Justine Tunney [Wed, 29 Mar 2023 20:51:37 +0000 (13:51 -0700)]
Make loading weights 10-100x faster
This is a breaking change that's going to give you three benefits:
1. Your inference commands should load 100x faster
2. You may be able to safely load models 2x larger
3. You can run many concurrent inference processes
This was accomplished by changing the file format so we can mmap()
weights directly into memory without having to read() or copy them
thereby ensuring the kernel can make its file cache pages directly
accessible to our inference processes; and secondly, that the file
cache pages are much less likely to get evicted (which would force
loads to hit disk) because they're no longer competing with memory
pages that were needlessly created by gigabytes of standard i/o.
The new file format supports single-file models like LLaMA 7b, and
it also supports multi-file models like LLaMA 13B. Our Python tool
now merges the foo.1, foo.2, etc. files back into a single file so
that the C++ code which maps it doesn't need to reshape data every
time. That's made llama.cpp so much simpler. Much of its load code
has now been deleted.
Furthermore, this change ensures that tensors are aligned properly
on a 32-byte boundary. That opens the door to seeing if we can get
additional performance gains on some microprocessors, by using ops
that require memory alignment.
Lastly note that both POSIX and the Windows platform are supported
Casey Primozic [Thu, 30 Mar 2023 17:53:35 +0000 (10:53 -0700)]
Remove unused variable (#607)
* It seems some new warning were added recently that exposed this. I wrote the code that included this unused variable originally and it is indeed not needed.
DooWoong Lee (David) [Tue, 28 Mar 2023 17:02:34 +0000 (02:02 +0900)]
py : removed unused `model` variable and verified that the code functions correctly with `vocab_only` setting. Also confirmed that the code works as expected after running with reduced memory usage due to deletion of no-longer-needed variable. (#547)
anzz1 [Tue, 28 Mar 2023 14:09:55 +0000 (17:09 +0300)]
main.cpp fixes, refactoring (#571)
- main: entering empty line passes back control without new input in interactive/instruct modes
- instruct mode: keep prompt fix
- instruct mode: duplicate instruct prompt fix
- refactor: move common console code from main->common
anzz1 [Sat, 25 Mar 2023 21:38:11 +0000 (23:38 +0200)]
CMake / CI additions (#497)
* CMake: Add AVX512 option
* CI: Add AVX/AVX512 builds (Windows)
(AVX512 tests can only be run when the worker happens to support it, building works anyway)
* CMake: Fix sanitizer linkage ( merged #468 )
* CI: Add sanitizer builds (Ubuntu)
* CI: Fix release tagging
(change @zendesk/action-create-release to @anzz1/action-create-release until upstream PR Added commitish as input zendesk/action-create-release#32 is merged)
rabidcopy [Fri, 24 Mar 2023 15:22:39 +0000 (10:22 -0500)]
fix instruct mode (#445)
changes to EOS behavior in interactive and reverse prompt handling broke instruct mode by erroneously injecting instruct mode's reverse prompt and an extra newline.
comex [Fri, 24 Mar 2023 15:19:05 +0000 (08:19 -0700)]
Support calling mlock() on loaded model data on Linux and macOS (#453)
* Support calling mlock() on loaded model data on Linux and macOS
This is enabled by a new --mlock command line option.
Using mlock() disables swapping and memory compression for the model
data. Doing so can be useful on systems where the model takes up a
large fraction of system RAM. In my experience, macOS is quite eager to
start compressing llama.cpp's memory, which then makes it halt for a few
seconds while it decompresses, even with a model that uses "only" 25GB
out of 32GB.
Of course, this comes at the cost of forcing the system to swap or
compress other processes' memory instead, so it needs to be used with
care and shouldn't be enabled by default.
In theory it should be possible to support this on Windows as well using
VirtualLock(), but I'm not much of a Windows user.
rabidcopy [Thu, 23 Mar 2023 20:22:47 +0000 (15:22 -0500)]
Replace EOS with newline to prevent context/memory being flushed by EOS in interactive mode (#333)
* Improve interactive mode's coherence after EOS
Aims to improve coherence and ability to resume the interactive session when the user is given input back after an end of text token is reached.
Not sure what token 13 is or why it seems to help. See conversation for examples.
* Make newline token a constant
* dynamically determine newline token
* relocate previous newline token const
* cleanup whitespace
* print a new line on end of text in interactive
this may need to be looked into further when not using a reverse prompt
* only print manual newline with reverse prompt
fix formatting of reverse prompts so they don't end up at the end of the current line while not introducing unnecessary new lines otherwise
* alternate approach to replace end of text tokens
* Inject the reverse prompt again after eos in interactive mode
* tokenize reverse prompt when needed
makes this PR compatible with https://github.com/ggerganov/llama.cpp/pull/330
* tokenize and inject only first reverse prompt
thanks to tjohnman
* tokenize first reverse prompt once
* add newline token
* add newline token
* tokenize/inject reverse prompt for refactor
this doesn't seem right though
* tokenize nothing for antiprompt if no reverse
* Update main.cpp
* Update main.cpp
* tokenize and inject reverse prompt as needed
this doesn't seem to work if the reverse prompt is tokenized outside earlier on
* not needed
* remove newline token
* remove newline token
* tokenize newline token
* add space to comment
* Update main.cpp
Co-authored-by: Georgi Gerganov <redacted>
---------
Co-authored-by: Slaren <redacted> Co-authored-by: Georgi Gerganov <redacted>
overview / pkg / ggml / sources / llama.cpp / log