 50h100a
|
9022c6d869
remove progress_bar imports
|
2 ماه پیش |
|
50h100a
|
9576096b9d
iterate over weights normally
|
2 ماه پیش |
 AlpinDale
|
4ec08af18b
chore: update fused MoE weight loading (#700)
|
4 ماه پیش |
|
AlpinDale
|
0e558e9b2f
fix: loading chameleon model with TP>1 (#695)
|
4 ماه پیش |
|
AlpinDale
|
3f712cd287
feat: add progress bar for loading individual weight modules (#640)
|
4 ماه پیش |
|
AlpinDale
|
f1d0b77c92
[0.6.0] Release Candidate (#481)
|
4 ماه پیش |
|
AlpinDale
|
9d81716bfd
[v0.5.3] Release Candidate (#388)
|
8 ماه پیش |
|
AlpinDale
|
f8dfac6372
chore: attention refactor and upstream sync apr01 (#365)
|
9 ماه پیش |
|
AlpinDale
|
da223153c6
feat&fix: cohere support and missing GPU blocks (#333)
|
10 ماه پیش |
|
AlpinDale
|
e42a78381a
feat: switch from pylint to ruff (#322)
|
10 ماه پیش |
|
AlpinDale
|
e31c6f0b45
feat: refactor modeling logic and support more models (#274)
|
10 ماه پیش |
|
AlpinDale
|
7d6ba53602
feat: fused top-k kernels for MoE (#273)
|
10 ماه پیش |
|
AlpinDale
|
224b87b484
feat: add fused mixtral moe support (#238)
|
10 ماه پیش |
|
AlpinDale
|
842912d022
feat: on-the-fly gguf conversion (#250)
|
11 ماه پیش |
|
AlpinDale
|
d9b65e6c5f
feat: DeepSeek MoE support (#237)
|
11 ماه پیش |
|
AlpinDale
|
c3a221eb02
feat: GGUF, QuIP#, and Marlin support (#228)
|
11 ماه پیش |
|
AlpinDale
|
8fa608aeb7
feat: replace Ray with NCCL for control plane comms (#221)
|
11 ماه پیش |
|
AlpinDale
|
97f37c1cb2
fix: use tensor parallel for quantized mixtral (#213)
|
11 ماه پیش |
 AlpinDale
|
193287b2ef
fix: mixtral unused import
|
1 سال پیش |
|
AlpinDale
|
53d391e1f2
merge 'dev' into 'main'
|
1 سال پیش |
|
AlpinDale
|
7e72ce0a73
feat: mixtral tensor parallelism (#193)
|
1 سال پیش |
|
AlpinDale
|
b9b295d74e
chore: backlogs 1 (#191)
|
1 سال پیش |
|
AlpinDale
|
f013d714c0
chore: merge dev branch into main (#177)
|
1 سال پیش |
 g4rg
|
fe57bb7ad2
feat: add rope scaling to mixtral (#174)
|
1 سال پیش |
|
AlpinDale
|
7d91e9e0f2
feat: CUDA graphs (#172)
|
1 سال پیش |
|
AlpinDale
|
725be3e0de
feat: mixtral HF with expert parallelism (#167)
|
1 سال پیش |
|
AlpinDale
|
730357c7d5
chore: implement lazy module loader for models (#165)
|
1 سال پیش |
|
AlpinDale
|
2755a48d51
merge dev branch into main (#153)
|
1 سال پیش |
|
AlpinDale
|
87277c76e4
feat: Mixtral 8x7B support (#155)
|
1 سال پیش |
