vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Commit graph

  • 1117d06607
         opencl : fix element-wise multiplication (#3656) b1395 shibe2 2023-10-18 16:09:22 +04:00
  • c1ca1de2ac fixed support for old falcon models Concedo 2023-10-18 17:20:44 +08:00
  • 700951dbd4 Merge branch 'master' into concedo_experimental Concedo 2023-10-18 16:33:09 +08:00
  • ad2727d091
         Merge branch 'master' into speculative-tree speculative-tree Georgi Gerganov 2023-10-18 10:38:03 +03:00
  • ef9f76e389 Fixed segfault when prompt was too long Christopher Oezbek 2023-10-18 09:12:20 +02:00
  • 0b2ec84246
         Merge c16b80ba4c into cb33f43a2a coezbek 2023-10-18 06:03:08 +00:00
  • c16b80ba4c
         Merge branch 'ggerganov:master' into master coezbek 2023-10-18 08:03:04 +02:00
  • 53b7cdf8a3 Merge branch 'concedo' into concedo_experimental Concedo 2023-10-18 13:51:13 +08:00
  • 5a1f178091
         Add error context with 'raise from' wonjun Jang 2023-10-18 01:30:38 +00:00
  • 9db276f0c2 metal : fix block_q_n_dot_y Jhen 2023-10-18 09:02:58 +08:00
  • a5b26b660b
         Move transformers dependency to local code wonjun Jang 2023-10-18 01:01:50 +00:00
  • 35fd37430f fix zig build FSSRepo 2023-10-17 18:04:26 -04:00
  • c02c52efb5 fix multiple clients FSSRepo 2023-10-17 17:54:56 -04:00
  • d2b1fac6c7 fix make bui;d errors FSSRepo 2023-10-17 17:18:56 -04:00
  • ed0c11cb83 multimodal support enabled by default FSSRepo 2023-10-17 16:58:20 -04:00
  • 6c277eaab5 update api like OpenAI FSSRepo 2023-10-17 16:53:38 -04:00
  • 58f8ae9bfe readme change FSSRepo 2023-10-17 16:32:19 -04:00
  • fa0f22f14f Merge remote-tracking branch 'upstream/master' FSSRepo 2023-10-17 16:31:33 -04:00
  • cb33f43a2a
         fix embeddings when using CUDA (#3657) b1394 slaren 2023-10-17 22:24:50 +02:00
  • aa2268f4cd sync README.md changes FSSRepo 2023-10-17 16:21:05 -04:00
  • e1675d133c
         llama : avoid fprintf in favor of LLAMA_LOG (#3538) b1393 Georgi Gerganov 2023-10-17 22:34:26 +03:00
  • 76b05fc4a0 Avoid a string copy Kerfuffle 2023-10-17 12:13:46 -06:00
  • 8796025b46 Make gguf SpecialVocab vocab size-aware KerfuffleV2 2023-10-17 05:02:41 -06:00
  • 3a007e2c81 Fix convert.py error msg when added tokens are out of range KerfuffleV2 2023-10-17 04:30:13 -06:00
  • 22b914e0ba Minor code cleanups. KerfuffleV2 2023-10-17 03:54:37 -06:00
  • 4079668cda Account for GGUF_KEY_KEY only setting when the key exists KerfuffleV2 2023-10-15 09:48:30 -06:00
  • 32383bbd1c Killll meeeeee KerfuffleV2 2023-10-15 08:16:18 -06:00
  • 14be9d9141 Fix BPE newline check, only I could break something so simple KerfuffleV2 2023-10-15 08:00:53 -06:00
  • d1075f6e08 Add validation for special token ids to llama.cpp KerfuffleV2 2023-10-15 07:25:21 -06:00
  • 1c9c215feb fix embeddings when using CUDA slaren 2023-10-17 20:35:01 +02:00
  • 8402566a7c
         readme : update hot-topics & models, detail windows release in usage (#3615) BarfingLemurs 2023-10-17 14:13:21 -04:00
  • b1b0b518a5
         Merge branch 'master' into master Georgi Gerganov 2023-10-17 21:12:32 +03:00
  • 6e34d31c44
         Update README.md (#479) LostRuins 2023-10-18 01:24:14 +08:00
  • 6043bfa684 OpenCL: Fix element-wise multiplication shibe2 2023-10-11 21:32:15 +04:00
