[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

comps/llms/local/llm_hf.py

@@ -12,20 +12,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import sys
 import os
+import sys
+from typing import List
 
+import lm_eval.api.registry
+import torch
+from docarray import BaseDoc
 from fastapi.responses import StreamingResponse
+from GenAIEval.evaluation.lm_evaluation_harness.lm_eval.models.huggingface import HFLM, GaudiHFModelAdapter
 from langchain_community.llms import HuggingFaceEndpoint
 
 from comps import GeneratedDoc, ServiceType, opea_microservices, opea_telemetry, register_microservice
 
-from docarray import BaseDoc
-from typing import List
-
-import torch
-import lm_eval.api.registry
-from GenAIEval.evaluation.lm_evaluation_harness.lm_eval.models.huggingface import HFLM, GaudiHFModelAdapter
 lm_eval.api.registry.MODEL_REGISTRY["hf"] = HFLM
 lm_eval.api.registry.MODEL_REGISTRY["gaudi-hf"] = GaudiHFModelAdapter
 
@@ -42,12 +41,12 @@ class LLMCompletionDoc(BaseDoc):
 device = os.getenv("DEVICE", "")
 
 llm = lm_eval.api.registry.get_model(model).create_from_arg_string(
-        model_args,
-        {
-            "batch_size": 1, # dummy
-            "max_batch_size": None,
-            "device": device,
-        },
+    model_args,
+    {
+        "batch_size": 1,  # dummy
+        "max_batch_size": None,
+        "device": device,
+    },
 )
 
 
@@ -58,13 +57,10 @@ class LLMCompletionDoc(BaseDoc):
     host="0.0.0.0",
     port=9000,
 )
-
 @opea_telemetry
 def llm_generate(input: LLMCompletionDoc):
     global llm
-    batched_inputs = torch.tensor(input.batched_inputs,
-            dtype=torch.long,
-            device=llm.device)
+    batched_inputs = torch.tensor(input.batched_inputs, dtype=torch.long, device=llm.device)
     with torch.no_grad():
         # TODO, use model.generate.
         logits = llm.model(batched_inputs).logits
@@ -74,9 +70,12 @@ def llm_generate(input: LLMCompletionDoc):
     greedy_tokens = logits.argmax(dim=-1)
     logprobs = torch.gather(logits, 2, batched_inputs[:, 1:].unsqueeze(-1)).squeeze(-1)
 
-    return {"greedy_tokens": greedy_tokens.detach().cpu().tolist(),
-            "logprobs": logprobs.detach().cpu().tolist(),
-            "batched_inputs": input.batched_inputs}
+    return {
+        "greedy_tokens": greedy_tokens.detach().cpu().tolist(),
+        "logprobs": logprobs.detach().cpu().tolist(),
+        "batched_inputs": input.batched_inputs,
+    }
+
 
 if __name__ == "__main__":
     opea_microservices["opea_service@llm_hf"].start()