GitLab

We don't check the return status after computing the graph, which
can silently lead to bad outputs if we try to keep going and future
computation succeeds. This appears to happens in certain cases on
Apple M2 devices.

Fixes #11070

Verified these fail on 0.9.1 and pass on HEAD.

Removes a test under benchmark/ that is unused

Reverts PR #11115. The original change was mistakingly reverted instead of #10822

This reverts commit aaa78180.

This reverts commit 6b04cad7.

Fixes issue where tool calls that don't expect any parameters were
not being parsed. This also fixes two additional issues: one where
2+ tool calls would not be correctly parsed, and cases where tool calls
with invalid parameters would still get parsed

* ggml: test write gguf order
* ggml: fix write tensor order

* incremental gguf parser
* gguf: update test to not rely on gguf on disc
* re-use existing create gguf
* read capabilities from gguf kv
* kv exists
* update tests
* s/doneFunc/successFunc/g
* new buffered reader

---------
Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

The current splitDim function only operates on tensors that are split evenly which isn't always the case, e.g. a QKV tensor. This change allows the function to be used for arbitrary splits

if tokenizer.json is already copied, skip tokenizer.model

while nn.Linear.Forward isn't applicable for sparse MLP, it's still
a nice container for the tensors

When a user elects to keep the existing app, the
new Ollama is named `Ollama 2.app`
This fixes the app startup flow to handle this naming pattern.

Give the desktop app a hint to start fast.

This reverts commit 09430011.

When starting the app in the background, start it hidden.

Fix an array out of bounds crash

move thinking logic into its own package

export ThinkingParser

This enables matching up devices and information reported by the backend
with system management libraries such as nvml to get accurate free
memory reporting.

"POST predict" basically means that the runner has crashed, which
can have many reasons. However, many people think this is a specific
error and either report only this message or group together unrelated
bugs. This replaces it with a more friendly and helpful message.

- Both `/api/generate` and `/api/chat` now accept a `"think"`
  option that allows specifying whether thinking mode should be on or
  not
- Templates get passed this new option so, e.g., qwen3's template can
  put `/think` or `/no_think` in the system prompt depending on the
  value of the setting
- Models' thinking support is inferred by inspecting model templates.
  The prefix and suffix the parser uses to identify thinking support is
  also automatically inferred from templates
- Thinking control & parsing is opt-in via the API to prevent breaking
  existing API consumers. If the `"think"` option is not specified, the
  behavior is unchanged from previous versions of ollama
- Add parsing for thinking blocks in both streaming/non-streaming mode
  in both `/generate` and `/chat`
- Update the CLI to make use of these changes. Users can pass `--think`
  or `--think=false` to control thinking, or during an interactive
  session they can use the commands `/set think` or `/set nothink`
- A `--hidethinking` option has also been added to the CLI. This makes
  it easy to use thinking in scripting scenarios like
  `ollama run qwen3 --think --hidethinking "my question here"` where you
  just want to see the answer but still want the benefits of thinking
  models

If OLLAMA_AUTH is set, sign each request w/ a timestamp and pass the signature in the token header

Computing an attention mask for a large context and max batch is
expensive - over 100ms. Models like Gemma3 that have multiple types
of caches and custom attention masks need to do this 4 times, so this
adds approximately 500ms to startup time when using 128k context

When we are reserving the worst case graph, we don't need the mask,
only its shape, so we can skip this.

Signed-off-by: Kyle Steere <kyle.steere@chainguard.dev>