Commit Graph

Select branches

Hide Pull Requests

nik

Mono Color

• 66199c9f03 ggml : use a simple std::thread in AMX without OpenMP (#20074) Adrien Gallouët 2026-03-04 11:57:09 +01:00
• c99909dd0b impl : use 6 digits for tensor dims (#20094) ddh0 2026-03-04 02:53:38 -06:00
• cb8f4fa3f8 Fix locale-dependent float printing in GGUF metadata (#17331) SamareshSingh 2026-03-04 02:30:40 -06:00
• 54910bd4f3 completion : Fix a typo in warning message (#20082) standby24x7 2026-03-04 14:44:49 +09:00
• ecd99d6a9a docs: Fix intel documentation link (#20040) Mickael Desgranges 2026-03-03 14:50:00 +01:00
• 137435ff15 kleidiai : add sme fp16 compute path for q4_0 gemm on aarch64 (#20043) Charles Xu 2026-03-03 10:40:26 +01:00
• 24350fdf9b opencl: add optimized q4_1 mm kernel for adreno (#19840) shaofeiqi 2026-03-02 19:49:41 -08:00
• 49a7564ac1 ggml webgpu: fix workgroup dispatch limit for large batch sizes (#19965) Abhijit Ramesh 2026-03-02 19:35:11 -08:00
• 4d828bd1ab ggml webgpu: Clean up per-thread parameter buffer pool and job submission logic (#19772) Nikhil Jain 2026-03-02 10:23:34 -08:00
• 36a7a6589c ggml-webgpu: Support non-contiguous src0 and overlapping src0/src1 in binary ops (#19850) Masashi Yoshimura 2026-03-03 00:59:53 +09:00
• feefb92836 vulkan: tune MMVQ for Intel Windows (#19988) Ruben Ortlam 2026-03-02 15:58:25 +01:00
• ec88c3ceea scripts : improve get-wikitext-2.sh (#19952) Adrien Gallouët 2026-03-02 15:40:49 +01:00
• 2afcdb9777 ggml-cpu: optimise s390x multiply extend instructions (#20032) Aaron Teo 2026-03-02 16:23:56 +08:00
• 319146247e vulkan: improve partial offloading performance on AMD (#19976) Ruben Ortlam 2026-03-01 17:32:14 +01:00
• 66d65ec29b cuda: cap grid.y at 65535 in non-contiguous dequantize/convert kernels (#19999) oobabooga 2026-03-01 02:40:22 -03:00
• 05728db18e vendors : update miniaudio library to 0.11.24 (#19914) Dmitry Atamanov 2026-02-28 20:10:01 +05:00
• 4720819d45 vendor : update cpp-httplib to 0.35.0 (#19969) Adrien Gallouët 2026-02-28 13:53:56 +01:00
• d979f2b176 tests : model metadata loading from huggingface (#19796) Bartowski 2026-02-28 04:44:38 -05:00
• ecbcb7ea9d CUDA: add CDNA3 MFMA support for flash attention MMA kernel (#19806) Jayant Lohia 2026-02-28 00:07:26 +05:30
• 3e6ab244ad server: Add pragma once to server-context.h (#19944) Roj234 2026-02-28 01:28:36 +08:00
• 5596a35791 server: Mirroring /v1/responses to /responses to match /v1/chat/completions pattern (#19873) Sami Kama 2026-02-27 08:44:42 -08:00
• 8d3b962f47 ci : use ubuntu-latest for gguf-publish workflow (#19951) Daniel Bevenius 2026-02-27 14:42:24 +01:00
• d903f30e25 ggml-cpu: add repack for mxfp4 (#19738) Aman Gupta 2026-02-27 18:15:09 +08:00
• 8387ffb28d gguf-py : dump version to 0.18.0 (#19950) Daniel Bevenius 2026-02-27 11:02:53 +01:00
• 2e7e638523 server : support multiple model aliases via comma-separated --alias (#19926) Pascal 2026-02-27 07:05:23 +01:00
• a8b192b6ec tests : enable test-chat out of tree build (#19558) Jan Patrick Lehr 2026-02-27 05:37:54 +01:00
• c17dce4f5c replace the magic nunber 768 by max work group size to support iGPU (#19920) Neo Zhang 2026-02-27 09:26:07 +08:00
• 88cf781f51 ggml-zendnn: update code for latest ZenDNN API (#19923) Vishal Singh 2026-02-27 06:13:41 +05:30
• 4e76d24f28 ggml : fix AMX and add batched support (#19925) Adrien Gallouët 2026-02-26 21:39:11 +01:00
• 723c71064d vulkan: fix fp16 Flash Attention on Windows AMD RDNA2 and below (#19921) Ruben Ortlam 2026-02-26 19:11:04 +01:00
• 37964f44f9 mtmd : fix padding of n_tokens (#19930) Georgi Gerganov 2026-02-26 18:39:49 +02:00
• 01cd448b8c server : fix ctx checkpoint restore logic (#19924) Georgi Gerganov 2026-02-26 18:20:16 +02:00
• 99bd67c9b2 kv-cache : fix can_shift() check to take into account M-RoPE (#19928) Georgi Gerganov 2026-02-26 18:08:54 +02:00
