should avoid calling hallucinated tools

  #8294 commented on Jan 8, 2025 • 0 new comments

• Ollama - Gentoo Linux support

  #8293 commented on Jan 8, 2025 • 0 new comments

• Speed ten times slower than llamafile

  #8305 commented on Jan 8, 2025 • 0 new comments

• ollama should use `/usr/local` to store models in Linux

  #8297 commented on Jan 8, 2025 • 0 new comments