Skip to content

Commit d71e880

Browse files
yufeiwu-nvyufeiwu-nv
authored
[https://nvbugs/6055474][test] Fix RTX-6000 with wrong moe backend (#12886)
Signed-off-by: yufeiwu-nv <[email protected]>
1 parent 62885f6 commit d71e880

2 files changed

Lines changed: 0 additions & 19 deletions

File tree

‎tests/integration/defs/perf/pytorch_model_config.py‎

Lines changed: 0 additions & 19 deletions
Original file line numberDiff line numberDiff line change
@@ -93,7 +93,6 @@ def get_model_yaml_config(model_label: str,
9393
'enable_padding': False
9494
},
9595
'moe_config': {
96-
'backend': 'TRTLLM',
9796
'max_num_tokens': 32768
9897
},
9998
'speculative_config': {
@@ -228,18 +227,6 @@ def get_model_yaml_config(model_label: str,
228227
'enable_attention_dp': True,
229228
}
230229
},
231-
# Qwen3 models with fp4 quantization on B200 with moe backend equal to TRTLLM
232-
{
233-
'patterns': [
234-
'qwen3_235b_a22b_fp4-bench-pytorch-float4-maxbs:512-maxnt:2048-input_output_len:1000,2000-con:8-ep:8-gpus:8',
235-
],
236-
'config': {
237-
'enable_attention_dp': False,
238-
'moe_config': {
239-
'backend': 'TRTLLM'
240-
}
241-
}
242-
},
243230
{
244231
'patterns': [
245232
'qwen3_4b-bench-pytorch-streaming-bfloat16-maxbs:4-kv_frac:0.6-input_output_len:500,100-reqs:200-con:4',
@@ -317,9 +304,6 @@ def get_model_yaml_config(model_label: str,
317304
'enable_padding': True,
318305
'max_batch_size': 720,
319306
},
320-
'moe_config': {
321-
'backend': 'TRTLLM'
322-
},
323307
'stream_interval': 10,
324308
'num_postprocess_workers': 4
325309
}
@@ -336,9 +320,6 @@ def get_model_yaml_config(model_label: str,
336320
'enable_padding': True,
337321
'max_batch_size': 720,
338322
},
339-
'moe_config': {
340-
'backend': 'TRTLLM'
341-
},
342323
'stream_interval': 10,
343324
'num_postprocess_workers': 4
344325
}

‎tests/integration/test_lists/dev/.gitignore‎

Whitespace-only changes.

0 commit comments

Comments
 (0)