1350 Commits

Author	SHA1	Message	Date
oobabooga	a43c210617	Improved past chats menu (#6158)	2024-06-25 00:07:22 -03:00
oobabooga	96ba53d916	Handle another fix after 57119c1b30	2024-06-24 15:51:12 -07:00
oobabooga	577a8cd3ee	Add TensorRT-LLM support (#5715)	2024-06-24 02:30:03 -03:00
oobabooga	536f8d58d4	Do not expose alpha_value to llama.cpp & rope_freq_base to transformers ... To avoid confusion	2024-06-23 22:09:24 -07:00
oobabooga	b48ab482f8	Remove obsolete "gptq_for_llama_info" message	2024-06-23 22:05:19 -07:00
oobabooga	5e8dc56f8a	Fix after previous commit	2024-06-23 21:58:28 -07:00
Louis Del Valle	57119c1b30	Update block_requests.py to resolve unexpected type error (500 error) (#5976)	2024-06-24 01:56:51 -03:00
CharlesCNorton	5993904acf	Fix several typos in the codebase (#6151)	2024-06-22 21:40:25 -03:00
GodEmperor785	2c5a9eb597	Change limits of RoPE scaling sliders in UI (#6142)	2024-06-19 21:42:17 -03:00
Guanghua Lu	229d89ccfb	Make logs more readable, no more \u7f16\u7801 (#6127)	2024-06-15 23:00:13 -03:00
Forkoz	1576227f16	Fix GGUFs with no BOS token present, mainly qwen2 models. (#6119) ... --------- Co-authored-by: oobabooga <112222186+oobabooga@users.noreply.github.com>	2024-06-14 13:51:01 -03:00
oobabooga	10601850d9	Fix after previous commit	2024-06-13 19:54:12 -07:00
oobabooga	0f3a423de1	Alternative solution to "get next logits" deadlock (#6106)	2024-06-13 19:34:16 -07:00
oobabooga	386500aa37	Avoid unnecessary calls UI -> backend, to make it faster	2024-06-12 20:52:42 -07:00
Forkoz	1d79aa67cf	Fix flash-attn UI parameter to actually store true. (#6076)	2024-06-13 00:34:54 -03:00
Belladore	3abafee696	DRY sampler improvements (#6053)	2024-06-12 23:39:11 -03:00
oobabooga	a36fa73071	Lint	2024-06-12 19:00:21 -07:00
oobabooga	2d196ed2fe	Remove obsolete pre_layer parameter	2024-06-12 18:56:44 -07:00
Belladore	46174a2d33	Fix error when bos_token_id is None. (#6061)	2024-06-12 22:52:27 -03:00
Belladore	a363cdfca1	Fix missing bos token for some models (including Llama-3) (#6050)	2024-05-27 09:21:30 -03:00
oobabooga	8df68b05e9	Remove MinPLogitsWarper (it's now a transformers built-in)	2024-05-27 05:03:30 -07:00
oobabooga	4f1e96b9e3	Downloader: Add --model-dir argument, respect --model-dir in the UI	2024-05-23 20:42:46 -07:00
oobabooga	ad54d524f7	Revert "Fix stopping strings for llama-3 and phi (#6043)" ... This reverts commit 5499bc9bc8.	2024-05-22 17:18:08 -07:00
oobabooga	5499bc9bc8	Fix stopping strings for llama-3 and phi (#6043)	2024-05-22 13:53:59 -03:00
oobabooga	9e189947d1	Minor fix after bd7cc4234d (thanks @belladoreai)	2024-05-21 10:37:30 -07:00
oobabooga	ae86292159	Fix getting Phi-3-small-128k-instruct logits	2024-05-21 10:35:00 -07:00
oobabooga	bd7cc4234d	Backend cleanup (#6025)	2024-05-21 13:32:02 -03:00
Philipp Emanuel Weidmann	852c943769	DRY: A modern repetition penalty that reliably prevents looping (#5677)	2024-05-19 23:53:47 -03:00
oobabooga	9f77ed1b98	--idle-timeout flag to unload the model if unused for N minutes (#6026)	2024-05-19 23:29:39 -03:00
altoiddealer	818b4e0354	Let grammar escape backslashes (#5865)	2024-05-19 20:26:09 -03:00
Tisjwlf	907702c204	Fix gguf multipart file loading (#5857)	2024-05-19 20:22:09 -03:00
A0nameless0man	5cb59707f3	fix: grammar not support utf-8 (#5900)	2024-05-19 20:10:39 -03:00
Samuel Wein	b63dc4e325	UI: Warn user if they are trying to load a model from no path (#6006)	2024-05-19 20:05:17 -03:00
chr	6b546a2c8b	llama.cpp: increase the max threads from 32 to 256 (#5889)	2024-05-19 20:02:19 -03:00
oobabooga	a38a37b3b3	llama.cpp: default n_gpu_layers to the maximum value for the model automatically	2024-05-19 10:57:42 -07:00
oobabooga	a4611232b7	Make --verbose output less spammy	2024-05-18 09:57:00 -07:00
oobabooga	e9c9483171	Improve the logging messages while loading models	2024-05-03 08:10:44 -07:00
oobabooga	e61055253c	Bump llama-cpp-python to 0.2.69, add --flash-attn option	2024-05-03 04:31:22 -07:00
oobabooga	51fb766bea	Add back my llama-cpp-python wheels, bump to 0.2.65 (#5964)	2024-04-30 09:11:31 -03:00
oobabooga	dfdb6fee22	Set llm_int8_enable_fp32_cpu_offload=True for --load-in-4bit ... To allow for 32-bit CPU offloading (it's very slow).	2024-04-26 09:39:27 -07:00
oobabooga	70845c76fb	Add back the max_updates_second parameter (#5937)	2024-04-26 10:14:51 -03:00
oobabooga	6761b5e7c6	Improved instruct style (with syntax highlighting & LaTeX rendering) (#5936)	2024-04-26 10:13:11 -03:00
oobabooga	4094813f8d	Lint	2024-04-24 09:53:41 -07:00
oobabooga	64e2a9a0a7	Fix the Phi-3 template when used in the UI	2024-04-24 01:34:11 -07:00
oobabooga	f0538efb99	Remove obsolete --tensorcores references	2024-04-24 00:31:28 -07:00
Colin	f3c9103e04	Revert walrus operator for params['max_memory'] (#5878)	2024-04-24 01:09:14 -03:00
oobabooga	9b623b8a78	Bump llama-cpp-python to 0.2.64, use official wheels (#5921)	2024-04-23 23:17:05 -03:00
oobabooga	f27e1ba302	Add a /v1/internal/chat-prompt endpoint (#5879)	2024-04-19 00:24:46 -03:00
oobabooga	e158299fb4	Fix loading sharted GGUF models through llamacpp_HF	2024-04-11 14:50:05 -07:00
wangshuai09	fd4e46bce2	Add Ascend NPU support (basic) (#5541)	2024-04-11 18:42:20 -03:00