• b68d75165a llama: Add option to merge gate and exp weights (#19139) Aman Gupta 2026-02-26 21:01:08 +08:00
• ffaafde16f ggml-virtgpu: improve the reliability of the code (#19846) Kevin Pouget 2026-02-26 13:00:57 +01:00
• efba35a860 server: fix load-on-startup not respected in ini file (#19897) drrros 2026-02-26 14:32:31 +03:00
• 9b62913b40 jinja : correct default size for string slices (#19913) Eric Zhang 2026-02-26 19:28:09 +08:00
• 66287bdaac model : add Jina Embeddings v5 Nano (partial EuroBERT) support (#19826) Maximilian Werk 2026-02-26 12:14:09 +01:00
• 1ca3d1de15 gguf : avoid too many file size calls (#19919) Georgi Gerganov 2026-02-26 12:46:32 +02:00
• bd72300591 server : fix typo in server README.md (#19900) yggdrasil75 2026-02-26 05:26:16 -05:00
• 2943210c1e support permuted, remove check s0/s10 (#19889) Neo Zhang 2026-02-26 10:27:20 +08:00
• 3769fe6eb7 vulkan: check for memory overlap before doing fusion (#19768) Jeff Bolz 2026-02-25 11:25:38 -06:00
• 832aa94762 common : add more aliases for sampler CLI params (#19797) ddh0 2026-02-25 09:34:25 -06:00
• 3af34b9ff5 ci : update the ROCm/HIP toolchain versions [no ci] (#19891) Slobodan Josic 2026-02-25 15:54:49 +01:00
• f20469d919 server : enable multi-modal prompt caching (#19877) Georgi Gerganov 2026-02-25 15:15:42 +02:00
• d7d826b3c1 server : support multi-modal context checkpoints (#19849) Georgi Gerganov 2026-02-25 15:14:27 +02:00
• c747294b2d scripts: update corpus of compare-logprobs (#19326) Xuan-Son Nguyen 2026-02-25 12:57:34 +01:00
• 8fdf269dad ci : update Windows ROCm build to 26.Q1 [no ci] (#19810) Mario Limonciello 2026-02-25 05:30:19 -06:00
• a96a1120b4 gguf : fix ftell/fseek for Windows (#19870) Aldehir Rojas 2026-02-24 22:58:11 -06:00
• 244641955f models : fix graph splits (#19866) Georgi Gerganov 2026-02-25 00:01:13 +02:00
• 47eb12b953 server: fix query params lost when proxying requests in multi-model router mode (#19854) Pascal 2026-02-24 21:46:06 +01:00
• 418dea39ce ggml/gguf : prevent integer overflows (#19856) Georgi Gerganov 2026-02-24 20:17:11 +02:00
• da426cb250 model : update label for LFM2-24B-A2B (#19848) Tarek Dakhran 2026-02-24 14:27:42 +01:00
• c830f99cfa server : support max_completion_tokens request property (#19831) Radoslav Gerganov 2026-02-24 10:30:00 +02:00
• aa6f918c1c Vulkan Scalar Flash Attention Refactor (#19625) Ruben Ortlam 2026-02-24 08:35:48 +01:00
• 8c2c0108dd vulkan: fix coopmat1 without bf16 support (#19793) Jeff Bolz 2026-02-24 00:48:32 -06:00
• 3ea5360c00 vulkan: fix data race in mul_mat_id shader (#19790) Jeff Bolz 2026-02-24 00:43:12 -06:00
• 39fb81f875 hexagon refactor all Ops to use local context struct (#19819) Max Krasnyansky 2026-02-23 16:32:14 -08:00
• 5eb0ea32f0 feat: Add code blocks full height setting to parameter sync service (#19835) Aleksander Grygier 2026-02-23 22:30:13 +01:00
• b68a83e641 vendor : update cpp-httplib to 0.34.0 (#19830) Adrien Gallouët 2026-02-23 21:05:48 +01:00
• d8aeb65cee tests : fix typos in comments in test-backend-sampler [no ci] (#19824) Daniel Bevenius 2026-02-23 17:12:02 +01:00
• 9051663d5d webui: Add setting to have full height Code Blocks in Chat Messages (#19829) Aleksander Grygier 2026-02-23 14:16:50 +01:00
• 72b44c0d21 model-conversion : merge inspect-org-model.py with tensor-info.py (#19823) Daniel Bevenius 2026-02-23 14:15:16 +01:00
• bc160d3582 ggml-cpu: arm64: q5_K repack gemm and gemv (and generic) implementations (dotprod) (#19356) Alberto Cabrera Pérez 2026-02-23 12:42:52 +00:00
• 2b6dfe824d llama : remove write/read of output ids/logits/embeddings (#18862) Daniel Bevenius 2026-02-23 07:04:30 +01:00
• e8e261699a cli : provide model with text filename (#19783) Sigbjørn Skjæret 2026-02-22 22:33:49 +01:00
