Commit graph

  • 610a03a8c4
    contrib : add naming guidelines Georgi Gerganov 2025-01-10 17:42:43 +02:00
  • 36074d1eb8
    minor [no ci] Georgi Gerganov 2025-01-10 15:54:53 +02:00
  • 940139cd29
    cont Georgi Gerganov 2025-01-10 15:36:23 +02:00
  • 81d852902e
    glahpp: add a space at the end of file Akarshan Biswas 2025-01-10 18:52:49 +05:30
  • 1af85b5118
    SYCL: Add Gated Linear attention kernel Akarshan Biswas 2025-01-10 18:44:48 +05:30
  • 1d9f1f2778
    cont Georgi Gerganov 2025-01-10 15:06:41 +02:00
  • 1586ed5061
    llama : update API names to use correct prefix Georgi Gerganov 2025-01-10 14:40:29 +02:00
  • ba8a1f9c5b
    examples : add README.md to tts example [no ci] (#11155) Daniel Bevenius 2025-01-10 13:16:16 +01:00
  • b6ebd4fc8c fix: ggml: Fix compile error Junil Kim 2025-01-10 19:00:47 +09:00
  • ff3fcabc72
    convert : add --print-supported-models option (#11172) Daniel Bevenius 2025-01-10 11:30:53 +01:00
  • d1af0e9b75
    lora : update API names (#11167) Georgi Gerganov 2025-01-09 22:23:27 +02:00
  • aeeb9420a3 vocab : minor tokenization optimizations (#11160) Georgi Gerganov 2025-01-09 17:12:54 +02:00
  • 0b1bb82fa6
    vocab : minor tokenization optimizations (#11160) Georgi Gerganov 2025-01-09 17:12:54 +02:00
  • a857dc50af
    vocab : more pimpl (#11165) Georgi Gerganov 2025-01-10 10:28:37 +02:00
  • 45aab64e93
    hparams : move vocab params to llama_vocab (#11159) Georgi Gerganov 2025-01-09 16:44:49 +02:00
  • c725f691ea
    llama : add struct llama_vocab to the API (#11156) Georgi Gerganov 2025-01-09 15:28:52 +02:00
  • 609ec7e0a0
    llama : functions -> methods (#11110) Georgi Gerganov 2025-01-06 16:13:01 +02:00
  • 1921b9d39c refactor: ggml: Improve vulkan-shaders-gen toolchain setup Junil Kim 2024-12-24 12:52:04 +09:00
  • 4a17b483c9 refactor: ggml: Improve vulkan-shaders-gen toolchain setup Junil Kim 2024-11-28 12:40:04 +09:00
  • 17b80f080d fix: ggml: fix vulkan-shaders-gen build Junil Kim 2024-11-22 11:37:14 +09:00
  • b48d763583
    vocab : more pimpl (#11165) Georgi Gerganov 2025-01-10 10:28:37 +02:00
  • 446fec5023
    hparams : move vocab params to llama_vocab (#11159) Georgi Gerganov 2025-01-09 16:44:49 +02:00
  • df1c467a72
    llama : add struct llama_vocab to the API (#11156) Georgi Gerganov 2025-01-09 15:28:52 +02:00
  • 7ecf2aab3d
    vocab : more pimpl (#11165) Georgi Gerganov 2025-01-10 10:28:37 +02:00
  • c67e9858b4
    hparams : move vocab params to llama_vocab (#11159) Georgi Gerganov 2025-01-09 16:44:49 +02:00
  • bfe781a42d
    vocab : fix bug (eos -> bos) Georgi Gerganov 2025-01-09 22:40:28 +02:00
  • dfd319c890
    model : fix Phi MoE conflicts Georgi Gerganov 2025-01-09 14:33:32 +02:00
  • cee3648ee3
    llama : vocab cleanup Georgi Gerganov 2025-01-08 22:17:15 +02:00
  • 9dd71e078f
    llama : vocab pimpl cont Georgi Gerganov 2025-01-08 21:42:47 +02:00
  • 615bea8629
    llama : vocabl private charsmap Georgi Gerganov 2025-01-08 20:55:30 +02:00
  • 885495ccd1
    llama : vocab load Georgi Gerganov 2025-01-08 20:49:26 +02:00
  • 695a0037db
    llama : vocab fix names Georgi Gerganov 2025-01-08 20:03:07 +02:00
  • 2c9f20d4bb
    llama : vocab pimpl Georgi Gerganov 2025-01-08 19:47:51 +02:00
