Abhijit Ramesh [Tue, 3 Mar 2026 03:35:11 +0000 (19:35 -0800)]
ggml webgpu: fix workgroup dispatch limit for large batch sizes (#19965)
* ggml-webgpu: fix workgroup dispatch limit for large batch sizes
WebGPU limits workgroup sizes to 65535 per dimension. Large MUL_MAT
operations with batch sizes exceedeing this limi would fail.
* add compute_2d_workgroups() helper to split total workgroup ID across
X/Y dimensions
* update mul_mat_reg_tile.wgsl to reconstruct linear workgroup ID from 2D
dispatch
* update mul_mat_subgroup_matrix.wgsl to reconstruct linear workgroup ID
from 2D dispatch
* update mul_mat.wgsl to compute global index from 2D workgroup
coordinates
* refactor all three mul_mat dispatch paths to use the shared helper
* ggml-webgpu: add bounds checking for over-dispatched workgroups
2D workgroup dispatch can over-dispatch when total workgroups don't
divide evenly into the 65535 per-dimension limit. Extra workgroups
would compute invalid batch indices, causing memory corruption.
* add batch_idx bound check to mul_mat_reg_tile.wgsl and
mul_mat_subgroup_matrix.wgsl to prevent over-dispatched workgroups
from accessing invalid memory
* fixes test failures with large batch sizes (eg., bs=[128, 1024])
* ggml-webgpu: add back TODO for spliting large sizes into batches
* address review: replace FATTN_WARP_SIZE with constexpr, improve dispatch
- Replace #define FATTN_WARP_SIZE with constexpr int warp_size =
ggml_cuda_get_physical_warp_size() in each device function
- Use ne[1]*gqa_ratio threshold for MMA vs tile dispatch. Benchmarked
crossover on MI300X @ d32768 with power-of-2 GQA models:
hsk=64 (Llama 1B, gqa=4): MMA wins at eff >= 128 (+11%)
hsk=128 (Llama 3B, gqa=4): MMA wins at eff >= 128 (+4%)
Unified threshold: eff_nq >= 128 for all head sizes.
- Remove VEC fallback; small batches fall through to tile kernel
* Update ggml/src/ggml-cuda/fattn.cu
* use ggml_cuda_info().devices warp_size instead of hardcoded check
Daniel Bevenius [Fri, 27 Feb 2026 13:42:24 +0000 (14:42 +0100)]
ci : use ubuntu-latest for gguf-publish workflow (#19951)
This commit changes the runner for the gguf-publish workflow from
ubuntu-slim back to ubuntu-latest, which was updated in Commit 142cbe2ac68978e5dec3a2e19c1b64ef1c5740b1 ("ci : use new 1vCPU runner for
lightweight jobs (#19107)").
The motivation for this is that the action used in the workflow depends
on the docker daemon, which does not seem not available in the
ubuntu-slim runner. This is currently causing an error in the workflow
and preventing the gguf-publish workflow from running successfully.
Today was the the first time since the original change (I think) that
publish task has been run which may be why the issue was not noticed
before.
Pascal [Fri, 27 Feb 2026 06:05:23 +0000 (07:05 +0100)]
server : support multiple model aliases via comma-separated --alias (#19926)
* server : support multiple model aliases via comma-separated --alias
* server : update --alias description and regenerate docs
* server : multiple model aliases and tags
- address review feedback from ngxson
- --alias accepts comma-separated values (std::set, no duplicates)
- --tags for informational metadata (not used for routing)
- aliases resolve transparently in router via get_meta/has_model
- /v1/models exposes aliases and tags fields
* regenerate docs
* nits
* server : use first alias as model_name for backward compat
address review feedback from ngxson
* server : add single-model test for aliases and tags
Jan Patrick Lehr [Fri, 27 Feb 2026 04:37:54 +0000 (05:37 +0100)]
tests : enable test-chat out of tree build (#19558)
The binary relies on model files that it tries to find. However, when
configuring the build directory to be parallel to the source tree those
heuristics fail.
This sets the working directory for the test executable to be the
source-tree which resolves this issue.
Vishal Singh [Fri, 27 Feb 2026 00:43:41 +0000 (06:13 +0530)]
ggml-zendnn: update code for latest ZenDNN API (#19923)
- adapt ggml-zendnn.cpp to the new lowoha::matmul interface
- update the ZenDNN git tag in CMake to the latest release (ZenDNN‑2026‑WW08)
- add static lib support in CMake
Kevin Pouget [Thu, 26 Feb 2026 12:00:57 +0000 (13:00 +0100)]
ggml-virtgpu: improve the reliability of the code (#19846)
* ggml-virtgpu-backend: validate the consistency of the received objects
This patch adds consistency checks in the
ggml-virtgpu-backend (running on the host side) to ensure that the
data received from the guest is consistent (valid pointers, valid
sizes and offsets).
