fix: turbomind backend config in cli serve (#3784)

Author: PeymanRM

lmdeploy/cli/serve.py

@@ -359,6 +359,8 @@ def api_server(args):
                                                    cache_block_seq_len=args.cache_block_seq_len,
                                                    enable_prefix_caching=args.enable_prefix_caching,
                                                    max_prefill_token_num=args.max_prefill_token_num,
+                                                   num_tokens_per_iter=args.num_tokens_per_iter,
+                                                   max_prefill_iters=args.max_prefill_iters,
                                                    communicator=args.communicator,
                                                    hf_overrides=args.hf_overrides)
         chat_template_config = get_chat_template(args.chat_template)