vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

server: #5655 - continue to update other slots on embedding concurrent request.

Browse source 
server: tests: add multi users embeddings as fixed
This commit is contained in:
Pierrick HYMBERT 2024-02-24 13:01:48 +01:00
parent 525213d2f5
commit 09b77b4c9e
3 changed files with 25 additions and 34 deletions
Download patch file Download diff file Expand all files Collapse all files

2
examples/server/server.cpp
View file

@ -1836,7 +1836,7 @@ struct llama_server_context
send_embedding(slot);
slot.release();
slot.i_batch = -1;
return true;
continue;
}
completion_token_output result;

34
examples/server/tests/features/issues.feature
View file

@ -1,36 +1,4 @@
# List of ongoing issues
@bug
Feature: Issues
# Issue #5655
Scenario: Multi users embeddings
Given a server listening on localhost:8080
And a model file stories260K.gguf
And a model alias tinyllama-2
And 42 as server seed
And 64 KV cache size
And 2 slots
And continuous batching
And embeddings extraction
Then the server is starting
Then the server is healthy
Given a prompt:
"""
Write a very long story about AI.
"""
And a prompt:
"""
Write another very long music lyrics.
"""
And a prompt:
"""
Write a very long poem.
"""
And a prompt:
"""
Write a very long joke.
"""
Given concurrent embedding requests
Then the server is busy
Then the server is idle
Then all embeddings are generated
# No confirmed issue at the moment

23
examples/server/tests/features/parallel.feature
View file

@ -8,6 +8,7 @@ Feature: Parallel
And 42 as server seed
And 64 KV cache size
And 2 slots
And embeddings extraction
And continuous batching
Then the server is starting
Then the server is healthy
@ -75,3 +76,25 @@ Feature: Parallel
Then the server is busy
Then the server is idle
Then all prompts are predicted
Scenario: Multi users embeddings
Given a prompt:
"""
Write a very long story about AI.
"""
And a prompt:
"""
Write another very long music lyrics.
"""
And a prompt:
"""
Write a very long poem.
"""
And a prompt:
"""
Write a very long joke.
"""
Given concurrent embedding requests
Then the server is busy
Then the server is idle
Then all embeddings are generated