Commit Graph

  • 8171e621fc sync : ggml (Metal fixes, new ops, tests) (#1633) Georgi Gerganov 2023-12-13 21:55:03 +02:00
  • ec03661b20 cmake : target windows 8 or above for prefetchVirtualMemory in llama-talk (#1617) Kreijstal 2023-12-12 12:35:00 +01:00
  • 6335933a5b cmake : Fix bug in httplib.h for mingw (#1615) Kreijstal 2023-12-10 18:47:52 +01:00
  • 885b5563d0 metal : fix ggml_metal_log vargs (#1606) Finn Voorhees 2023-12-08 11:50:50 +00:00
  • 9521ba6801 whisper.objc : disable timestamps for real-time transcription Georgi Gerganov 2023-12-08 13:43:37 +02:00
  • 29511d33c7 whisper : more debug messages + fix fallback logic Georgi Gerganov 2023-12-08 13:43:03 +02:00
  • 7bc4d22337 metal : fix soft_max kernel src1 argument (#1602) Georgi Gerganov 2023-12-08 13:39:32 +02:00
  • afce6fa113 sync : ggml (new ops, new backend, etc) (#1602) Georgi Gerganov 2023-12-07 22:27:19 +02:00
  • 3163090d89 server : pass max-len argument to the server (#1574) Oleg Sidorov 2023-12-05 22:01:45 +01:00
  • f0efd0202d ios : Remove #if arch(arm) check for using Metal (#1561) Finn Voorhees 2023-12-05 01:14:26 +00:00
  • 3c28d1a571 ggml : Fix 32-bit compiler warning (#1575) Digipom 2023-12-03 09:15:28 -05:00
  • e369243ebd ggml : re-enable blas for src0 != F32 (#1583) Georgi Gerganov 2023-12-01 23:57:52 +02:00
  • a0ec3fac54 Server : Add support for .vtt format to Whisper server (#1578) Aleksander Andrzejewski 2023-12-01 00:44:26 +01:00
  • 15c4fdce45 chess : tuning performance gg/chess Georgi Gerganov 2023-11-30 10:50:47 +02:00
  • 70741ba794 wchess: c++17 -> c++11 Fraxy V 2023-11-30 08:37:54 +02:00
  • bb723282cc wchess: off/on prompt Fraxy V 2023-11-30 01:17:29 +02:00
  • dc5513a709 wchess: prompt Fraxy V 2023-11-29 19:30:57 +02:00
  • ffc244845b wchess : dynamic grammar Fraxy V 2023-11-29 18:53:28 +02:00
  • 8962a6bd67 wchess: preparing dyn grammar Fraxy V 2023-11-29 15:29:16 +02:00
  • d313034b9c wchess grammar tweaks Fraxy V 2023-11-29 09:25:45 +02:00
  • 8b0b0acff3 wchess : remove vad Fraxy V 2023-11-28 19:03:17 +02:00
  • 02ade14f67 wchess minor Fraxy V 2023-11-28 16:21:46 +02:00
  • 8dba8204eb Merge pull request #1 from ggerganov/gg/wchess fraxy-v 2023-11-28 15:45:17 +02:00
  • 6559b538e5 server : backport .srt output format (#1565) Oleg Sidorov 2023-11-28 14:42:58 +01:00
  • 73d5005880 cmake : install required ggml.h header (#1568) Gregor Jasny 2023-11-28 14:41:49 +01:00
  • 4260d4fc70 wchess : minor gg/wchess Georgi Gerganov 2023-11-28 15:10:18 +02:00
  • ee65df7982 wchess : add clear_audio callback Georgi Gerganov 2023-11-28 13:37:26 +02:00
  • 6b094b6dfe server : set default CORS headers to allow all (#1567) Kasumi 2023-11-28 17:55:20 +08:00
  • 641f2f4282 readme : update help (#1560) Hang 2023-11-27 18:04:08 +08:00
  • bfacd9f8ce CI : Add CUDA 11.8.0 support (#1554) bobqianic 2023-11-27 10:03:16 +00:00
  • c8b3bc6a0d cuda : use CUBLAS_COMPTE_F32 insted of CUBLAS_COMPUTE_F16 cuda-cublas-opts Georgi Gerganov 2023-11-27 11:57:07 +02:00
  • f52e74d4dc CI : Rectify the Clang-Related workflow issues (#1551) bobqianic 2023-11-27 09:35:37 +00:00
  • 23c21e92eb server : automatically convert audio on the server (#1539) Ismatulla Mansurov 2023-11-27 02:28:34 -07:00