* ggml-virtgpu-backend: add fallback/skips for optional ggml backend methods
these three methods are optional in the GGML interface. `get_max_size`
was already properly defaulted, but `backend sychronize` and `butf
get_max_size` would have segfaulted the backend if not implemented.
* ggml-virtgpu-backend: fix log format missing argument
* ggml-virtgpu-backend: improve the abort message
* ggml-virtgpu-backend: more safety checks
* ggml-virtgpu-backend: new error code
* ggml-virtgpu-backend: initialize all the error codes
* ggml-virtgpu: add a missing comment generated by the code generator
* ggml-virtgpu: add the '[virtgpu]' prefix to the device/buffer names
* ggml-virtgpu: apir_device_buffer_from_ptr: improve the error message
* ggml-virtgpu: shared: make it match the latest api_remoting.h of Virglrenderer APIR
(still unmerged)
* ggml-virtgpu: update the code generator to have dispatch_command_name in a host/guest shared file
* ggml-virtgpu: REMOTE_CALL: fail if the backend returns an error
* docs/backend/VirtGPU.md: indicate that the RAM+VRAM size is limed to 64 GB with libkrun
* ggml-virtgpu: turn off clang-format header ordering for some of the files
Compilation breaks when ordered alphabetically.
* ggml-virtgpu: clang-format
* ggml-virtgpu/backend/shared/api_remoting: better comments for the APIR return codes
Maximilian Werk [Thu, 26 Feb 2026 11:14:09 +0000 (12:14 +0100)]
model : add Jina Embeddings v5 Nano (partial EuroBERT) support (#19826)
* WIP: Add EuroBERT support with autoformatting changes
This commit includes:
- EuroBERT model implementation for GGUF conversion
- C++ backend support for EuroBERT architecture
- Unintended autoformatting changes to Python files
Saving before reverting formatting-only changes.
* feat: add back eos assert when not last token pooling
* feat: removed duplicated code and cleanup
* feat: removed not working architectures and unnecessary check
* fix: typo
* fix: dynamic pooling config
* feat: added an example model for eurobert
* feat: proper llama-vocab implementation for jina-v5
Mario Limonciello [Wed, 25 Feb 2026 11:30:19 +0000 (05:30 -0600)]
ci : update Windows ROCm build to 26.Q1 [no ci] (#19810)
* Update build command to build llama-* tools not just ggml-hip
* Update rocWMMA headers to 7.2
* Add GFX1150 target
* Correct library paths for AMD libraries in 26.Q1
Daniel Bevenius [Mon, 23 Feb 2026 13:15:16 +0000 (14:15 +0100)]
model-conversion : merge inspect-org-model.py with tensor-info.py (#19823)
This commit replaces/merges the inspect-org-model.py script with the
contents tensor-info.py script. The merged script has also been updated
to also print tensor sizes which was the only thing that was not done
before (by tensor-info.py that is).
The motivation for this is that tensor-info.py does not load the tensor
weights which can be time consuming for larger models. And also now that
both are doing almost the same thing it makes sense to just have one and
not two scripts to maintain.
This commit updates the session handing in the completion tool to handle
the that logits are no longer stored in the session file. Instead, we
need to replay the last token to get the logits for sampling.
* common : add common_prompt_batch_decode function
This commit adds a new function which is responsible for decoding prompt
and optionally handle the saving for session data.
* update save-state.cpp to use llama_state_load_file
This commit updates the save-load-state example to utilize the new
llama_state_load_file function for loading the model state from a file.
And it also replays the last token after loading since this state is now
stored before the last token is processed.
* examples : set n_seq_max = 2 for ctx3
This commit updates the save-load-state example to set the n_seq_max
parameter to 2 when initializing the ctx3 context.
The motivation for this change is that using 1 as n_parallel/n_seq_max
the context only supports one sequence, but the test laster tries to
use a second sequence which results in the following error:
```console
main : loaded state with 4 tokens
main : seq 0 copied, 225760 bytes
main : kv cache cleared
find_slot: seq_id=1 >= n_seq_max=1 Try using a bigger --parallel value
state_read_meta: failed to find available cells in kv cache
```
This seems to only happen for recurrent/hybrid models.
Gaurav Garg [Sat, 21 Feb 2026 09:39:36 +0000 (15:09 +0530)]
Improve CUDA graph capture (#19754)
* Improve CUDA graph capture
Currently, CUDA graphs are eagerly enabled on the first call to ggml_backend_cuda_graph_compute. If the graph properties keep changing (4+ consecutive updates), the graph is permanently disabled. This is suboptimal because:
- The first call always incurs CUDA graph capture overhead even if the graph is unstable
- Once permanently disabled, CUDA graphs never re-enable even after the graph stabilizes (e.g., switching from prompt processing to decode)
The new approach delays CUDA graph activation until warmup completes: the same cgraph must be called at least twice with matching properties before CUDA graph capture begins. This avoids wasted capture overhead on volatile graphs and allows graphs to become eligible once they stabilize.