  • f4b6969b1d
    llama : vocab Georgi Gerganov 2025-01-08 16:00:34 +02:00
  • 7cf1ae4afb
    llama : remove unicode.h from llama-model.cpp Georgi Gerganov 2025-01-08 15:02:35 +02:00
  • c1d6ae9bd8
    Revert "ninja multi-config -> ninja" slaren 2025-01-07 20:28:57 +01:00
  • d3cbd43cc6
    test slaren 2025-01-07 20:21:29 +01:00
  • dfdc4d786a
    ninja multi-config -> ninja slaren 2025-01-07 18:41:09 +01:00
  • 6860c4bef3
    test Georgi Gerganov 2025-01-07 17:22:07 +02:00
  • 5c8d759a3f
    llama : fix llm_type enum names Georgi Gerganov 2025-01-07 16:04:56 +02:00
  • a15e22537f
    llama : pimpl llama_model Georgi Gerganov 2025-01-07 15:36:39 +02:00
  • a16daa9552
    llama : move load tensors to llama_model Georgi Gerganov 2025-01-06 17:00:16 +02:00
  • 662dd05016
    llama : add llama_model methods Georgi Gerganov 2025-01-06 16:13:01 +02:00
  • fd2672b952 squash! convert : add --print-supported-models option Daniel Bevenius 2025-01-10 08:14:39 +01:00
  • beae79455b convert : add --print-supported-models option Daniel Bevenius 2025-01-10 08:06:47 +01:00
  • 91ab9ed858 update test case Te993 2025-01-10 13:56:29 +08:00
  • 61777707ca add swift test cases Te993 2025-01-10 13:55:59 +08:00
  • d3eeeae218 support omnivlm for ios Te993 2025-01-10 13:53:23 +08:00
  • c3f9d25706
    Vulkan: Fix float16 use on devices without float16 support + fix subgroup_size_control validation error (#11161) b4458 0cc4m 2025-01-10 06:39:33 +01:00
  • ddc3c2208a initial sampling changes: VJHack 2025-01-09 23:04:28 -06:00
  • 305dc66649 vulkan: support copy from q4_0/q4_1/q5_0/q5_1/q8_0/iq4_nl to f32 Jeff Bolz 2025-01-09 22:20:32 -06:00
  • 845d572b87 little stuff Eve 2025-01-09 21:58:26 -05:00
  • 6145fc79e5 q2_k separate out Eve 2025-01-09 21:41:50 -05:00
  • 973bc4069f q3_k separate out calculation Eve 2025-01-09 21:06:05 -05:00
  • ee7136c6d1
    llama: add support for QRWKV6 model architecture (#11001) b4457 Molly Sophia 2025-01-10 09:58:08 +08:00
  • 324afba5cc better sanity check skipping for QRWKV6 in llama-quant Molly Sophia 2025-01-10 09:42:46 +08:00
  • d8a304c2ef Fix fused lerp weights loading with RWKV6 Molly Sophia 2025-01-10 08:41:32 +08:00
  • c6860cc734
    SYCL: Refactor ggml_sycl_compute_forward (#11121) b4456 Akarshan Biswas 2025-01-10 05:43:03 +05:30
  • 51b5ac507d make the caches happy Eve 2025-01-09 17:06:54 -05:00
  • 924bccc214 vulkan: support copy from f32 to q4_0/q4_1/q5_0/q5_1/q8_0/iq4_nl Jeff Bolz 2025-01-09 14:46:19 -06:00
  • d9b07a1690
    vocab : more pimpl Georgi Gerganov 2025-01-09 21:21:14 +02:00
  • 5b87db0802
    Merge branch 'ggerganov:master' into master Jianlin Shi 2025-01-09 12:14:13 -07:00
  • 543fd01eb9
    hparams : remove n_vocab_types Georgi Gerganov 2025-01-09 16:53:17 +02:00
  • c2008b568f
    hparams : remove n_vocab Georgi Gerganov 2025-01-09 16:44:49 +02:00
  • 0f0229736c
    model : avoid hardcoded chat template constant Georgi Gerganov 2025-01-09 20:02:45 +02:00
  • 983aa09b5c Merge branch 'master' into compilade/cuda-tq2_0 Francis Couture-Harpin 2025-01-09 13:02:09 -05:00
  • fb43d5e8b5 ggml-cuda : cleanup TQ2_0 Francis Couture-Harpin 2025-01-09 12:16:02 -05:00
