Only the last token's processing time is included in prompt processing,
giving an artificially high rate. In addition, the number of tokens
only included the tokens that miss the cache, instead of our historic
total tokens.
Currently, context length is unbounded - the cache will keep
growing forever independent of the model's trained context
length. This caps it and enforces semantics similar to most
cloud services:
- Long prompts will result in an error, not truncation.
- Generation that exceeds the context will be stopped
Errors that occur during pipeline processing are currently only
logged but not sent back to the client. Rather than using HTTP
status codes as we have historically done, this serializes errors
as messages to allow sending them at any time during the stream.
The MLX runner previously reported a static VRAM estimate that was
computed at load time and consisted only of the weights. This is
strictly less than the actual memory usage, as it does not include
the KV cache or compute graph.
When the entire prompt was already cached (e.g. repeated prompt),
findRemaining returned an empty slice, causing FromValues to panic
on an index-out-of-range accessing a zero-length byte slice.
Fix by always keeping at least one token to re-evaluate so the
pipeline can seed token generation. Also reject empty prompts
early rather than panicking.
Currently, a canceled request can result in computation continuing
in the background to completion. It can also trigger a deadlock
when there is nobody to read the output tokens and the pipeline
cannot continue to the next request.
Particularly in error cases, it can be difficult to ensure that
all pinned memory is unpinned, MLX buffers are released and cache
state is consistent. This encapsulates those pieces and sets up
proper deferrals so that this happens automatically on exit.
Pass subprocess stdout/stderr through to the parent's stderr directly
instead of re-wrapping each line with slog. The subprocess already
writes structured slog output, so the re-wrapping produced nested
timestamps, levels, and message fields that were hard to read.
Also downgrade verbose KV cache debug logs to trace level.
The KV cache previously used a tree structure which could
store multiple divergent sequences, which is good for cache
reuse. However, this is typically used in conjunction with
paged attention so each node in the tree can store just a
chunk of the KV cache and they can be stitched together later.
We don't currently do this, so the cache was storing copies of
the full cache for each past sequence.
This redundancy plus the lack of resource limits, caused significant
memory use as a conversation grew. Instead, this changes to store
a single entry for the cache, which can be prefix matched. Although
it is less ideal for multiple users, it largely matches Ollama's
current behavior. It can be improved as additional pieces are fleshed
out.
The previous approach tracked array lifecycles through reference
counting, where each array recorded its inputs and a reference count
that was decremented as dependents were freed. This is not really
necessary as MLX tracks references internally. It is also error
prone as it is easy to create new arrays and forget to free them
when the Go variable goes out of scope.
Instead, we can pin just the arrays we want (typically outputs and
specific intermediates, like the cache). All other arrays are freed
by default when we run sweep. This avoids most causes of memory leaks
while still giving the freedom to save what we want.
The recent change in #14322 added tryLoadByName() which attempts to
load libmlxc.dylib via rpath before searching directories. This is an
optimization for Homebrew installations where rpath is correctly set.
However, when rpath isn't set (which is the common case for app bundle
installations), dlopen fails and the CHECK macro prints an error to
stderr:
ERROR - dynamic.c:21 - CHECK failed: handle->ctx != NULL
This error is misleading because it's an expected failure path - the
code correctly falls back to searching the executable directory and
loads the library successfully. The error message causes user confusion
and makes it appear that something is broken.
Replace the CHECK macro with a simple return code so the C code fails
silently. The Go code already handles error logging appropriately:
tryLoadByName() fails silently (intentional fallback), while
tryLoadFromDir() logs via slog.Error() when explicit path loading fails.
This change adds a new x/tokenizer package which includes:
* New BPE and SentencePiece tokenizers
* Removing the dependency on the imagegen tokenizers
* Fixes to multibyte decoding in the pipeline
* Various correctness and benchmark tests
Not included in this PR is the WordPiece tokenizer for BERT models which will be
added when we add embedding models. The imagegen tokenizers will also be removed in
a follow-up PR.
The existing code manually searches directories for libmlxc.* and passes
full paths to dlopen, bypassing the binary's rpath. This means MLX
libraries installed via package managers (e.g., Homebrew) aren't found
even when rpath is correctly set at link time.
This change adds a fallback that tries loading via rpath first (using
just the library name), before falling back to the existing directory
search. This follows standard Unix/macOS conventions and works with any
installation that sets rpath.
Fixes library loading on macOS with Homebrew-installed mlx-c without
requiring OLLAMA_LIBRARY_PATH environment variable.
Co-authored-by: Natl <nat@MacBook-Pro.local>
* add ability to disable cloud
Users can now easily opt-out of cloud inference and web search by
setting
```
"disable_ollama_cloud": true
```
in their `~/.ollama/server.json` settings file. After a setting update,
the server must be restarted.
Alternatively, setting the environment variable `OLLAMA_NO_CLOUD=1` will
also disable cloud features. While users previously were able to avoid
cloud models by not pulling or `ollama run`ing them, this gives them an
easy way to enforce that decision. Any attempt to run a cloud model when
cloud is disabled will fail.
