apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

[FDConfig]Remove total_block_num/dtype/block_size/enc_dec_block_num in ParallelConfig (#4400)

Browse Source 
* delete some attr in parallel config

* delete comment

---------

Co-authored-by: root <root@yqlcc01-sys-rpm12rzmwjd.yqlcc01.baidu.com>
This commit is contained in:
YuanRisheng
2025-10-16 20:00:37 +08:00
committed by GitHub
parent b87e2c6184
commit 0355235fb9
19 changed files with 51 additions and 66 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/worker/xpu_worker.py
+2 -2
View File
@@ -56,7 +56,7 @@ class XpuWorker(WorkerBase):
self.device_ids = self.parallel_config.device_ids.split(",")
self.device = f"xpu:{self.local_rank % self.max_chips_per_node}"
paddle.device.set_device(self.device)
paddle.set_default_dtype(self.parallel_config.dtype)
paddle.set_default_dtype(self.model_config.dtype)
gc.collect()
paddle.device.xpu.empty_cache()
@@ -122,7 +122,7 @@ class XpuWorker(WorkerBase):
used_memory = xpu_get_used_global_memory(int(self.device_ids[self.local_rank]))
available_kv_cache_memory = total_available_memory - used_memory
model_block_memory_used = self.cal_theortical_kvcache()
available_kv_cache_memory += model_block_memory_used * self.parallel_config.total_block_num
available_kv_cache_memory += model_block_memory_used * self.cache_config.total_block_num
if self.parallel_config.use_ep:
available_kv_cache_memory = int(available_kv_cache_memory * 0.6)