565 commits

Author	SHA1	Message	Date
Concedo	73b99a7266	add premade chat completions adapter	2024-06-27 00:13:06 +08:00
Concedo	e42bc5d677	add negative prompt support to chat completions adapter	2024-06-26 11:12:24 +08:00
Concedo	151ff95a67	Merge branch 'upstream' into concedo_experimental ... # Conflicts: # CMakeLists.txt # Makefile # README.md # ggml-cuda.cu # ggml-cuda/common.cuh	2024-06-25 19:25:14 +08:00
Concedo	13398477a1	fix ubatch, autoselect vulkan dgpu if possible	2024-06-22 00:23:46 +08:00
Nexesenex	153527745b	Augmented benchmark stats (#929) ... * Augmented benchmark stats v1 * output instead of coherence * populate bench flags as a flags field instead of multiple lines --------- Co-authored-by: Concedo <39025047+LostRuins@users.noreply.github.com>	2024-06-18 21:30:36 +08:00
Concedo	ba9ef4d01b	fix to allow clblast to work even after blas backend splitoff	2024-06-17 15:02:55 +08:00
Concedo	623390e4ab	allow sdui when img model not loaded, allow sdclamped to provide a custom clamp size (+1 squashed commits) ... Squashed commits: [957c9c9c] allow sdui when img model not loaded, allow sdclamped to provide a custom clamp size	2024-06-14 16:58:50 +08:00
Concedo	e69da9c9d8	strings rename kobold lite to koboldai lite	2024-06-13 20:00:28 +08:00
Concedo	49e4c3fd7b	adjust lite default port, disable double BOS warning, whisper and SD go quiet when horde mode is set too	2024-06-13 15:10:35 +08:00
Concedo	02357eadf8	Merge commit '7672adeec7' into concedo_experimental ... # Conflicts: # CMakeLists.txt # Makefile # kompute-shaders/op_rope_f16.comp # kompute-shaders/op_rope_f32.comp # kompute-shaders/rope_common.comp # tests/test-backend-ops.cpp # tests/test-grad0.cpp # tests/test-rope.cpp	2024-06-09 15:35:51 +08:00
Concedo	813cf829b5	allow selecting multigpu on vulkan	2024-06-06 18:36:56 +08:00
Concedo	10b148f4c2	added skip bos for tokenize endpoint	2024-06-05 10:49:11 +08:00
Concedo	a541a3d509	quantkv will not trigger if fa is off or ctx shift is on	2024-06-03 19:14:22 +08:00
Concedo	efee37a708	gui for quantkv	2024-06-03 18:25:57 +08:00
Concedo	10a1d628ad	added new binding fields for quant k and quant v	2024-06-03 14:35:59 +08:00
Concedo	267ee78651	change max payload to 32mb	2024-06-02 16:44:19 +08:00
Concedo	b0a7d1aba6	fixed makefile (+1 squashed commits) ... Squashed commits: [ef6ddaf5] try fix makefile	2024-06-02 15:21:48 +08:00
Concedo	9e64f0b5af	added whisper file upload mode	2024-06-02 12:04:56 +08:00
Concedo	a65e0800ab	update docs, added gui for whisper	2024-06-01 02:01:49 +08:00
Concedo	961c789c91	wav file resampling	2024-05-30 13:41:58 +08:00
Concedo	62ab344b1e	transcribe api is functional	2024-05-30 00:07:53 +08:00
Concedo	f24aef8792	initial whisper integration	2024-05-29 23:13:11 +08:00
Concedo	dd59303ae1	Merge branch 'concedo_experimental' of https://github.com/LostRuins/koboldcpp into concedo_experimental	2024-05-28 18:25:13 +08:00
Concedo	38d4d743bb	add flash attn and quiet mode to quick launch	2024-05-28 18:25:00 +08:00
jojorne	dc53e30785	Why not search for cuda_path as well? (#865) ... Let's add dll directory for cuda on Windows too.	2024-05-27 21:38:17 +08:00
Concedo	27e784a42d	up ver	2024-05-25 00:03:22 +08:00
Concedo	fac6373b13	fix tools	2024-05-24 23:50:08 +08:00
Concedo	09adfa70ad	limit default threads to max 8 to deal with ecores	2024-05-22 14:47:57 +08:00
Concedo	618e60c279	model download if its a url	2024-05-21 18:56:11 +08:00
Concedo	2cbf39cba2	disable ui resize on macos	2024-05-17 15:56:10 +08:00
Concedo	1db3421c52	multiple minor fixes	2024-05-17 15:47:53 +08:00
Concedo	6d9d846bdd	prevent mixing lora and quant	2024-05-16 00:29:03 +08:00
Concedo	08993696c3	try apply lora on load	2024-05-15 22:53:23 +08:00
Concedo	44443edfda	rep pen slope works (+1 squashed commits) ... Squashed commits: [535ad566] experiment with rep pen range	2024-05-15 17:20:57 +08:00
Concedo	5ce2fdad24	taesd for sdxl, add lora loading done	2024-05-14 23:02:56 +08:00
Concedo	2ee808a747	Merge branch 'upstream' into concedo_experimental ... # Conflicts: # .github/workflows/build.yml # CMakeLists.txt # README.md # ci/run.sh # llama.cpp # models/ggml-vocab-llama-bpe.gguf.inp # models/ggml-vocab-llama-bpe.gguf.out # requirements.txt # scripts/compare-llama-bench.py # scripts/sync-ggml.last # tests/CMakeLists.txt # tests/test-backend-ops.cpp # tests/test-grammar-integration.cpp # tests/test-tokenizer-1-bpe.cpp	2024-05-14 19:28:47 +08:00
Concedo	5d15f8f76a	vae test	2024-05-14 19:17:01 +08:00
Concedo	4807b66907	wip sd	2024-05-13 23:23:16 +08:00
Concedo	bd95ee7d9a	temporary version for archiving	2024-05-13 21:53:58 +08:00
Concedo	d8a52321da	ditched the coherent flag	2024-05-13 20:38:51 +08:00
Concedo	f4746572d9	wildcare sdui url	2024-05-12 11:09:59 +08:00
Concedo	eff01660e4	re-added smart context due to people complaining	2024-05-11 17:25:03 +08:00
Concedo	702be65ed1	don't show embedded sdui if no model	2024-05-11 08:56:56 +08:00
Concedo	1effe16861	fixed horde worker flag	2024-05-11 01:17:04 +08:00
Concedo	7967377ebc	fix for sdui showing when sdmodel not loaded, and not showing when remote tunnel is used.	2024-05-10 23:40:20 +08:00
Concedo	69570daf31	tidy argparse	2024-05-10 17:28:08 +08:00
Concedo	dbe72b959e	tidy up and refactor code to support old flags	2024-05-10 16:50:53 +08:00
Concedo	eccc2ddca2	better warnings	2024-05-10 11:27:40 +08:00
Concedo	6f23ca24fb	deprecated some old flags	2024-05-10 10:57:52 +08:00
Concedo	868446bd1a	replace sdconfig and hordeconfig	2024-05-09 22:43:50 +08:00