The app's old "airplane mode" setting, which did a similar thing for
hiding cloud models within the app is now unified with this new cloud
disabled mode. That setting has been replaced with a "Cloud" toggle,
which behind the scenes edits `server.json` and then restarts the
server.
* gate cloud models across TUI and launch flows when cloud is disabled
Block cloud models from being selected, launched, or written to
integration configs when cloud mode is turned off:
- TUI main menu: open model picker instead of launching with a
disabled cloud model
- cmd.go: add IsCloudModelDisabled checks for all Selection* paths
- LaunchCmd: filter cloud models from saved Editor configs before
launch, fall through to picker if none remain
- Editor Run() methods (droid, opencode, openclaw): filter cloud
models before calling Edit() and persist the cleaned list
- Export SaveIntegration, remove SaveIntegrationModel wrapper that
was accumulating models instead of replacing them
* rename saveIntegration to SaveIntegration in config.go and tests
* cmd/config: add --model guarding and empty model list fixes
* Update docs/faq.mdx
Co-authored-by: Jeffrey Morgan <jmorganca@gmail.com>
* Update internal/cloud/policy.go
Co-authored-by: Jeffrey Morgan <jmorganca@gmail.com>
* Update internal/cloud/policy.go
Co-authored-by: Jeffrey Morgan <jmorganca@gmail.com>
* Update server/routes.go
Co-authored-by: Jeffrey Morgan <jmorganca@gmail.com>
* Revert "Update internal/cloud/policy.go"
This reverts commit 8bff8615f9.
Since this error shows up in other integrations, we want it to be
prefixed with Ollama
* rename cloud status
* more status renaming
* fix tests that weren't updated after rename
---------
Co-authored-by: ParthSareen <parth.sareen@ollama.com>
Co-authored-by: Jeffrey Morgan <jmorganca@gmail.com>
This change fixes an issue where GGML based models (for either the Ollama runner or
the legacy llama.cpp runner) would try to load the mlx library. That would panic
and the model fails to start.
This change adds a new MLX based runner which includes:
* Method-based MLX bindings
* Subprocess-based MLX runner (x/mlxrunner)
* KV cache with tree management
* A basic sampler
The GLM4-MoE-Lite model has been ported to use the new bindings.
---------
Co-authored-by: Michael Yang <git@mxy.ng>
This change includes:
- changes to the safetensors metadata format
- changes to the create command to properly create the blobs with the new format
- changes to load the new format
- fixes ollama show to properly show each tensor
When context length is clamped to the model's trained context length,
ollama ps now shows the actual clamped value instead of the originally
configured value.
- Fix panic in ollama show for image gen models (safe type assertion)
- Add vision capability for Flux2KleinPipeline models at create time
- Flatten transparent PNG images onto white background for better results
Remove static VRAM estimation (EstimateVRAM, CheckMemoryRequirements)
which wasn't helpful. Instead, report the actual tensor weight size
from the manifest for ollama ps.
- Remove memory estimation check from runner startup
- Remove EstimateVRAM, CheckMemoryRequirements, modelVRAMEstimates
- Add TotalTensorSize() to get actual weight size from manifest
- Use weight size for Server.vramSize instead of estimates
Note: This is better than showing 0 or inaccurate estimates, but the
weight size is a drastic underestimation of actual memory usage since
it doesn't account for activations, intermediate tensors, or MLX
overhead. Future work should query real-time memory from MLX
(e.g., MetalGetActiveMemory) for accurate reporting.
Remove the Qwen image generation and image editing model packages
to clean up the codebase. These models will be reintroduced later.
- Delete x/imagegen/models/qwen_image/ (10 files)
- Delete x/imagegen/models/qwen_image_edit/ (5 files)
- Remove related CLI flags and imports from cmd/engine/main.go
- Update comments in cache/step.go to remove Qwen-specific references
Add --quantize fp4 support to ollama create for image generation models
(flux2, z-image-turbo), using MLX's affine 4-bit quantization.
Changes:
- Add fp4 to validation in CreateImageGenModel
- Add FP4 case to quantizeTensor (group_size=32, bits=4, affine mode)
- Add GetQuantization() to WeightSource interface for dynamic params
- Update LoadLinearLayer to use quantization params from model metadata
* MLX - dynamic loading of mlx-c
Create a wrapper layer to indirect the dependency on mlx-c so
the main ollama binary does not have a load-time dependency on mlx-c, mlx, and on linux, cuda. Lazy load the library via dlopen
so we can adjust the path to ensure the dependencies are found
and fail gracefully if not present.
* review comments
* fix broken tests
* x: make `ollama create --experimental` import from safetensors
This change allows pulling in safetensors models into the new experimental model format, and also
fixes the `ollama show` command to be able to correctly display the model information.
* gofumpt the linter
* gofumpt the linter again
* validate the model name
- Install mlx.metallib for arm64 builds (required for Metal GPU acceleration)
- Apply rpath settings to all macOS builds, not just x86_64
- Add CMAKE_BUILD_WITH_INSTALL_RPATH to avoid install_name_tool errors
- Update build_darwin.sh to copy, sign, and package the metallib
