ollama/llm
Daniel Hiltgen cecac8bf1b
mlxrunner: add Inkling 2 Mini (#19799)
* mlxrunner: add Inkling 2 Mini

* mlx: fuse Inkling 2 Mini residual mixing on Metal

Fuse RMS normalization and query scoring in one kernel. Apply the
softmax weights in a second kernel to avoid storing per-query products.

* create: expose Inkling 2 Mini context length in model metadata

Copy model_max_length to max_position_embeddings during import so
ollama show reports the checkpoint's context length. Reject configs
that specify conflicting limits.

* mlxrunner: reduce Metal dispatches in Inkling 2 Mini decode

Reuse the prefill mask shader to compute relative position bias in one
dispatch for single-query attention. On M3 Ultra with Metal 3, compute
the loop controller's gain and forget projections in one matrix multiply.
2026-11-20 18:38:44 -08:00
..
exit_status.go runner: Remove CGO engines, use llama-server exclusively for GGML models (#16031) 2026-05-29 13:35:47 -07:00
exit_status_other.go runner: Remove CGO engines, use llama-server exclusively for GGML models (#16031) 2026-05-29 13:35:47 -07:00
exit_status_windows.go runner: Remove CGO engines, use llama-server exclusively for GGML models (#16031) 2026-05-29 13:35:47 -07:00
gbnf.go llm: apply a format after the thinking in one llama-server request 2026-09-22 11:24:11 -07:00
gbnf_test.go llm: apply a format after the thinking in one llama-server request 2026-09-22 11:24:11 -07:00
llama_binary.go create: add server-side MLX imports and drop GGUF conversion (#14969) 2026-09-14 20:32:36 -07:00
llama_binary_test.go create: add server-side MLX imports and drop GGUF conversion (#14969) 2026-09-14 20:32:36 -07:00
llama_server.go Merge pull request #19748 from ollama/claude/sched-shift-no-reload-18950 2026-11-18 14:53:02 -08:00
llama_server_speech.go llm: LFM3 GGUF fixes and llama-server speech (#19344) 2026-10-27 18:56:22 -07:00
llama_server_speech_test.go llm: LFM3 GGUF fixes and llama-server speech (#19344) 2026-10-27 18:56:22 -07:00
llama_server_systemone.go llama: update to b11521 and use native decision support 2026-10-11 23:25:23 -07:00
llama_server_test.go Merge pull request #19748 from ollama/claude/sched-shift-no-reload-18950 2026-11-18 14:53:02 -08:00
llm_darwin.go
llm_linux.go
llm_windows.go win: lint fix (#10571) 2025-05-05 11:08:12 -07:00
media.go runner: Remove CGO engines, use llama-server exclusively for GGML models (#16031) 2026-05-29 13:35:47 -07:00
metal_retry.go runner: Remove CGO engines, use llama-server exclusively for GGML models (#16031) 2026-05-29 13:35:47 -07:00
model_split.go create: add server-side MLX imports and drop GGUF conversion (#14969) 2026-09-14 20:32:36 -07:00
model_split_test.go create: add server-side MLX imports and drop GGUF conversion (#14969) 2026-09-14 20:32:36 -07:00
rocm_default.go runner: Remove CGO engines, use llama-server exclusively for GGML models (#16031) 2026-05-29 13:35:47 -07:00
rocm_windows.go llama-server followups (#16353) 2026-06-01 10:44:21 -07:00
runner_job_other.go win: prevent orphaned runners (#18500) 2026-10-15 12:41:45 -07:00
runner_job_windows.go win: prevent orphaned runners (#18500) 2026-10-15 12:41:45 -07:00
runner_job_windows_test.go win: prevent orphaned runners (#18500) 2026-10-15 12:41:45 -07:00
score.go mlx: add LFM3 and d1 models (#19333) 2026-10-27 08:47:18 -07:00
server.go mlxrunner: add Inkling 2 Mini (#19799) 2026-11-20 18:38:44 -08:00
server_wait_test.go runner: Remove CGO engines, use llama-server exclusively for GGML models (#16031) 2026-05-29 13:35:47 -07:00
speech.go llm: LFM3 GGUF fixes and llama-server speech (#19344) 2026-10-27 18:56:22 -07:00
status.go llama.cpp: version update (#18577) 2026-09-22 13:58:49 -07:00
status_test.go llama.cpp: version update (#18577) 2026-09-22 13:58:49 -07:00
vulkan_windows.go llm: use host Vulkan loader on Windows (#16869) 2026-06-24 10:35:48 -07:00