upstream-archive/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-07-29 06:00:58 -05:00
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • 673cfef9aa CANN: Fix build error with GCC 13 (#11990) hipudding 2025-02-28 15:23:47 +08:00
  • fbeda9002d vulkan: matmul dequantization improvements (#12015) b4786 Eve 2025-02-28 07:20:08 +00:00
  • 581650b7ca vulkan: improve im2col (#11826) b4785 Daniele 2025-02-28 06:52:51 +00:00
  • 828effd9d7 kv-cache : basic abstraction Georgi Gerganov 2025-02-27 15:54:44 +02:00
  • 82675a0180 Merge branch 'master' into gg/llama-kv-cache Georgi Gerganov 2025-02-27 15:10:18 +02:00
  • 952feedfca context : disable encoder embd tensor for now Georgi Gerganov 2025-02-27 15:07:10 +02:00
  • b95c8af37c cmake: Fix ggml backend dependencies and installation (#11818) b4784 Vladimir Vuksanovic 2025-02-27 08:42:48 +01:00
  • c9ecf620d6 Merge branch 'master' into compilade/mamba2 Francis Couture-Harpin 2025-02-26 15:32:20 -05:00
  • 7037e94852 vulkan: subgroup size test Daniele 2025-02-09 23:30:13 +00:00
  • a800ae46da llava : add struct for FFI bindgen (#12079) b4783 Ting Lou 2025-02-26 22:26:52 +08:00
  • 69050a11be Refactor gguf scripts to improve metadata handling (#11909) gguf-v0.16.0 Sigbjørn Skjæret 2025-02-26 14:04:48 +01:00
  • 3567ee3a94 gguf-py: enable reading non-native endian files (#12081) Aleksei Nikiforov 2025-02-26 12:39:27 +01:00
  • 53e4db1012 readme : update infra list (#9096) Kante Yin 2025-02-26 15:49:36 +08:00
  • d7cfe1ffe0 docs: add docs/function-calling.md to lighten server/README.md's plight (#12069) Olivier Chafik 2025-02-25 18:52:56 +00:00
  • a82c9e7c23 vulkan: fix assertion when qy_needs_dequant (#12068) b4778 Jeff Bolz 2025-02-25 09:30:21 -06:00
  • 4efe989886 context : pass embeddings tensor from encoder to decoder Georgi Gerganov 2025-02-25 16:11:17 +02:00
  • 401af80b54 server: handle echo=false on /v1/completions (#12060) b4777 rhjdvsgsgks 2025-02-25 11:52:52 +00:00
  • c132239bfb add OP sigmoid (#12056) b4776 Judd 2025-02-25 19:32:20 +08:00
  • 393fca629e ggml-cpu: Fix build with sve (#12059) b4775 Molly Sophia 2025-02-25 19:28:22 +08:00
  • 61d4f39dfe vulkan: implement more backpropagation operators (#11914) b4774 Rémy O 2025-02-25 12:04:45 +01:00
  • 0b52745649 server: support add_generation_prompt query param (#12062) b4773 Olivier Chafik 2025-02-25 10:40:22 +00:00
  • e2b3294f2c context : fix enc-dec state save/load Georgi Gerganov 2025-02-25 12:14:34 +02:00
  • e5bc5f8e02 context : enc-dec is now working Georgi Gerganov 2025-02-25 12:10:34 +02:00
  • 4d1051a40f Add Doc for Converting Granite Vision -> GGUF (#12006) Alex Brooks 2025-02-25 02:46:05 -07:00
  • 3e9a2860e9 llama : expose llama_model_n_head_kv in the API (#11997) b4771 Vitali Lovich 2025-02-25 01:29:33 -08:00
  • 58d07a8043 metal : copy kernels for quant to F32/F16 conversions (#12017) b4770 Gian-Carlo Pascutto 2025-02-25 10:27:58 +01:00
  • 34a846b584 opencl: fix for small models (#11950) b4769 lhez 2025-02-24 13:47:07 -08:00
  • be58e30017 enc-dec : compose wip Georgi Gerganov 2025-02-24 15:16:45 +02:00
  • 7a2c913e66 llava : Add Granite Vision Support (#11794) b4768 Alex Brooks 2025-02-24 09:09:51 -07:00
  • a1b1dea33b Merge branch 'master' into xsn/private_batch_api Xuan Son Nguyen 2025-02-24 17:01:30 +01:00
  • 4bf7ca3943 llama_decode_ext Xuan Son Nguyen 2025-02-24 17:01:20 +01:00
  • 08d5986290 [SYCL] Optimize mul_mat for Q4_0 on Intel GPU (#12035) b4767 Neo Zhang Jianyu 2025-02-24 22:33:23 +08:00
  • 9cd78f11a1 context : explicit llama_context_i abstract interface Georgi Gerganov 2025-02-24 13:38:11 +02:00