This also fixes issues such as https://github.com/ggml-org/llama.cpp/discussions/19708
* Update ggml/src/ggml-cuda/ggml-cuda.cu
Co-authored-by: Johannes Gäßler <redacted>
* Remove EM dashes
* Update ggml/src/ggml-cuda/ggml-cuda.cu
Co-authored-by: Aman Gupta <redacted>
---------
Co-authored-by: Johannes Gäßler <redacted> Co-authored-by: Aman Gupta <redacted>
Jesse Posner [Thu, 19 Feb 2026 21:40:52 +0000 (13:40 -0800)]
common : fix Step-3.5-Flash format detection and thinking support (#19635)
* common : fix Step-3.5-Flash format detection and thinking support
Step-3.5-Flash uses the same XML-style tool call format as Qwen3-Coder
(<tool_call><function=...><parameter=...>) but its Jinja template lacks
the bare <function> and plural <parameters> markers that the detection
logic previously required. This caused it to fall through to Hermes 2
Pro, which doesn't call func_args_not_string(), so arguments stayed as
JSON strings and templates using arguments|items crashed.
Additionally, the Qwen3-Coder-XML format handler had no thinking support.
Models like Step-3.5-Flash that unconditionally emit <think> in their
generation prompt need the same thinking_forced_open handling that
Nemotron v3 and Hermes 2 Pro already have, otherwise reasoning_content
is never separated from content in API responses.
Changes:
- Relax Qwen3-Coder XML detection to only require the 3 shared markers
- Tighten Nemotron v3 branch to also require bare <function> and plural
<parameters>, preventing Step-3.5-Flash from being misrouted via <think>
- Add thinking_forced_open support to Qwen3-Coder-XML init function
- Add <think>/</think> to preserved tokens
- Fix build_grammar_xml_tool_call to handle thinking_forced_open in the
grammar root rule, allowing </think> before tool calls
- Add Step-3.5-Flash chat template and format detection test
Builds on: https://github.com/ggml-org/llama.cpp/pull/19283
Step-3.5-Flash uses the same XML tool call format as Qwen3-Coder and
Nemotron 3 Nano (<tool_call>/<function=...>/<parameter=...>) but with
unconditional <think> output. Route it to the Nemotron v3 PEG parser
for streaming and schema-aware parameter parsing.
Detection: templates with <think> + XML tool tags use Nemotron v3 PEG
parser; templates without <think> (Qwen3-Coder) use GBNF grammar.
* chat : remove dead thinking code from qwen3_coder_xml
Remove thinking handling code that became unreachable after routing
Step-3.5-Flash to the Nemotron v3 PEG parser. Qwen3-Coder has no
<think> in its template, so the thinking_forced_open logic, preserved
tokens, and grammar prefix were dead paths.
Note: JAIS-2 requires F32 precision accumulators for numerical stability
and uses standard attention (not flash attention) on CUDA backends.
* fix: run convert_hf_to_gguf_update.py for jais-2 tokenizer hash
* fix: use NEOX RoPE type for JAIS2
* fix: remove Q/K permutation (NEOX RoPE doesn't need it)
* fix: enable flash attention for JAIS2 (fixed by #19115)
* fix: add dedicated JAIS2 pre-tokenizer type and control vector support
- Add LLAMA_VOCAB_PRE_TYPE_JAIS2 with cascading whitespace regex
- Include original regex from tokenizer.json as comment
- Add build_cvec call for control vector support
Tarek Dakhran [Thu, 19 Feb 2026 11:18:57 +0000 (12:18 +0100)]
mtmd : chat : Fix extra \n between text and media marker (#19595)
* mtmd : chat : Fix extra \n between text and media marker
Thanks to @tugot17 for detecting and reporting the issue.
For vision models (e.g. LFM2.5-VL-1.6B and Qwen/Qwen3-VL-4B-Instruct) `llama-mtmd-cli` produces identical output to HF implementation.
However `llama-server` doesn't. I traced it down to extra newline
inserted after `<__media__>`.
This happens in `to_json_oaicompat`, that treats media markers as text
and joins all parts with `\n` separator.
PR introduces new type `media_marker` and uses it for media markers.
Extra logic is added to prevent insertion of newlines before and after
media markers.
With this change number of input tokens is identical to HF
implementation and as a result the output is also identical.
I explored other ways to address the issue
* remove completely `\n` between text parts in `to_json_oaicompat`
* merge text messages in server-common.cpp before sending them to `to_json_oaicompat`
Please propose alternative ways of fixing this issue.
* Refactor to use explicite per type ifs
* Update common/chat.cpp
Co-authored-by: Piotr Wilkin (ilintar) <redacted>
* Update common_chat_templates_apply_legacy
overview / pkg / ggml / sources / llama.cpp / log