[Frontend] Separate pooling APIs in offline inference (#11129)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2024-12-13 10:40:07 +00:00
committed by GitHub
parent f93bf2b189
commit eeec9e3390
21 changed files with 669 additions and 304 deletions
+11 -5
View File
@@ -9,14 +9,20 @@ prompts = [
]
# Create an LLM.
# You should pass task="embed" for embedding models
model = LLM(
model="intfloat/e5-mistral-7b-instruct",
task="embed", # You should pass task="embed" for embedding models
task="embed",
enforce_eager=True,
)
# Generate embedding. The output is a list of PoolingRequestOutputs.
outputs = model.encode(prompts)
# Generate embedding. The output is a list of EmbeddingRequestOutputs.
outputs = model.embed(prompts)
# Print the outputs.
for output in outputs:
print(output.outputs.embedding) # list of 4096 floats
for prompt, output in zip(prompts, outputs):
embeds = output.outputs.embedding
embeds_trimmed = ((str(embeds[:16])[:-1] +
", ...]") if len(embeds) > 16 else embeds)
print(f"Prompt: {prompt!r} | "
f"Embeddings: {embeds_trimmed} (size={len(embeds)})")