polish code with new pre-commit rule (#2923)

2025-12-24 13:28:13 +08:00 · 2025-07-19 23:19:27 +08:00
parent b8676d71a8
commit 25698d56d1
424 changed files with 14307 additions and 13518 deletions
--- a/benchmarks/yaml/eb45-21B-vl-128k-wint4-h800-tp1.yaml
+++ b/benchmarks/yaml/eb45-21B-vl-128k-wint4-h800-tp1.yaml
@@ -7,4 +7,4 @@ tensor_parallel_size: 1
 enable_chunked_prefill: True
 max_num_batched_tokens: 384
 quantization: wint4
-reasoning_parser: ernie-45-vl
+reasoning_parser: ernie-45-vl
--- a/benchmarks/yaml/eb45-32k-w4a8c8-tp4_decode.yaml
+++ b/benchmarks/yaml/eb45-32k-w4a8c8-tp4_decode.yaml
@@ -12,4 +12,4 @@ rdma_comm_ports: "7671,7672,7673,7674"
 pd_comm_port: "2334"
 max_num_batched_tokens: 384
 max_num_partial_prefills: 3
-max_long_partial_prefills: 3
+max_long_partial_prefills: 3
--- a/benchmarks/yaml/eb45-32k-w4a8c8-tp4_prefill.yaml
+++ b/benchmarks/yaml/eb45-32k-w4a8c8-tp4_prefill.yaml
@@ -9,4 +9,4 @@ cache_queue_port: 55664
 engine_worker_queue_port: 6677
 cache_transfer_protocol: "rdma,ipc"
 rdma_comm_ports: "7675,7676,7677,7678"
-pd_comm_port: "2333"
+pd_comm_port: "2333"
--- a/benchmarks/yaml/eb45-32k-wint4-h800-dp8_prefill.yaml
+++ b/benchmarks/yaml/eb45-32k-wint4-h800-dp8_prefill.yaml
@@ -10,4 +10,4 @@ engine_worker_queue_port: 6677
 num_gpu_blocks_override: 1024
 cache_transfer_protocol: "rdma"
 rdma_comm_ports: "7671,7672,7673,7674,7675,7676,7677,7678"
-pd_comm_port: "2334"
+pd_comm_port: "2334"
--- a/benchmarks/yaml/eb45-32k-wint4-mtp-tp4-decode.yaml
+++ b/benchmarks/yaml/eb45-32k-wint4-mtp-tp4-decode.yaml
@@ -10,4 +10,4 @@ splitwise_role: decode
 engine_worker_queue_port: 6678
 cache_transfer_protocol: "rdma,ipc"
 rdma_comm_ports: "7671,7672,7673,7674"
-pd_comm_port: "2334"
+pd_comm_port: "2334"
--- a/benchmarks/yaml/eb45-32k-wint4-mtp-tp4-prefill.yaml
+++ b/benchmarks/yaml/eb45-32k-wint4-mtp-tp4-prefill.yaml
@@ -9,4 +9,4 @@ cache_queue_port: 55664
 engine_worker_queue_port: 6677
 cache_transfer_protocol: "rdma,ipc"
 rdma_comm_ports: "7675,7676,7677,7678"
-pd_comm_port: "2333"
+pd_comm_port: "2333"
--- a/benchmarks/yaml/eb45-32k-wint4-tp4_decode.yaml
+++ b/benchmarks/yaml/eb45-32k-wint4-tp4_decode.yaml
@@ -12,4 +12,4 @@ rdma_comm_ports: "7671,7672,7673,7674"
 pd_comm_port: "2334"
 max_num_batched_tokens: 384
 max_num_partial_prefills: 3
-max_long_partial_prefills: 3
+max_long_partial_prefills: 3
--- a/benchmarks/yaml/eb45-32k-wint4-tp4_prefill.yaml
+++ b/benchmarks/yaml/eb45-32k-wint4-tp4_prefill.yaml
@@ -9,4 +9,4 @@ cache_queue_port: 55664
 engine_worker_queue_port: 6677
 cache_transfer_protocol: "rdma,ipc"
 rdma_comm_ports: "7675,7676,7677,7678"
-pd_comm_port: "2333"
+pd_comm_port: "2333"
--- a/benchmarks/yaml/qwen3moe235b-32k-wint4-h800-tp4.yaml
+++ b/benchmarks/yaml/qwen3moe235b-32k-wint4-h800-tp4.yaml
@@ -3,4 +3,4 @@ max_num_seqs: 75
 gpu_memory_utilization: 0.85
 kv_cache_ratio: 0.75
 quantization: wint4
-tensor_parallel_size: 4
+tensor_parallel_size: 4
--- a/benchmarks/yaml/qwen3moe235b-32k-wint8-h800-tp4.yaml
+++ b/benchmarks/yaml/qwen3moe235b-32k-wint8-h800-tp4.yaml
@@ -3,4 +3,4 @@ max_num_seqs: 25
 gpu_memory_utilization: 0.9
 kv_cache_ratio: 0.75
 quantization: wint8
-tensor_parallel_size: 4
+tensor_parallel_size: 4
--- a/benchmarks/yaml/request_yaml/quick_benchmark.yaml
+++ b/benchmarks/yaml/request_yaml/quick_benchmark.yaml
@@ -1,3 +1,3 @@
 metadata:
  min_tokens: 32
-max_tokens: 33
+max_tokens: 33
--- a/benchmarks/yaml/request_yaml/qwen2-32k.yaml
+++ b/benchmarks/yaml/request_yaml/qwen2-32k.yaml
@@ -5,4 +5,4 @@ metadata:
 max_tokens: 12288
 repetition_penalty: 1.05
 frequency_penalty: 0
-presence_penalty: 0
+presence_penalty: 0
--- a/benchmarks/yaml/request_yaml/qwen3-32k.yaml
+++ b/benchmarks/yaml/request_yaml/qwen3-32k.yaml
@@ -5,4 +5,4 @@ metadata:
 max_tokens: 12288
 repetition_penalty: 1.0
 frequency_penalty: 0
-presence_penalty: 1.5
+presence_penalty: 1.5
--- a/benchmarks/yaml/request_yaml/vLLM_default.yaml
+++ b/benchmarks/yaml/request_yaml/vLLM_default.yaml
@@ -8,4 +8,4 @@ frequency_penalty: 0
 presence_penalty: 0
 skip_special_tokens: false
 chat_template_kwargs:
-  enable_thinking: true
+  enable_thinking: true
--- a/benchmarks/yaml/x1-32k-wint8-p800-tp8.yaml
+++ b/benchmarks/yaml/x1-32k-wint8-p800-tp8.yaml
@@ -3,4 +3,4 @@ max_num_seqs: 64
 gpu_memory_utilization: 0.9
 tensor_parallel_size: 8
 quantization: wint8
-reasoning_parser: ernie-x1
+reasoning_parser: ernie-x1