 Alex Cheema
|
edfa53a4c2
Merge pull request #646 from exo-explore/modelasyncload
|
7 mēneši atpakaļ |
 Alex Cheema
|
4a5b80a958
make sure mlx stuff is on separate thread non blocking
|
7 mēneši atpakaļ |
|
Alex Cheema
|
92d1bc01de
Merge pull request #645 from exo-explore/modelasyncload
|
7 mēneši atpakaļ |
|
Alex Cheema
|
6662d5668c
load mlx model shard on mlx thread so it doesnt block
|
7 mēneši atpakaļ |
|
Alex Cheema
|
a0d673fa3a
Merge pull request #640 from exo-explore/simpledownload
|
7 mēneši atpakaļ |
|
Alex Cheema
|
7c649085a1
fix eta/speed for resuming an existing download, using the session downloaded bytes
|
7 mēneši atpakaļ |
|
Alex Cheema
|
90e0e2761f
ignore not_started progress updates
|
7 mēneši atpakaļ |
|
Alex Cheema
|
265586f7b4
set timeout on get too
|
7 mēneši atpakaļ |
|
Alex Cheema
|
4748bb7dc7
increase file download timeout to 30min
|
7 mēneši atpakaļ |
|
Alex Cheema
|
ae770db4f3
increase download chunks to 1MB
|
7 mēneši atpakaļ |
|
Alex Cheema
|
82f75d0ccf
increase hf download http timeout 15 mins for large downloads
|
7 mēneši atpakaļ |
|
Alex Cheema
|
295f41c5cc
increase bench job timeout to give enough time to download
|
7 mēneši atpakaļ |
|
Alex Cheema
|
19a27c5bfd
HF_HOME -> EXO_HOME
|
7 mēneši atpakaļ |
|
Alex Cheema
|
d7ca9b7732
show each node id in the tinychat topology viz
|
7 mēneši atpakaļ |
|
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
|
7 mēneši atpakaļ |
|
Alex Cheema
|
21586063f6
use llama-3.2-1b in tinygrad test
|
7 mēneši atpakaļ |
|
Alex Cheema
|
277d63d860
special case when a model doesnt have a model index file, then use wildcard for allow_patterns
|
7 mēneši atpakaļ |
|
Alex Cheema
|
74379ef671
log download logs with DEBUG>=6 very verbose
|
7 mēneši atpakaļ |
|
Alex Cheema
|
3c7bd48aa3
get rid of some more hf bloat
|
7 mēneši atpakaļ |
|
Alex Cheema
|
1df023023e
remove a lot of hf bloat
|
7 mēneši atpakaļ |
|
Alex Cheema
|
b89495f444
rewrite ShardDownloader, simplify significantly
|
7 mēneši atpakaļ |
|
Alex Cheema
|
903950f64e
Merge pull request #638 from exo-explore/deepseekv3fix
|
7 mēneši atpakaļ |
|
Alex Cheema
|
a3766f538a
add exception for mlx-community/DeepSeek-R1-3bit and mlx-community/DeepSeek-V3-3bit in tokenizers test
|
7 mēneši atpakaļ |
|
Alex Cheema
|
9711d632e0
Merge pull request #637 from exo-explore/deepseekv3fix
|
7 mēneši atpakaļ |
|
Alex Cheema
|
82ef086010
add deepseek-v3-3bit and deepseek-r1-3bit
|
7 mēneši atpakaļ |
|
Alex Cheema
|
55ea366932
fix post_init deepseek v3
|
7 mēneši atpakaļ |
|
Alex Cheema
|
63318983de
Merge pull request #631 from sigseg5/main
|
7 mēneši atpakaļ |
 sigseg5
|
fb841a1f50
Adjust truncate size in history list for text without any spaces
|
7 mēneši atpakaļ |
|
sigseg5
|
4512366580
Fix bubble behavior when user passes long text without any spaces
|
7 mēneši atpakaļ |
|
sigseg5
|
9525c0e7a7
Add adaptive padding for user and assistant messages on width <= 1480px
|
7 mēneši atpakaļ |
