Commit from ustream PR vllm-project#7746

Validate the that the input prompts aren't empty

This avoids an async loop crash that takes down the server

Signed-off-by: Max de Bayser <mbayser@br.ibm.com>
Signed-off-by: Jefferson Fialho <jfialho@ibm.com>

vllm/engine/async_llm_engine.py

@@ -543,6 +543,7 @@ async def process_model_inputs_async(
                 inputs,
                 request_id=request_id,
             )
+            self._validate_enc_dec_inputs(model_inputs)
         else:
             if is_explicit_encoder_decoder_prompt(inputs):
                 raise ValueError("Cannot pass encoder-decoder prompt "
@@ -555,6 +556,7 @@ async def process_model_inputs_async(
                 lora_request=lora_request,
                 prompt_adapter_request=prompt_adapter_request,
             )
+            self._validate_dec_only_inputs(model_inputs)
 
         return self.input_processor(model_inputs)
 

vllm/engine/llm_engine.py

@@ -946,6 +946,7 @@ def process_model_inputs(
                 inputs,
                 request_id=request_id,
             )
+            self._validate_enc_dec_inputs(model_inputs)
         else:
             if is_explicit_encoder_decoder_prompt(inputs):
                 raise ValueError("Cannot pass encoder-decoder prompt "
@@ -958,6 +959,7 @@ def process_model_inputs(
                 lora_request=lora_request,
                 prompt_adapter_request=prompt_adapter_request,
             )
+            self._validate_dec_only_inputs(model_inputs)
 
         return self.input_processor(model_inputs)
 
@@ -1631,3 +1633,13 @@ def is_encoder_decoder_model(self):
 
     def is_embedding_model(self):
         return self.model_config.is_embedding_model
+
+    def _validate_dec_only_inputs(self, inputs: LLMInputs):
+        if "prompt_token_ids" not in inputs or len(
+                inputs["prompt_token_ids"]) == 0:
+            raise ValueError("Empty prompt")
+
+    def _validate_enc_dec_inputs(self, inputs: EncoderDecoderLLMInputs):
+        if "encoder_prompt_token_ids" not in inputs or\
+              len(inputs["encoder_prompt_token_ids"]) == 0:
+            raise ValueError("Empty prompt")