From ef79bea04d48ac6f433837640bec9660e0741760 Mon Sep 17 00:00:00 2001 From: Nikita Savelyev Date: Fri, 28 Feb 2025 16:32:20 +0100 Subject: [PATCH] Add processing of int4 configs for model id aliases --- optimum/intel/openvino/configuration.py | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/optimum/intel/openvino/configuration.py b/optimum/intel/openvino/configuration.py index 5a444bcc10..0cdd9d8f87 100644 --- a/optimum/intel/openvino/configuration.py +++ b/optimum/intel/openvino/configuration.py @@ -238,6 +238,15 @@ class OVQuantizationMethod(str, Enum): }, } +# Add configs for model id aliases +# The list below contains pairs of model ids: config for the second model id will be copied from the first model id. +model_id_aliases = [ + ("meta-llama/Meta-Llama-3.1-8B-Instruct", "meta-llama/Llama-3.1-8B-Instruct"), + ("meta-llama/Meta-Llama-3.1-8B", "meta-llama/Llama-3.1-8B"), +] +for m_id_1, m_id_2 in model_id_aliases: + _DEFAULT_4BIT_CONFIGS[m_id_2] = _DEFAULT_4BIT_CONFIGS[m_id_1] + _DEFAULT_4BIT_CONFIG = { "bits": 4, "ratio": 1.0,