A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 16:00 UTC
[Misc] format output for encoder_decoder.py (#16095)
Signed-off-by: reidliu41 <reid201711@gmail.com> Co-authored-by: reidliu41 <reid201711@gmail.com>
Reid committed
Apr 6, 2025 at 10:57 UTC
d8f094a92aad5b102a798de7d6da1454972620c7
1 file changed
+6
-5
examples/offline_inference/encoder_decoder.py
+6
-5
@@ -75,8 +75,6 @@ prompts = [
75
enc_dec_prompt1, enc_dec_prompt2, enc_dec_prompt3
76
] + zipped_prompt_list
77
78
-print(prompts)
79
-
78
# Create a sampling params object.
79
sampling_params = SamplingParams(
80
temperature=0,
@@ -91,10 +89,13 @@ sampling_params = SamplingParams(
89
outputs = llm.generate(prompts, sampling_params)
90
91
# Print the outputs.
94
-for output in outputs:
92
+print("-" * 50)
93
+for i, output in enumerate(outputs):
94
prompt = output.prompt
95
encoder_prompt = output.encoder_prompt
96
generated_text = output.outputs[0].text
98
- print(f"Encoder prompt: {encoder_prompt!r}, "
99
- f"Decoder prompt: {prompt!r}, "
97
+ print(f"Output {i+1}:")
98
+ print(f"Encoder prompt: {encoder_prompt!r}\n"
99
+ f"Decoder prompt: {prompt!r}\n"
100
f"Generated text: {generated_text!r}")
101
+ print("-" * 50)