  • 40e5ce054f CLBlast: Fix temporary buffer size for f16 conversion (wsize) b1391 shibe2 2023-10-11 21:30:06 +04:00
  • a5e8c1d8c7
         train-text-from-scratch : fix assert failure in ggml-alloc (#3618) b1390 slaren 2023-10-17 19:00:58 +02:00
  • e74c705e15
         editorconfig : remove trailing spaces Georgi Gerganov 2023-10-17 19:52:53 +03:00
  • 3ad1e3f1a1
         server : documentation of JSON return value of /completion endpoint (#3632) coezbek 2023-10-17 18:51:02 +02:00
  • b7495d6631
         Update examples/server/README.md Georgi Gerganov 2023-10-17 19:50:47 +03:00
  • bd9451ca2a
         Merge branch 'master' into speculative-tree Georgi Gerganov 2023-10-17 19:31:40 +03:00
  • 1142013da4
         save-load-state : fix example + add ci test (#3655) b1387 Georgi Gerganov 2023-10-17 19:12:46 +03:00
  • 5fe268a4d9
         readme : add Aquila2 links (#3610) ldwang 2023-10-17 23:52:33 +08:00
  • b3838fe0f2
         ci : add test for save-load-state example Georgi Gerganov 2023-10-17 18:50:22 +03:00
  • 4e4ef09df0
         save-load-state : fix example (close #3606) Georgi Gerganov 2023-10-17 18:41:42 +03:00
  • 1a159553f9
         tokenizer : special token handling (#3538) b1385 staviq 2023-10-17 17:11:01 +02:00
  • fc82541b1d
         minor : comment + spacing Georgi Gerganov 2023-10-17 18:08:57 +03:00
  • 6f8fe88f10 fix for lite (+5 squashed commit) Concedo 2023-10-15 16:50:03 +08:00
  • 010c52ec59
         Merge branch 'master' into speculative-tree Georgi Gerganov 2023-10-17 17:24:11 +03:00
  • e6dd81f0bc
         speculative : fix the n_drafted fix + p constants Georgi Gerganov 2023-10-17 17:04:31 +03:00
  • f07cd35da4
         speculative : fix off-by-one for n_drafted Georgi Gerganov 2023-10-17 11:40:09 +03:00
  • 281ef73c25
         k-quants : fix quantization ranges (#3646) b1384 Georgi Gerganov 2023-10-17 09:19:28 +03:00
  • 79d4732c70 metal : block_q_n_dot_y for q5_1 (broken) Jhen 2023-10-17 13:45:08 +08:00
  • fce44a7628 metal : implement dequantize_q5_1 Jhen 2023-10-17 13:44:44 +08:00
  • a7a4887be9 metal : revert unnecessary change Jhen 2023-10-17 12:11:26 +08:00
  • 7ebd4acb0a metal : block_q_n_dot_y for block_q5_0 (broken) Jhen 2023-10-17 11:56:07 +08:00
  • 9c3e05d524 metal : implement dequantize_q5_0 Jhen 2023-10-17 11:55:34 +08:00
  • 940efa95fe
         llava : fix tokenization to not add bos between image embeddings and user prompt (#3645) b1383 Georgi Gerganov 2023-10-16 23:58:00 +03:00
  • 4d1804330e fix llava implementation FSSRepo 2023-10-16 16:31:17 -04:00
  • 317dc4bcc2
         k-quants : fix quantization ranges Georgi Gerganov 2023-10-16 22:12:29 +03:00
  • 20131fef6d set seed M. Yusuf Sarıgöz 2023-10-16 22:09:30 +03:00
  • d7eca255d7 context shift fixed FSSRepo 2023-10-16 14:43:10 -04:00
  • e0fb74c6ee
         llava : fix tokenization to not add bos after system prompt Georgi Gerganov 2023-10-16 21:26:04 +03:00
  • 2d9f11db28 fixed premature end due stop word FSSRepo 2023-10-16 12:36:05 -04:00
  • 373d782d42
         minor : comments + rename Georgi Gerganov 2023-10-16 18:17:31 +03:00
  • 1c626e2fe1
         speculative : minor refactor Georgi Gerganov 2023-10-16 12:47:37 +03:00
  • 360a333145
         common : add llama_batch_add() and llama_batch_clear() helpers Georgi Gerganov 2023-10-16 12:41:33 +03:00
