Another fix

src/petals/models/mixtral/model.py

@@ -158,6 +158,11 @@ class DistributedMixtralForCausalLM(FromPretrainedMixin, RemoteGenerationMixin,
 
 
 class DistributedMixtralForSequenceClassification(FromPretrainedMixin, MixtralForSequenceClassification):
+    _keys_to_ignore_on_load_missing = DistributedMixtralModel._keys_to_ignore_on_load_missing
+    _keys_to_ignore_on_load_unexpected = DistributedMixtralModel._keys_to_ignore_on_load_unexpected
+
+    config_class = DistributedMixtralConfig
+
     def __init__(self, config: DistributedMixtralConfig):
         MixtralPreTrainedModel.__init__(self, config)
         self.num_labels = config.num_labels