[Meta] Llama4 EAGLE Support (#20591)

Signed-off-by: qizixi <qizixi@meta.com>
Co-authored-by: qizixi <qizixi@meta.com>
This commit is contained in:
zhiweiz
2025-07-15 21:14:15 -07:00
committed by GitHub
parent 1eb2b9c102
commit c11013db8b
6 changed files with 255 additions and 15 deletions

View File

@ -84,6 +84,7 @@ def main():
gpu_memory_utilization=0.8,
speculative_config=speculative_config,
disable_log_stats=False,
max_model_len=16384,
)
sampling_params = SamplingParams(temperature=args.temp, max_tokens=args.output_len)