diff --git a/setup.cfg b/setup.cfg
index cf14434..c8dbc9a 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -40,7 +40,7 @@ install_requires =
     transformers>=4.32.0,<5.0.0  # if you change this, please also change version assert in petals/__init__.py
     speedtest-cli==2.1.3
     pydantic>=1.10,<2.0  # 2.0 is incompatible with hivemind yet
-    hivemind @ git+https://github.com/learning-at-home/hivemind
+    hivemind==1.1.10.post2
     tensor_parallel==1.0.23
     humanfriendly
     async-timeout>=4.0.2
diff --git a/src/petals/models/bloom/model.py b/src/petals/models/bloom/model.py
index cf83822..53e4a98 100644
--- a/src/petals/models/bloom/model.py
+++ b/src/petals/models/bloom/model.py
@@ -43,7 +43,7 @@ class DistributedBloomModel(FromPretrainedMixin, PTuneMixin, BloomModel):
         attention_mask: Optional[torch.Tensor] = None,
         head_mask: Optional[torch.LongTensor] = None,
         inputs_embeds: Optional[torch.LongTensor] = None,
-        use_cache: Optional[bool] = None,
+        use_cache: Optional[bool] = None,  # Not used here but needed for HF Transformers compatibility
         output_attentions: Optional[bool] = None,
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
@@ -63,7 +63,6 @@ class DistributedBloomModel(FromPretrainedMixin, PTuneMixin, BloomModel):
             attention_mask is None or (attention_mask == 1).all()
         ), f"Custom attention masks are not supported, {attention_mask=}"
         assert head_mask is None, f"Custom head masks are not supported, {head_mask=}"
-        assert use_cache is None or use_cache, f"{use_cache=} is not supported"
         assert not output_attentions, f"{output_attentions=} is not supported"
         assert not output_hidden_states, f"{output_hidden_states=} is not supported"
         assert return_dict is None or return_dict, f"{return_dict=} is not supported"
diff --git a/src/petals/models/llama/model.py b/src/petals/models/llama/model.py
index a9dfcc1..cf7d150 100644
--- a/src/petals/models/llama/model.py
+++ b/src/petals/models/llama/model.py
@@ -43,7 +43,7 @@ class DistributedLlamaModel(FromPretrainedMixin, PTuneMixin, LlamaModel):
         position_ids: Optional[torch.LongTensor] = None,
         past_key_values: Optional[RemotePastKeyValues] = None,
         inputs_embeds: Optional[torch.FloatTensor] = None,
-        use_cache: Optional[bool] = None,
+        use_cache: Optional[bool] = None,  # Not used here but needed for HF Transformers compatibility
         output_attentions: Optional[bool] = None,
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
@@ -65,7 +65,6 @@ class DistributedLlamaModel(FromPretrainedMixin, PTuneMixin, LlamaModel):
         assert (
             position_ids is None or (position_ids[:, 1:] - position_ids[:, :-1] == 1).all()
         ), f"Non-consecutive position_ids are not supported, {position_ids=}"
-        assert use_cache is None or use_cache, f"{use_cache=} is not supported"
         assert not output_attentions, f"{output_attentions=} is not supported"
         assert not output_hidden_states, f"{output_hidden_states=} is not supported"
         assert return_dict is None or return_dict, f"{return_dict=} is not supported"