  • 651adf4b66 gguf_convert_endian.py: implement byteswapping for q4_k and q6_k (#11349) Aleksei Nikiforov 2025-02-24 12:27:01 +01:00
  • 8303e8b0fb SYCL: Fix GGML_SYCL_DEBUG macro (#11995) b4765 Akarshan Biswas 2025-02-24 15:48:25 +05:30
  • 4a1054b552 context : reuse built_attn_mha Georgi Gerganov 2025-02-24 11:18:40 +02:00
  • a5a85a3bc0 context : fix recurrent reserve Georgi Gerganov 2025-02-24 08:59:12 +02:00
  • 0699a44c83 context : remove redundant virtual, protected -> private Georgi Gerganov 2025-02-23 20:02:11 +02:00
  • 6378112cb5 graph : remove the build_kv_... API from llama_graph_i Georgi Gerganov 2025-02-23 19:39:22 +02:00
  • 7ad0779f5d run: allow to customize prompt by env var LLAMA_PROMPT_PREFIX (#12041) b4764 Florent BENOIT 2025-02-23 18:15:51 +01:00
  • f777a73e18 Some llama-run cleanups (#11973) b4763 Eric Curtin 2025-02-23 13:14:32 +00:00
  • 372fa3a894 cont : enc should work now, next is dec Georgi Gerganov 2025-02-23 11:38:59 +02:00
  • af7747c95a ggml-cpu: Support s390x SIMD Instruction Set (#12019) b4762 Aaron Teo 2025-02-23 05:39:24 +08:00
  • a28e0d5eb1 CUDA: app option to compile without FlashAttention (#12025) b4761 Johannes Gäßler 2025-02-22 20:44:34 +01:00
  • 36c258ee92 llava: build clip image from pixels (#11999) b4760 Ting Lou 2025-02-22 22:28:28 +08:00
  • f3e64859ed ci : fix arm upload artifacts (#12024) b4759 Georgi Gerganov 2025-02-22 15:03:00 +02:00
  • f343850bd4 cont : fix archive name to use matrix gg-ci-fix-arm-b4760-f343850 Georgi Gerganov 2025-02-22 14:39:31 +02:00
  • 3f683b4088 ci : fix arm upload artifacts Georgi Gerganov 2025-02-22 13:53:39 +02:00
  • 5fa07c2f93 CUDA: optimize FA for GQA + large batches (#12014) Johannes Gäßler 2025-02-22 12:20:17 +01:00
  • 335eb04a91 ci : Build on Github-hosted arm64 runners (#12009) Rohanjames1997 2025-02-22 04:48:57 -06:00
  • cf756d6e0a server : disable Nagle's algorithm (#12020) b4756 Georgi Gerganov 2025-02-22 12:46:31 +02:00
  • d70908421f cuda: Add Q5_1, Q5_0, Q4_1 and Q4_0 to F32 conversion support. (#12000) b4755 Gian-Carlo Pascutto 2025-02-22 09:43:24 +01:00
  • de8b5a3624 llama.swiftui : add "Done" dismiss button to help view (#11998) b4754 Daniel Bevenius 2025-02-22 06:33:29 +01:00
  • 6f7fe74946 ggml-quants : improve imatrix behavior for TQ1_0, TQ2_0, Q4_0, Q5_0 Francis Couture-Harpin 2025-02-21 18:47:09 -05:00
  • d0060fc498 ggml-quants : better and faster make_qkxs_quants Francis Couture-Harpin 2025-02-21 15:05:03 -05:00
  • dd6b8408c9 ggml-quants : improve IQ4_NL, IQ4_XS, and Q3_K Francis Couture-Harpin 2025-02-21 13:14:32 -05:00
  • f5e80208c5 wip enc-dec Georgi Gerganov 2025-02-21 19:17:47 +02:00
  • c4c0a4d13c Merge branch 'master' into gg/llama-kv-cache Georgi Gerganov 2025-02-21 19:14:07 +02:00
  • 51f311e057 llama : skip loading unused tensors (#12004) b4753 Georgi Gerganov 2025-02-21 18:33:18 +02:00
  • 3753b30d65 context : fix n_outputs init Georgi Gerganov 2025-02-21 15:50:27 +02:00
  • f588a70da3 context : wrap input tensors in struct Georgi Gerganov 2025-02-21 15:08:25 +02:00
  • ebf1bdf97b context : add logs Georgi Gerganov 2025-02-21 14:35:23 +02:00
  • 586d5fe6eb doc: update contributing guidelines [no ci] (#11969) Johannes Gäßler 2025-02-21 12:51:25 +01:00
  • 548c230dff graph : remove worst_case from the API Georgi Gerganov 2025-02-21 12:10:57 +02:00
  • ecc8e3aeff CUDA: correct the lowest Maxwell supported by CUDA 12 (#11984) b4751 PureJourney 2025-02-21 19:21:05 +08:00
  • 2645a7d9a9 context : add save/load for recurrent context Georgi Gerganov 2025-02-21 10:28:42 +02:00
  • 0b3863ff95 MUSA: support ARM64 and enable dp4a .etc (#11843) Bodhi 2025-02-21 15:46:23 +08:00
