 Alex Cheema
|
85279007b3
hotfix edge case where we try to render before tokenizer is set
|
11 ماه پیش |
|
Alex Cheema
|
09a8468395
upgrade mlx to 0.17.0
|
11 ماه پیش |
|
Alex Cheema
|
1f9d16ec78
run tokenizers test in ci, run all models available
|
11 ماه پیش |
|
Alex Cheema
|
6243846eeb
ci logs
|
11 ماه پیش |
|
Alex Cheema
|
cfe980bdaa
simplify ci
|
11 ماه پیش |
|
Alex Cheema
|
9513c4fd17
ci tail log files
|
11 ماه پیش |
|
Alex Cheema
|
7a02acdcd5
fix ci output streaming
|
11 ماه پیش |
|
Alex Cheema
|
ad695696a5
run on every commit on main, reuqire approval on other branches
|
11 ماه پیش |
|
Alex Cheema
|
710e5a31e7
TODO for why use_fast=False is giving inconsistent behaviour (no spaces decoding invididual tokens) for Mistral-Large-Instruct-2407-4bit
|
11 ماه پیش |
|
Alex Cheema
|
e17e5f9a41
tests for tokenizers. unfortunately use_fast=False and use_fast=True give different behaviour
|
11 ماه پیش |
|
Alex Cheema
|
0d218e244e
use fast AutoProcessor fixes #164 tokenizer issues with mistral-large.
|
11 ماه پیش |
|
Alex Cheema
|
23ae5e92c5
hold circleci tests for approval on non-main branches
|
11 ماه پیش |
|
Alex Cheema
|
d54944f4ca
stream outputs from chatgpt api integration test
|
11 ماه پیش |
 Alex Cheema
|
1133e27ad3
Merge pull request #166 from exo-explore/formatting
|
11 ماه پیش |
|
Alex Cheema
|
f53056dede
more compact operator formatting
|
11 ماه پیش |
|
Alex Cheema
|
14f2846a9c
yapf set blank_line_before_nested_class_or_def to false
|
11 ماه پیش |
|
Alex Cheema
|
ea70c9fb76
reformat with yapf format.py
|
11 ماه پیش |
|
Alex Cheema
|
2e27076665
simplify formatting with yapf
|
11 ماه پیش |
|
Alex Cheema
|
417114fae4
fix mistral nemo
|
11 ماه پیش |
|
Alex Cheema
|
5101f03369
keep 4 in RotatingKVCache
|
11 ماه پیش |
|
Alex Cheema
|
6db73fab1c
laptop gpu device capabilites
|
11 ماه پیش |
|
Alex Cheema
|
647ffb94eb
increase cli generation timeout
|
11 ماه پیش |
|
Alex Cheema
|
dd24e7db1e
only ignore CancelledError inside stop
|
11 ماه پیش |
|
Alex Cheema
|
2e1233357c
ignore CancelledError when stopping the server
|
11 ماه پیش |
|
Alex Cheema
|
dfa3fdcf08
Merge pull request #161 from exo-explore/cli
|
11 ماه پیش |
|
Alex Cheema
|
ae35ada19b
fix headless mode with --disable-tui
|
11 ماه پیش |
|
Alex Cheema
|
b95916e0b5
show prompts and outputs in tui
|
11 ماه پیش |
|
Alex Cheema
|
e84304317c
add a cli that can be triggered with --run-model <model> --prompt <prompt>
|
11 ماه پیش |
|
Alex Cheema
|
65e0488ebe
logs for file filtering, grpc_discovery -> udp_discovery
|
11 ماه پیش |
|
Alex Cheema
|
cea9b48d24
update mlx-lm to 0.17.0, use lru caches for kv_cache with RotatingKVCache to optimise memory fixes #158
|
11 ماه پیش |
