Skip to content

[Cria][Lllama runner] Use caching temp allocator #3974

[Cria][Lllama runner] Use caching temp allocator

[Cria][Lllama runner] Use caching temp allocator #3974

Triggered via pull request December 4, 2025 15:46
Status Success
Total duration 1h 11m 27s
Artifacts 11

cuda.yml

on: pull_request
Matrix: export-model-cuda-artifact
Matrix: test-cuda-builds
Matrix: test-models-cuda
Matrix: test-model-cuda-e2e
check-all-cuda-builds
3s
check-all-cuda-builds
Fit to window
Zoom out
Zoom in

Artifacts

Produced during runtime
Name Size Digest
google-gemma-3-4b-it-cuda-non-quantized Expired
7.22 GB
sha256:9c6c747b25491a82facda55132cd0e24d32de2122e3c74e2e7df6803064ae3fd
google-gemma-3-4b-it-cuda-quantized-int4-tile-packed Expired
4.03 GB
sha256:c31f3202a38bcb8171263659abd609134845acb3af2015bb9b6021c9198d3d37
mistralai-Voxtral-Mini-3B-2507-cuda-non-quantized Expired
6.82 GB
sha256:f4221b2ac6ccb2dc85c8af62c8670adbe440634cc2471638d4e087c63b933972
mistralai-Voxtral-Mini-3B-2507-cuda-quantized-int4-tile-packed Expired
2.89 GB
sha256:ef6656ff7a86aecdb6abae8110a935aac0f1e3f592b17656f793052b569adcf5
mistralai-Voxtral-Mini-3B-2507-cuda-quantized-int4-weight-only Expired
6.14 GB
sha256:8e49698c19bd6425b48b875da5c02aceab8f67e44ab88a4b8e618a7461c6ce23
openai-whisper-large-v3-turbo-cuda-non-quantized Expired
1.17 GB
sha256:46ccbb7358288c0a9e3476a178ba0e905decad709dfe269ef6c507639afac52f
openai-whisper-large-v3-turbo-cuda-quantized-int4-tile-packed Expired
490 MB
sha256:2c3c745d5ec3d0f3ab2c919bdd92fb97e23bf0e6f8828f9a9a902bd2ce67baa6
openai-whisper-large-v3-turbo-cuda-quantized-int4-weight-only Expired
484 MB
sha256:8d505ca0f14557620826ba8b4df45a0857188c4480a53f0d48e49a2706a03b67
openai-whisper-small-cuda-non-quantized Expired
361 MB
sha256:b53e300ab0b7c2178e916383b58ca6cf86b89eefc36af67cceba11b2693557a5
openai-whisper-small-cuda-quantized-int4-tile-packed Expired
172 MB
sha256:c1de7ae12583fb3a258bf2056c80dce4fa7224fbb19fc0690cb7541de4c7768d
openai-whisper-small-cuda-quantized-int4-weight-only Expired
270 MB
sha256:bbe5f7ffa5289d5fd8e68f35d0fb2c21073e8f384d48c7bad315f6f90291bbe5