 Alex Cheema
|
0bebf8dfde
fix indent
|
1 rok temu |
|
Alex Cheema
|
55c4385db5
cleanup tmp files on failed download
|
1 rok temu |
|
Alex Cheema
|
788c49784c
retry fetch_file_list also
|
1 rok temu |
|
Alex Cheema
|
6b1c8635fc
ensure exo dir on start, retry with exp backoff on file downloads
|
1 rok temu |
|
Alex Cheema
|
e6b4f2993c
fix prompt output spacing in tui
|
1 rok temu |
|
Alex Cheema
|
3675804f4d
throttle repo progress events and only send them out if something changed
|
1 rok temu |
|
Alex Cheema
|
96f1aecb05
only in_progress if any given file is in_progress
|
1 rok temu |
|
Alex Cheema
|
23a5030604
even if part of a file is downloaded it may not be in_progress
|
1 rok temu |
|
Alex Cheema
|
31b56e862f
make a singleton thread pool executor for tinygrad since we always want it to run on the same thread
|
1 rok temu |
|
Alex Cheema
|
9f6c688d62
update tinygrad
|
1 rok temu |
|
Alex Cheema
|
4887be5103
parallelise model loading
|
1 rok temu |
|
Alex Cheema
|
141de0d011
increase chatgpt api response timeout to 900 seconds
|
1 rok temu |
 Alex Cheema
|
837ed5d980
Merge pull request #648 from exo-explore/modelasyncload
|
1 rok temu |
|
Alex Cheema
|
9c1bea97e8
fix embed_tokens for last layer in qwen models
|
1 rok temu |
|
Alex Cheema
|
af171f06fa
propagate prompts to other nodes so they can display them, cleaner prompt/output output
|
1 rok temu |
|
Alex Cheema
|
edfa53a4c2
Merge pull request #646 from exo-explore/modelasyncload
|
1 rok temu |
|
Alex Cheema
|
4a5b80a958
make sure mlx stuff is on separate thread non blocking
|
1 rok temu |
|
Alex Cheema
|
92d1bc01de
Merge pull request #645 from exo-explore/modelasyncload
|
1 rok temu |
|
Alex Cheema
|
6662d5668c
load mlx model shard on mlx thread so it doesnt block
|
1 rok temu |
|
Alex Cheema
|
a0d673fa3a
Merge pull request #640 from exo-explore/simpledownload
|
1 rok temu |
|
Alex Cheema
|
7c649085a1
fix eta/speed for resuming an existing download, using the session downloaded bytes
|
1 rok temu |
|
Alex Cheema
|
90e0e2761f
ignore not_started progress updates
|
1 rok temu |
|
Alex Cheema
|
265586f7b4
set timeout on get too
|
1 rok temu |
|
Alex Cheema
|
4748bb7dc7
increase file download timeout to 30min
|
1 rok temu |
|
Alex Cheema
|
ae770db4f3
increase download chunks to 1MB
|
1 rok temu |
|
Alex Cheema
|
82f75d0ccf
increase hf download http timeout 15 mins for large downloads
|
1 rok temu |
|
Alex Cheema
|
295f41c5cc
increase bench job timeout to give enough time to download
|
1 rok temu |
|
Alex Cheema
|
19a27c5bfd
HF_HOME -> EXO_HOME
|
1 rok temu |
|
Alex Cheema
|
d7ca9b7732
show each node id in the tinychat topology viz
|
1 rok temu |
|
Alex Cheema
|
b349e48b0d
fix visual bug where frontend would show the full hf repo size, but in some cases that includes redundant files so we should use the model index in those cases too
|
1 rok temu |
