 AlpinDale
|
89a2c6dee1
chore: refactor `MultiModalConfig` initialization and profiling (#745)
|
3 月之前 |
|
AlpinDale
|
81c28d2a7f
fix: use nvml to get consistent device names (#739)
|
3 月之前 |
|
AlpinDale
|
b03fa02397
refactor: base worker input refactor for multi-step (#683)
|
4 月之前 |
|
AlpinDale
|
bf88c8567e
feat: mamba model support (#674)
|
4 月之前 |
|
AlpinDale
|
62111fab17
feat: allow serving encoder-decoder models in the API server (#664)
|
4 月之前 |
|
AlpinDale
|
c147670c13
fix: clean up incorrect log in worker (#636)
|
4 月之前 |
|
AlpinDale
|
a0e446a17d
feat: initial encoder-decoder support with BART model (#633)
|
4 月之前 |
|
AlpinDale
|
f1d0b77c92
[0.6.0] Release Candidate (#481)
|
4 月之前 |
|
AlpinDale
|
9d81716bfd
[v0.5.3] Release Candidate (#388)
|
8 月之前 |
|
AlpinDale
|
e3252edd07
fix: remove event and stream, add typing (#382)
|
9 月之前 |
|
AlpinDale
|
f8dfac6372
chore: attention refactor and upstream sync apr01 (#365)
|
9 月之前 |
|
AlpinDale
|
e42a78381a
feat: switch from pylint to ruff (#322)
|
10 月之前 |
|
AlpinDale
|
9810daa699
feat: INT8 KV Cache (#298)
|
10 月之前 |
|
AlpinDale
|
4b80b42362
fix: memory leaks due to nccl cuda graphs (#275)
|
10 月之前 |
 Thomas Xin
|
43cf0e98a0
fix: worker initialization on WSL (#260)
|
10 月之前 |
|
AlpinDale
|
ea0f57b233
feat: allow further support for non-cuda devices (#247)
|
11 月之前 |
|
AlpinDale
|
31c95011a6
feat: FP8 E5M2 KV Cache (#226)
|
11 月之前 |
|
AlpinDale
|
641bb0f6e9
feat: add custom allreduce kernels (#224)
|
11 月之前 |
|
AlpinDale
|
c0aac15421
feat: S-LoRA support (#222)
|
11 月之前 |
|
AlpinDale
|
8fa608aeb7
feat: replace Ray with NCCL for control plane comms (#221)
|
11 月之前 |
|
AlpinDale
|
15a0454172
feat: FP8 KV Cache (#185)
|
1 年之前 |
|
AlpinDale
|
7d91e9e0f2
feat: CUDA graphs (#172)
|
1 年之前 |
|
AlpinDale
|
f5f9bc6a7c
fix: memory profiling (#166)
|
1 年之前 |
|
AlpinDale
|
653da510d1
chore: rewrite InputMetadata (#143)
|
1 年之前 |
|
AlpinDale
|
1aab8a7d6f
feat: speedup compilation times by 3x (#130)
|
1 年之前 |
|
AlpinDale
|
237d2ec28d
fix: CPU OOM for large models (#128)
|
1 年之前 |
|
AlpinDale
|
e7b6a2d5a0
chore: tensor parallel refactors part 2 (#116)
|
1 年之前 |
|
AlpinDale
|
f384f3ae60
fix: force v2 for ctxlen larger than 8192 (#100)
|
1 年之前 |
 AlpinDale
|
ae7d8df224
fix lint issues (again)
|
1 年之前 |
 50h100a
|
fa0ae5a2c9
feat: new mirostatv2 implementation (#96)
|
1 年之前 |
