]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
bert : add tests + fix quantization (#5475)
authorGeorgi Gerganov <redacted>
Tue, 13 Feb 2024 11:01:29 +0000 (13:01 +0200)
committerGitHub <redacted>
Tue, 13 Feb 2024 11:01:29 +0000 (13:01 +0200)
commit49cc1f7d67de2da99f3ac185f9ff1319b7bf35f8
tree46e731e1f972752260cf470cc226a05ad9cfea6a
parent99b8b43d7b185a6483f28cf798a2d968b2e16ca7
bert : add tests + fix quantization (#5475)

* llama : do not quantize pos embd and token type tensors

* ci : add BERT tests

ggml-ci

* ci : do not do BERT tests on low-perf nodes

ggml-ci
ci/run.sh
llama.cpp