  • 03f254193b wchess: hardcoded rules Fraxy V 2023-11-27 10:51:20 +02:00
  • 8f2d8eae10 wchess: basic chess rules Fraxy V 2023-11-27 10:41:04 +02:00
  • a44b21bce0 wchess: tidy up entry files Fraxy V 2023-11-25 11:34:06 +02:00
  • f07ff2aa6a chess -> wchess Fraxy V 2023-11-25 10:16:48 +02:00
  • 447d49530c whisper : remove trailing whitespaces Georgi Gerganov 2023-11-24 13:13:12 +02:00
  • 9d6ebd877c release : v1.5.1 v1.5.1 Georgi Gerganov 2023-11-24 12:41:55 +02:00
  • 0ba365f958 metal : add backend function to check device family support (#1547) Georgi Gerganov 2023-11-24 12:37:08 +02:00
  • 010c8ec3ab cuda : sync some minor stuff from llama.cpp (#1548) Georgi Gerganov 2023-11-24 12:36:21 +02:00
  • ffdb5c4735 whisper : fix typo Georgi Gerganov 2023-11-24 09:45:10 +02:00
  • a5881d619c server : add --print-realtime param (#1541) ecneladis 2023-11-24 08:35:02 +01:00
  • 34f70b3a56 whisper : add whisper_lang_str_full (#1546) bradmit 2023-11-24 18:33:13 +11:00
  • 8328d1900f fix(server): typo in temperature parameter (#1545) Okabintaro 2023-11-23 19:59:36 +01:00
  • d2bd5f0bdc metal : fix build (#1544) sandrohanea 2023-11-23 19:20:53 +01:00
  • 34209a37a2 readme : add server example Georgi Gerganov 2023-11-23 17:20:33 +02:00
  • 280e631bcf chess.wasm: poc of chess rules Fraxy V 2023-11-23 16:09:00 +02:00
  • 2f86da0d09 chess.wasm: add chessboard Fraxy V 2023-11-23 08:49:47 +02:00
  • 180e062eda go : fixed Makefile for MacOS ARM 64 (#1530) Gleicon Moraes 2023-11-22 13:08:11 -03:00
  • 5c7be85fdc Change temp file name for server application (#1535) Felix 2023-11-22 09:23:36 +01:00
  • 146169ec38 bench : pass memcpy threads from cli Georgi Gerganov 2023-11-21 22:27:22 +02:00
  • 9befab5ab9 bench : multi-thread memcpy (#1534) Georgi Gerganov 2023-11-21 22:07:30 +02:00
  • ee2971bf6a bench : multi-thread memcpy bench-memcpy Georgi Gerganov 2023-11-21 21:42:00 +02:00
  • 9ac88f2b57 Close file after writing in server application (#1533) Felix 2023-11-21 20:36:10 +01:00
  • a787f7f85c chess.wasm: encoder context value resulting in echoing Fraxy V 2023-11-21 20:42:20 +02:00
  • 46f5b6cb08 server : add video to readme Georgi Gerganov 2023-11-21 17:30:43 +02:00
  • c83a38e89d chess.wasm: go back to greedy Fraxy V 2023-11-21 16:56:22 +02:00
  • 758c951729 chess.wasm: grammar in emscripten Fraxy V 2023-11-21 16:30:44 +02:00
  • eff3570f78 server : add a REST Whisper server example with OAI-like API (#1380) Felix 2023-11-20 20:40:24 +01:00
  • fa19bc4195 whisper : update example in whisper.h (#1529) M. A. Ali 2023-11-20 20:52:27 +02:00
  • a01b2e0971 sdl : fix audio callback (#1523) Georgi Gerganov 2023-11-20 13:16:38 +02:00
  • 8159a9ab99 whisper : reuse whisper_decode_with_state (#1521) Georgi Gerganov 2023-11-20 13:16:11 +02:00
  • 7516d9c16d ci : redistribute CUDA DLLs (#1522) Tamotsu Takahashi 2023-11-19 19:43:22 +09:00
  • 46cc26d1b9 whisper : fix with_state methods to use the correct state (#1519) sandrohanea 2023-11-19 10:25:30 +01:00
  • f784f9fa12 whisper : fix overriding the audio context Georgi Gerganov 2023-11-19 10:32:32 +02:00
  • ca23f8ee6d cuda : assert ggml_add sources to be contiguous Georgi Gerganov 2023-11-19 10:32:08 +02:00
