 AlpinDale
|
c20073824a
cleanup
|
пре 7 месеци |
|
AlpinDale
|
85a865cc00
feat: fp8 quant
|
пре 7 месеци |
|
AlpinDale
|
8be299e78b
fix: lora load check
|
пре 7 месеци |
|
AlpinDale
|
096d9eb6c5
enhance nvlink detection
|
пре 7 месеци |
|
AlpinDale
|
fb7825df8f
squash logprobs
|
пре 7 месеци |
|
AlpinDale
|
66b7bc4415
sliding window in prefix kernel
|
пре 7 месеци |
|
AlpinDale
|
42998e423c
better quant verification
|
пре 7 месеци |
|
AlpinDale
|
483c95a2f8
fix ops in gptq and awq
|
пре 7 месеци |
|
AlpinDale
|
8f9cb7235c
chore: allow multiple served model names
|
пре 7 месеци |
|
AlpinDale
|
fc80f57967
fix: correct file name for qwen2 moe
|
пре 7 месеци |
|
AlpinDale
|
f894f7b176
Revert "reduce dedupe by wrapping in general worker class"
|
пре 8 месеци |
|
AlpinDale
|
082b0b03bc
Revert "actually run the workers"
|
пре 8 месеци |
|
AlpinDale
|
36cf32649d
actually run the workers
|
пре 8 месеци |
|
AlpinDale
|
9fff6fb507
reduce dedupe by wrapping in general worker class
|
пре 8 месеци |
|
AlpinDale
|
b92bddafe9
time.monotonic() -> time.time()
|
пре 8 месеци |
 AlpinDale
|
0178b4d976
docker: add AWS Neuron Docker image
|
пре 8 месеци |
|
AlpinDale
|
949f0445de
readme: update installation command
|
пре 8 месеци |
 Naomiusearch
|
893158a4ed
fix: quants installation on ROCm (#469)
|
пре 8 месеци |
|
AlpinDale
|
5ee79a1692
readme: update for 0.5.3
|
пре 8 месеци |
|
AlpinDale
|
9d81716bfd
[v0.5.3] Release Candidate (#388)
|
пре 8 месеци |
 Krovius
|
205c8e4106
fix: kobold api /tokencount (#424)
|
пре 9 месеци |
 50h100a
|
f663d3fccc
Merge pull request #397 from 50h100a/pr_samplerasserts
|
пре 9 месеци |
 50h100a
|
85ae23ac3c
Missed .items() and assert
|
пре 9 месеци |
|
50h100a
|
43c9858854
Merge pull request #244 from PygmalionAI/faster_topk
|
пре 9 месеци |
|
50h100a
|
1da7fd64bc
Merge pull request #396 from 50h100a/pr_notneuron
|
пре 9 месеци |
|
50h100a
|
0634b8a3a6
fix memory pinning conditional
|
пре 9 месеци |
|
50h100a
|
a0b21a3e42
Merge pull request #358 from 50h100a/pr_samplers
|
пре 9 месеци |
|
50h100a
|
bd564148e2
Merge branch 'main' of https://github.com/PygmalionAI/aphrodite-engine into ffs
|
пре 9 месеци |
|
50h100a
|
d3dd170a7d
merge main
|
пре 9 месеци |
|
AlpinDale
|
78d66f16d1
Chunked Prefill Part 1 (#384)
|
пре 9 месеци |
