3597 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
oobabooga	9ab7365b56	Read rope_theta for DBRX model (thanks turboderp)	2024-04-01 20:25:31 -07:00
oobabooga	db5f6cd1d8	Fix ExLlamaV2 loaders using unnecessary "bits" metadata	2024-03-30 21:51:39 -07:00
oobabooga	624faa1438	Fix ExLlamaV2 context length setting (closes #5750)	2024-03-30 21:33:16 -07:00
oobabooga	70c58b5fc2	Bump ExLlamaV2 to 0.0.17	2024-03-30 21:08:26 -07:00
oobabooga	c37f792afa	Better way to handle user_bio default in the API (alternative to bdcf31035f)	2024-03-29 10:54:01 -07:00
oobabooga	9653a9176c	Minor improvements to Parameters tab	2024-03-29 10:41:24 -07:00
oobabooga	3ce0d9221b	Bump transformers to 4.39	2024-03-28 19:40:31 -07:00
oobabooga	e0e28ecb0b	Set the gradio 4 allowed_paths	2024-03-28 15:10:54 -07:00
oobabooga	723f912c16	Fix the "typing dots" position in latest Gradio version	2024-03-28 12:57:35 -07:00
oobabooga	35da6b989d	Organize the parameters tab (#5767)	2024-03-28 16:45:03 -03:00
dependabot[bot]	3609ea69e4	Bump aqlm[cpu,gpu] from 1.1.0 to 1.1.2 (#5728)	2024-03-26 16:36:16 -03:00
Bartowski	9ad116a6e2	Add config for hyperion and hercules models to use chatml (#5742)	2024-03-26 16:35:29 -03:00
wldhx	7cbafc0540	docker: Remove obsolete CLI_ARGS variable (#5726)	2024-03-26 16:34:53 -03:00
Yiximail	bdcf31035f	Set a default empty string for user_bio to fix #5717 issue (#5722)	2024-03-26 16:34:03 -03:00
Yiximail	8c9aca239a	Fix prompt incorrectly set to empty when suffix is empty string (#5757)	2024-03-26 16:33:09 -03:00
oobabooga	2a92a842ce	Bump gradio to 4.23 (#5758)	2024-03-26 16:32:20 -03:00
oobabooga	49b111e2dd	Lint	2024-03-17 08:33:23 -07:00
oobabooga	d890c99b53	Fix StreamingLLM when content is removed from the beginning of the prompt	2024-03-14 09:18:54 -07:00
oobabooga	d828844a6f	Small fix: don't save truncation_length to settings.yaml ... It should derive from model metadata or from a command-line flag.	2024-03-14 08:56:28 -07:00
oobabooga	2ef5490a36	UI: make light theme less blinding	2024-03-13 08:23:16 -07:00
oobabooga	40a60e0297	Convert attention_sink_size to int (closes #5696)	2024-03-13 08:15:49 -07:00
oobabooga	edec3bf3b0	UI: avoid caching convert_to_markdown calls during streaming	2024-03-13 08:14:34 -07:00
oobabooga	8152152dd6	Small fix after 28076928ac	2024-03-11 19:56:35 -07:00
oobabooga	28076928ac	UI: Add a new "User description" field for user personality/biography (#5691)	2024-03-11 23:41:57 -03:00
oobabooga	63701f59cf	UI: mention that n_gpu_layers > 0 is necessary for the GPU to be used	2024-03-11 18:54:15 -07:00
oobabooga	46031407b5	Increase the cache size of convert_to_markdown to 4096	2024-03-11 18:43:04 -07:00
oobabooga	9eca197409	Minor logging change	2024-03-11 16:31:13 -07:00
oobabooga	afadc787d7	Optimize the UI by caching convert_to_markdown calls	2024-03-10 20:10:07 -07:00
oobabooga	056717923f	Document StreamingLLM	2024-03-10 19:15:23 -07:00
oobabooga	15d90d9bd5	Minor logging change	2024-03-10 18:20:50 -07:00
oobabooga	abcdd0ad5b	API: don't use settings.yaml for default values	2024-03-10 16:15:52 -07:00
oobabooga	a102c704f5	Add numba to requirements.txt	2024-03-10 16:13:29 -07:00
oobabooga	b3ade5832b	Keep AQLM only for Linux (fails to install on Windows)	2024-03-10 09:41:17 -07:00
oobabooga	67b24b0b88	Bump llama-cpp-python to 0.2.56	2024-03-10 09:07:27 -07:00
oobabooga	763f9beb7e	Bump bitsandbytes to 0.43, add official Windows wheel	2024-03-10 08:30:53 -07:00
oobabooga	52a34921ef	Installer: validate the checksum for the miniconda installer on Windows	2024-03-09 16:33:12 -08:00
oobabooga	cf0697936a	Optimize StreamingLLM by over 10x	2024-03-08 21:48:28 -08:00
oobabooga	afb51bd5d6	Add StreamingLLM for llamacpp & llamacpp_HF (2nd attempt) (#5669)	2024-03-09 00:25:33 -03:00
oobabooga	9271e80914	Add back AutoAWQ for Windows ... https://github.com/casper-hansen/AutoAWQ/issues/377#issuecomment-1986440695	2024-03-08 14:54:56 -08:00
oobabooga	549bb88975	Increase height of "Custom stopping strings" UI field	2024-03-08 12:54:30 -08:00
oobabooga	238f69accc	Move "Command for chat-instruct mode" to the main chat tab (closes #5634)	2024-03-08 12:52:52 -08:00
oobabooga	d0663bae31	Bump AutoAWQ to 0.2.3 (Linux only) (#5658)	2024-03-08 17:36:28 -03:00
oobabooga	0e6eb7c27a	Add AQLM support (transformers loader) (#5466)	2024-03-08 17:30:36 -03:00
oobabooga	2681f6f640	Make superbooga & superboogav2 functional again (#5656)	2024-03-07 15:03:18 -03:00
oobabooga	bae14c8f13	Right-truncate long chat completion prompts instead of left-truncating ... Instructions are usually at the beginning of the prompt.	2024-03-07 08:50:24 -08:00
Bartowski	104573f7d4	Update cache_4bit documentation (#5649) ... --------- Co-authored-by: oobabooga <112222186+oobabooga@users.noreply.github.com>	2024-03-07 13:08:21 -03:00
oobabooga	bef08129bc	Small fix for cuda 11.8 in the one-click installer	2024-03-06 21:43:36 -08:00
oobabooga	303433001f	Fix a check in the installer	2024-03-06 21:13:54 -08:00
oobabooga	bde7f00cae	Change the exllamav2 version number	2024-03-06 21:08:29 -08:00
oobabooga	2ec1d96c91	Add cache_4bit option for ExLlamaV2 (#5645)	2024-03-06 23:02:25 -03:00