Reduce chunk_size and add timeout to fix issue (#1822)

* Fix issue
* Add timeout for UT script.

Signed-off-by: ZePan110 <ze.pan@intel.com>

---------

Signed-off-by: ZePan110 <ze.pan@intel.com>

Author: ZePan110

tests/llms/test_llms_doc-summarization_vllm_on_intel_hpu.sh

@@ -135,23 +135,23 @@ function validate_microservices() {
         'text' \
         "docsum-vllm-gaudi" \
         "docsum-vllm-gaudi" \
-        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5.", "max_tokens":32, "language":"en", "summary_type": "truncate", "chunk_size": 2000}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5.", "max_tokens":32, "language":"en", "summary_type": "truncate", "chunk_size": 1500, "timeout":200}'
 
     echo "Validate map_reduce mode..."
     validate_services \
         "$URL" \
         'text' \
         "docsum-vllm-gaudi" \
         "docsum-vllm-gaudi" \
-        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5.", "max_tokens":32, "language":"en", "summary_type": "map_reduce", "chunk_size": 2000, "stream":false, "timeout":200}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5.", "max_tokens":32, "language":"en", "summary_type": "map_reduce", "chunk_size": 1500, "stream":false, "timeout":200}'
 
     echo "Validate refine mode..."
     validate_services \
         "$URL" \
         'text' \
         "docsum-vllm-gaudi" \
         "docsum-vllm-gaudi" \
-        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5.", "max_tokens":32, "language":"en", "summary_type": "refine", "chunk_size": 2000, "timeout":200}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5.", "max_tokens":32, "language":"en", "summary_type": "refine", "chunk_size": 1500, "timeout":200}'
 }
 
 function stop_docker() {