  • 005949109d
         prompts : add assistant.txt Georgi Gerganov 2023-10-16 12:41:14 +03:00
  • fd64f04fc2 fix long prompt than ctx proposed in #3639 FSSRepo 2023-10-15 19:07:18 -04:00
  • b727e022d6 fix ci make build undefined ref errors FSSRepo 2023-10-15 18:53:48 -04:00
  • ce961a304b some ci fixes FSSRepo 2023-10-15 18:46:01 -04:00
  • 9035978aae
         Merge pull request #6 from damian0815/fssrepo_mac_fixes Steward Garcia 2023-10-15 18:38:52 -04:00
  • f47fd17b73
         Merge branch 'ggerganov:master' into master Steward Garcia 2023-10-15 18:23:47 -04:00
  • 5b34bfa2e6
         swift : try to fix build Georgi Gerganov 2023-10-16 00:39:57 +03:00
  • b8acb6c9b8
         swift : fix build Georgi Gerganov 2023-10-16 00:20:03 +03:00
  • b5554b9e05
         sampling : fix malloc Georgi Gerganov 2023-10-16 00:09:24 +03:00
  • 0d96efabb5
         batched : fix n_seq_id Georgi Gerganov 2023-10-16 00:03:41 +03:00
  • 5becac802f
         Fix formatting coezbek 2023-10-15 22:46:03 +02:00
  • 7e48e21b1f
         examples : fix build after sampling refactoring Georgi Gerganov 2023-10-15 23:28:41 +03:00
  • 1841ca0ae0 Fixed loadPrompt() when prompt length exceeds context. Christopher Oezbek 2023-10-15 22:23:33 +02:00
  • 4a7f43f28c
         speculative : refactor sampling Georgi Gerganov 2023-10-15 22:30:59 +03:00
  • 1e1490f1ab
         move "Running on Windows" section below "Prepare data and run" BarfingLemurs 2023-10-15 15:18:45 -04:00
  • 32a67cbd16
         speculative : reuse the n_parallel CLI param Georgi Gerganov 2023-10-15 19:35:59 +03:00
  • e513abe37e always use "GGUF" as beginng of GGUF file chenqiny 2023-10-15 23:59:53 +08:00
  • 5a91551899 update headerdoc Damian Stewart 2023-10-15 17:25:13 +02:00
  • d64891b6cf cleanup and refactor *again* Damian Stewart 2023-10-15 17:17:18 +02:00
  • ff12b8fbd6 move back vocab_size where it was 김승덕/Infrastructure그룹(YA) 2023-10-15 21:30:09 +09:00
  • 9eb2b4f39c
         Merge branch 'ggerganov:master' into master seungduk.kim.2304 2023-10-15 21:25:55 +09:00
  • a861879256 Add q2_k and q3_k support 0cc4m 2023-10-15 14:16:05 +02:00
  • 177845089f
         Check tokenizer.json when tokenizer.model is not exist. wonjun Jang 2023-10-15 11:34:25 +00:00
  • e876aec189
         Add newline between classes wonjun Jang 2023-10-15 10:37:41 +00:00
  • 6fc75811d5
         Added documentation of JSON return value of /completion endpoint coezbek 2023-10-15 12:03:05 +02:00
  • 1f16e5f234
         remove byte_encoder wonjun Jang 2023-10-15 09:46:48 +00:00
  • 57a8f146a0 Address undeclared identifiers by adding conditional compilation dane madsen 2023-10-15 19:45:33 +10:00
  • 6ec856b3ef
         remove debug code wonjun Jang 2023-10-15 09:44:07 +00:00
  • c7b636e950
         change add_meta_vocab fucntion wonjun Jang 2023-10-15 09:43:22 +00:00
  • ea9f35f082
         add bytes_to_unicode function wonjun Jang 2023-10-15 09:42:03 +00:00
  • f888d2ea13
         Update convert.py wonjun Jang 2023-10-15 09:40:14 +00:00
  • b0e00cb868
         Update convert.py wonjun Jang 2023-10-15 09:38:58 +00:00
  • f7e377d62f
         Add HFVocab into convert.py wonjun Jang 2023-10-15 09:32:22 +00:00
  • ee0681f0d9 convert some asserts into non-terminating since they are ovezealous Concedo 2023-10-15 16:12:20 +08:00