  • ee02ad02c5 clip : fix visual encoders with no CLS (#11982) b4749 Alex Brooks 2025-02-20 23:11:03 -07:00
  • 08011c2ca1 context : add llama_kv_cache_recurrent prototype Georgi Gerganov 2025-02-20 20:54:18 +02:00
  • c392e5094d server (webui): Fix Premature Submission During IME Conversion (#11971) momonga 2025-02-21 03:43:22 +09:00
  • ad870c49f4 context : fix causal input for cache-less case Georgi Gerganov 2025-02-20 19:52:42 +02:00
  • b1554be1d7 context : add cache-less llama_context Georgi Gerganov 2025-02-20 15:18:45 +02:00
  • c5d91a7400 ggml-cpu: Add CPU backend support for KleidiAI library (#11390) b4747 Charles Xu 2025-02-20 14:06:51 +01:00
  • 072280ea6b Merge branch 'master' into gg/llama-kv-cache Georgi Gerganov 2025-02-20 14:26:43 +02:00
  • 4806498bf1 ggml: aarch64: implement SVE kernels for q3_K_q8_K vector dot (#11917) b4746 Prashant Vithule 2025-02-20 15:38:32 +05:30
  • 0d559580a0 run : add --chat-template-file (#11961) b4745 Michael Engel 2025-02-20 09:35:11 +01:00
  • d04e7163c8 doc: add links to ggml examples [no ci] (#11958) Johannes Gäßler 2025-02-19 20:45:17 +01:00
  • f95b04a21c model : fix order kvq -> qkv Georgi Gerganov 2025-02-19 18:47:37 +02:00
  • 2eacb4c1bf graph : simplify attention api Georgi Gerganov 2025-02-19 18:43:49 +02:00
  • e17e4b72d1 context : add llama_context_recurrent Georgi Gerganov 2025-02-19 14:56:01 +02:00
  • 5f11a5502a kv-cache : remove llama_kv_cache_i Georgi Gerganov 2025-02-19 14:36:27 +02:00
  • d07c621393 common : add llama.vim preset for Qwen2.5 Coder (#11945) b4743 Daniel Bevenius 2025-02-19 12:29:52 +01:00
  • abd4d0bc4f speculative : update default params (#11954) b4742 Georgi Gerganov 2025-02-19 13:29:42 +02:00
  • 0f2bf55502 speculative : do not discard the last drafted token gg/speculative-update Georgi Gerganov 2025-02-19 09:21:39 +02:00
  • 965ad1c08a speculative : update default params Georgi Gerganov 2025-02-19 08:20:10 +02:00
  • 9626d9351a llama : fix indentation in llama-grammar [no ci] (#11943) Daniel Bevenius 2025-02-19 06:16:23 +01:00
  • b58934c183 server : (webui) Enable communication with parent html (if webui is in iframe) (#11940) igardev 2025-02-19 00:01:44 +02:00
  • f5cedbcaaa kv-cache : prepare for abstraction Georgi Gerganov 2025-02-18 21:26:42 +02:00
  • 63e489c025 tool-call: refactor common chat / tool-call api (+ tests / fixes) (#11900) b4739 Olivier Chafik 2025-02-18 18:03:23 +00:00
  • 63ac128563 server : add TEI API format for /rerank endpoint (#11942) b4738 Xuan-Son Nguyen 2025-02-18 14:21:41 +01:00
  • 2bffc2d514 model : pass llama_graph_i as ptr Georgi Gerganov 2025-02-18 14:57:26 +02:00
  • 9e50456e19 context : minor simplify Georgi Gerganov 2025-02-18 14:53:02 +02:00
  • befe14f06f llama : reorder encode/decode in sources Georgi Gerganov 2025-02-18 14:47:53 +02:00
  • bc6f187e9c cont : use returend tensors from the graph build Georgi Gerganov 2025-02-18 14:24:17 +02:00
  • 172f61690c cont : return important tensors Georgi Gerganov 2025-02-18 13:48:43 +02:00
  • c23590319a graph : add llama_graph_result Georgi Gerganov 2025-02-18 11:16:53 +02:00
  • 5137da7b8c scripts: corrected encoding when getting chat template (#11866) (#11907) MoonRide303 2025-02-18 10:30:16 +01:00
  • 09aaf4f1f5 docs : Fix duplicated file extension in test command (#11935) xiaobing318 2025-02-18 17:12:49 +08:00
  • f0d3ff2388 Merge branch 'master' into gg/llama-kv-cache Georgi Gerganov 2025-02-18 10:14:37 +02:00
  • 73e2ed3ce3 CUDA: use async data loading for FlashAttention (#11894) b4735 Johannes Gäßler 2025-02-17 14:03:24 +01:00