[XPU] Update doc and add scripts for downloading dependencies (#2845)

* [XPU] update xvllm download

* update supported models

* fix xpu model runner in huge memory with small model

* update doc

fastdeploy/worker/xpu_model_runner.py

@@ -774,7 +774,6 @@ class XPUModelRunner(ModelRunnerBase):
         del self.share_inputs["caches"]
         if self.forward_meta is not None:
             del self.forward_meta.caches
-        del self.share_inputs["block_tables"]
         paddle.device.xpu.empty_cache()
 
     def cal_theortical_kvcache(self):
@@ -817,11 +816,6 @@ class XPUModelRunner(ModelRunnerBase):
         # Reset block table and kv cache with global block num
         self.initialize_kv_cache()
 
-        self.share_inputs["block_tables"] = paddle.full(
-            [self.parallel_config.max_num_seqs, self.num_gpu_blocks],
-            -1,
-            dtype="int32")
-
         # Reset free list
         free_list = list(
             range(