• 5452d736f8 jinja: correct stats for tojson and string filters (#19785) Xuan-Son Nguyen 2026-02-22 21:08:23 +01:00
• ed4837891d common : fix improper trimming in XML parser on complete message (#19805) Aldehir Rojas 2026-02-22 10:34:54 -06:00
• cacc371f99 Fix wrong cli-argument in documentation (#19804) Kilian Krampf 2026-02-22 16:26:33 +01:00
• ae2368e74e model : add Kanana-2 model support (#19803) HelloKS 2026-02-23 00:15:02 +09:00
• 9f0684f003 ci : fix rocm archive name [no ci] (#19808) Sigbjørn Skjæret 2026-02-22 16:14:37 +01:00
• 34ec1c3f18 server : merge contiguous Responses input items into a single assistant message (#19773) Aldehir Rojas 2026-02-22 07:11:31 -06:00
• e877ad8bd9 ci : fix rocm release path [no ci] (#19784) Sigbjørn Skjæret 2026-02-22 08:07:46 +01:00
• 35715657cb Update ROCm docker container to 7.2 release (#19418) Mario Limonciello 2026-02-21 14:53:39 -06:00
• f75c4e8bf5 Add a build target to generate ROCm artifacts using ROCm 7.2 (#19433) Mario Limonciello 2026-02-21 12:56:26 -06:00
• 99156f3a5f vendor : update cpp-httplib to 0.33.1 (#19778) Adrien Gallouët 2026-02-21 19:12:31 +01:00
• a0c91e8f9f Improve CUDA graph capture (#19754) Gaurav Garg 2026-02-21 15:09:36 +05:30
• 07968d53e4 fix: UI single model selection in router mode (#19767) crsawyer 2026-02-21 02:28:39 -06:00
• ba3b9c8844 hexagon : fix build release (#19444) (#19587) Mengsheng Wu 2026-02-20 16:40:00 -08:00
• 94b0200a01 common : merge qwen3-coder and nemotron nano 3 parsers (#19765) Aldehir Rojas 2026-02-20 16:22:22 -06:00
• b908baf182 ggml-cpu: add RVV vec dot kernels for quantization types (#18784) Taimur Ahmad 2026-02-20 16:30:07 +05:00
• 492bc31978 quantize : add --dry-run option (#19526) ddh0 2026-02-20 02:20:16 -06:00
• 77d6ae4ac8 test: mul_mat tests with huge batch size (#19519) Jeff Bolz 2026-02-19 18:08:25 -08:00
• 10b26ee23a WebUI hide models in router mode (#19374) crsawyer 2026-02-19 15:53:42 -06:00
• 3dadc88b58 common : fix Step-3.5-Flash format detection and thinking support (#19635) Jesse Posner 2026-02-19 13:40:52 -08:00
• 39e4b1dc9b common : fix gpt-oss Jinja error when assistant message has both content and thinking with tool calls (#19704) abhijitb11 2026-02-19 12:59:20 -08:00
• 11c325c6e0 ggml-webgpu: Add unary op (SQR, SQRT, SIN, COS) support. (#19700) Masashi Yoshimura 2026-02-20 01:18:30 +09:00
• 237958db33 model: Add PaddleOCR-VL model support (#18825) megemini 2026-02-20 00:05:25 +08:00
• abb9f3c42b vulkan: fix MMQ shader push constants and multi-dispatch (#19732) Ruben Ortlam 2026-02-19 14:59:16 +01:00
• da348c9dfb models : fix qwen3.5 beta/gate shapes (#19730) Georgi Gerganov 2026-02-19 15:19:53 +02:00
• e6267a9359 mtmd: build_attn modified, flash_attn on/off via ctx_params (#19729) Saba Fallah 2026-02-19 13:50:29 +01:00
• 2bf318fd2f model : add JAIS-2 architecture support (#19488) 3 a l i 2026-02-19 16:30:17 +04:00
• c78e682245 CUDA: fix kernel selection logic for tile FA (#19686) Johannes Gäßler 2026-02-19 12:42:58 +01:00
• c5897995a7 mtmd : chat : Fix extra \n between text and media marker (#19595) Tarek Dakhran 2026-02-19 12:18:57 +01:00
• 03fd9d3bb4 webui: Fix Attachments not being included in completion request (#19731) Aleksander Grygier 2026-02-19 10:27:38 +01:00
• 8004f3a8d1 model : add tokenizer from LFM2.5-Audio-1.5B (#19687) Tarek Dakhran 2026-02-19 09:54:48 +01:00
• eacb4b67a2 llama : use output_resolve_row() in get_logits_ith/get_embeddings_ith (#19663) Daniel Bevenius 2026-02-19 09:48:08 +01:00
• c0d0430340 model : full modern bert support (#18330) Ryan Mangeno 2026-02-19 02:52:21 -05:00
• 3bb2fcc856 llamafile: powerpc: add FP16 MMA path for Q4/Q8 matmul (#19709) shalinib-ibm 2026-02-19 11:58:53 +05:30
• 27326bfce1 models : dedup qwen35 graphs (#19660) Georgi Gerganov 2026-02-19 08:17:49 +02:00