  • 914a82da4d Fix validation error about subgroup_size_control extension 0cc4m 2025-01-09 16:24:00 +01:00
  • 5a392192c1 Vulkan: Remove float16 use in shaders 0cc4m 2025-01-09 15:57:14 +01:00
  • 1929d27954
    SYCL: Some device info print refactoring and add details of XMX availability Akarshan Biswas 2025-01-09 12:07:40 +05:30
  • c2a0a2a02d
    SYCL: add function name to noop debug Akarshan Biswas 2025-01-07 17:04:51 +05:30
  • a28663fcb2
    SYCL: add back GGML_USED(dst) to ggml_sycl_cpy Akarshan Biswas 2025-01-07 16:58:09 +05:30
  • 31f3626b16
    SYCL: refactor ggml_sycl_compute_forward Akarshan Biswas 2025-01-07 16:50:29 +05:30
  • d8931a701c
    llama.android : update to new API Georgi Gerganov 2025-01-09 16:03:09 +02:00
  • 330bd07b82
    llama : llama_n_vocab() now uses struct llama_vocab Georgi Gerganov 2025-01-09 15:57:57 +02:00
  • 68db76595e
    llama : update llama_chat_apply_template Georgi Gerganov 2025-01-09 15:47:13 +02:00
  • f2df367e09 squash! examples : add README.md to tts example [no ci] Daniel Bevenius 2025-01-09 14:36:27 +01:00
  • 22b31cd16d
    llama : expose llama_vocab in the API Georgi Gerganov 2025-01-09 15:28:52 +02:00
  • 98d4e55f5a Style: Adds missing newline Andreas Kieslinger 2025-01-09 13:09:38 +00:00
  • dd95edfcfb Refactor: Removes code permanently excluded from compilation to increase readability. Andreas Kieslinger 2025-01-09 12:44:14 +00:00
  • aefcffabb1
    model : fix Phi MoE conflicts Georgi Gerganov 2025-01-09 14:33:32 +02:00
  • ad1923a0ce
    llama : vocab cleanup Georgi Gerganov 2025-01-08 22:17:15 +02:00
  • f784700c31
    llama : vocab pimpl cont Georgi Gerganov 2025-01-08 21:42:47 +02:00
  • 0f14663a4a
    llama : vocabl private charsmap Georgi Gerganov 2025-01-08 20:55:30 +02:00
  • c949316ca4
    llama : vocab load Georgi Gerganov 2025-01-08 20:49:26 +02:00
  • 40df96e009
    llama : vocab fix names Georgi Gerganov 2025-01-08 20:03:07 +02:00
  • 190f371001
    llama : vocab pimpl Georgi Gerganov 2025-01-08 19:47:51 +02:00
  • 2b150e0a6c
    llama : vocab Georgi Gerganov 2025-01-08 16:00:34 +02:00
  • 6fa9007059
    llama : remove unicode.h from llama-model.cpp Georgi Gerganov 2025-01-08 15:02:35 +02:00
  • f0db5ce0af
    Revert "ninja multi-config -> ninja" slaren 2025-01-07 20:28:57 +01:00
  • 141c40cd0b
    test slaren 2025-01-07 20:21:29 +01:00
  • 5db92f2e82
    ninja multi-config -> ninja slaren 2025-01-07 18:41:09 +01:00
  • e696addb4e
    test Georgi Gerganov 2025-01-07 17:22:07 +02:00
  • a48412f92b
    llama : fix llm_type enum names Georgi Gerganov 2025-01-07 16:04:56 +02:00
  • fffa6b15c4
    llama : pimpl llama_model Georgi Gerganov 2025-01-07 15:36:39 +02:00
  • c2a3fd648e
    llama : move load tensors to llama_model Georgi Gerganov 2025-01-06 17:00:16 +02:00
  • e188b476e6
    llama : add llama_model methods Georgi Gerganov 2025-01-06 16:13:01 +02:00
  • 0cdc133919 Refactor: Moves node graph checks and copy ops into individual function for improved readability. Andreas Kieslinger 2025-01-09 12:16:37 +00:00
  • 9091993a5e squash! examples : add README.md to tts example [no ci] Daniel Bevenius 2025-01-09 12:49:13 +01:00
  • 1204f97270
    doc: add cuda guide for fedora (#11135) Tei Home 2025-01-09 19:32:06 +08:00