776 Commits (5a74b552f900da3e481fdba6712f9c3853482c70)

 

All Branches   

Search

Author	SHA1	Message	Date
Adam Treat	ff257eb52c	Add MPT info to the download list and fix it so that isDefault will work even if the required version isn't there.	1 year ago
Adam Treat	5d95085cbe	Move this script and rename.	1 year ago
Adam Treat	8eeca20fd7	Simplify.	1 year ago
Adam Treat	8d295550eb	Don't keep this in memory when it is not needed.	1 year ago
Adam Treat	7094fd0788	Gracefully handle when we have a previous chat where the model that it used has gone away.	1 year ago
Adam Treat	ad82aaebb1	Copy pasta.	1 year ago
Adam Treat	9c66308922	Fix for special im_end token in mpt-7b-chat model.	1 year ago
Adam Treat	a4bec78ec6	Allow these to load for gptj too.	1 year ago
Aaron Miller	821b28a4fa	mpt: allow q4_2 quantized models to load	1 year ago
Aaron Miller	49fc7b315a	mpt tokenizer: better special token handling ... closer to the behavior of huggingface `tokenizers`, do not attempt to handle additional tokens as if they were part of the original vocabulary as this cannot prevent them from being split into smaller chunks - handle added tokens *before* the regular tokenizing pass note this is still necessary even with a "proper" tokenizer implementation	1 year ago
Adam Treat	9da4fac023	Fix gptj to have lower memory requirements for kv cache and add versioning to the internal state to smoothly handle such a fix in the future.	1 year ago
Adam Treat	c7f5280f9f	Fix the version.	1 year ago
Adam Treat	be9e748abe	Remove as upstream has removed.	1 year ago
Adam Treat	126dd97b0a	This shouldn't have snuck in.	1 year ago
Adam Treat	3a8ad1f700	Update to the alibi version that Zach made.	1 year ago
Adam Treat	90b2bcfebe	Match Helly's impl of kv cache.	1 year ago
Adam Treat	368886015d	Use F16 for kv cache on mpt.	1 year ago
Adam Treat	00804c4e3e	Fix for special tokens.	1 year ago
Adam Treat	98e19ebc25	Fix up mpt.	1 year ago
Zach Nussbaum	712aeb8866	fix: helly changes	1 year ago
Zach Nussbaum	d14b93222f	fix: model loading	1 year ago
Zach Nussbaum	28f0f76b9f	fix: convert script working	1 year ago
Zach Nussbaum	d928540a08	feat: load model	1 year ago
Zach Nussbaum	f8f248c18a	chore: import for mpt	1 year ago
Zach Nussbaum	e3f17c8e82	feat: mpt convert from hf to ggml	1 year ago
Zach Nussbaum	285e57ca68	feat: build works + tokenizer	1 year ago
Zach Nussbaum	199a585ad1	feat: add ln 2, rename vars	1 year ago
Zach Nussbaum	21f2aa4911	feat: mpt wip	1 year ago
Adam Treat	a066cba17d	Scaffolding for the mpt <-> ggml project.	1 year ago
Adam Treat	da5b057041	Only generate three words max.	1 year ago
Adam Treat	2b76fa6b20	Restore defaults for repeat penalty too.	1 year ago
Adam Treat	ee016e10ab	Send info on how many are running into this error.	1 year ago
Adam Treat	e0c9d7f8e0	Fail early/gracefully if incompatible hardware detected. And default to universal builds on mac.	1 year ago
Adam Treat	4bcc88b051	Convert the old format properly.	1 year ago
Adam Treat	fb464bb60e	Add debug for chatllm model loading and fix order of getting rid of the ... dummy chat when no models are restored.	1 year ago
Adam Treat	3a039c8dc1	Deserialize one at a time and don't block gui until all of them are done.	1 year ago
Adam Treat	fc8c158fac	Use last lts for this.	1 year ago
Adam Treat	280ad04c63	The GUI should come up immediately and not wait on deserializing from disk.	1 year ago
Adam Treat	ec7ea8a550	Move the location of the chat files to the model download directory and add a magic+version.	1 year ago
Aaron Miller	516a7ffa23	add name to LICENSE	1 year ago
Adam Treat	eb294d5623	Bump the version and save up to an order of magnitude of disk space for chat files.	1 year ago
Adam Treat	6ba0a1b693	Turn off saving chats to disk by default as it eats so much disk space.	1 year ago
Adam Treat	ba76cecbdf	Bump the version.	1 year ago
Adam Treat	b7b2ff8bab	Add reverse prompt support for gptj too.	1 year ago
Adam Treat	c2a81e5692	Add about dialog.	1 year ago
Adam Treat	cd83723ed7	Persistent state for gpt-j models too.	1 year ago
Adam Treat	a548448fcf	Don't crash if state has not been set.	1 year ago
Richard Guo	561acf81d7	Update monorepo_plan.md	1 year ago
Aaron Miller	56e9fd7e63	include <cstdint> in llmodel.h	1 year ago
Adam Treat	01e582f15b	First attempt at providing a persistent chat list experience. ... Limitations: 1) Context is not restored for gpt-j models 2) When you switch between different model types in an existing chat the context and all the conversation is lost 3) The settings are not chat or conversation specific 4) The sizes of the chat persisted files are very large due to how much data the llama.cpp backend tries to persist. Need to investigate how we can shrink this.	1 year ago