  • e2f0eba2d4 ios : sync submodule Georgi Gerganov 2023-11-17 10:42:04 +02:00
  • d4353e48f7 sync : ggml (ggml-alloc + linker + gguf fixes) (#1501) Georgi Gerganov 2023-11-17 10:00:07 +02:00
  • ec96d68402 whisper : quantize encoder only quantize-encoder Georgi Gerganov 2023-11-16 16:19:02 +02:00
  • bebf0da983 quantize : add support for K-quant types Georgi Gerganov 2023-11-16 16:18:24 +02:00
  • 848e54f3ad bench : fix memcpy bench size Georgi Gerganov 2023-11-16 10:59:32 +02:00
  • 7883d1cae4 talk-llama : improve quote and backtick handling (#1364) Sam Pullara 2023-11-16 00:34:05 -08:00
  • ccc85b4ff8 talk-llama : enable GPU by default Georgi Gerganov 2023-11-15 21:32:25 +02:00
  • c7606b47df models : add info about distilled models Georgi Gerganov 2023-11-15 21:10:13 +02:00
  • d38af151a1 release : v1.5.0 v1.5.0 Georgi Gerganov 2023-11-15 21:02:52 +02:00
  • 94267df08e bench-all : add distil models Georgi Gerganov 2023-11-15 20:49:12 +02:00
  • 8713c67133 js : latest whisper.js Georgi Gerganov 2023-11-15 20:10:16 +02:00
  • 57a60639bb bench-all : indentations Georgi Gerganov 2023-11-15 20:01:15 +02:00
  • bfbaa4dce5 whisper : make large version explicit + fix data size units (#1493) Georgi Gerganov 2023-11-15 19:42:25 +02:00
  • 1d79e78402 java : fix test (#1492) Georgi Gerganov 2023-11-15 17:42:53 +02:00
  • b6c5f49b78 whisper : add batched decoding (#1486) Georgi Gerganov 2023-11-15 16:12:52 +02:00
  • 270b1e48db cuda : sync llama.cpp fixes batched Georgi Gerganov 2023-11-15 15:52:06 +02:00
  • b7c82a37b1 whisper : enable beam-search by default Georgi Gerganov 2023-11-15 15:36:45 +02:00
  • 4c245ea108 whisper : avoid starting sampling threads with bs=1 Georgi Gerganov 2023-11-15 13:41:38 +02:00
  • 820f45895e whisper : add check for max number of decoders Georgi Gerganov 2023-11-15 13:28:16 +02:00
  • 6c8a003acf whisper : add comment about the KV cache size Georgi Gerganov 2023-11-14 23:00:57 +02:00
  • ae1bd69041 bench : add batch size 5 bench Georgi Gerganov 2023-11-14 22:45:08 +02:00
  • 3ed9af34f2 whisper : fix decoders initializer Georgi Gerganov 2023-11-14 22:14:52 +02:00
  • 9006946e4b whisper : speed-up sampling Georgi Gerganov 2023-11-14 22:05:31 +02:00
  • d77603578b whisper : clear kv cache when using whisper_decode API Georgi Gerganov 2023-11-14 21:04:33 +02:00
  • b2123cb463 whisper : fix mem leak again + remove oboslete function Georgi Gerganov 2023-11-14 19:12:21 +02:00
  • 3d24e35f49 whisper : fix memory leak in whisper_batch Georgi Gerganov 2023-11-14 17:14:22 +02:00
  • 91096daa1a whisper : full batched decoding support Georgi Gerganov 2023-11-14 16:57:28 +02:00
  • 8b943f9843 whisper : move kv_self to whisper_state Georgi Gerganov 2023-11-14 11:04:15 +02:00
  • 3cbaaed060 whisper : add whisper_batch Georgi Gerganov 2023-11-13 22:39:43 +02:00
  • d4231649e6 java : use tiny.en for tests (#1484) Georgi Gerganov 2023-11-13 16:53:55 +02:00
  • 3e5c7feeff whisper : add grammar-based sampling (#1229) Evan Jones 2023-11-13 03:51:34 -05:00
  • c23598e4ca talk-llama : add n_gpu_layers parameter (#1475) rlapray 2023-11-13 09:04:16 +01:00
  • 54a08bde29 examples : add whisper.android.java for compatibility with older Android versions using Java (#1382) Tong Li 2023-11-12 06:31:58 -10:00