Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Server: add test for num slots, fails on master #6950

Merged
merged 1 commit into from
May 1, 2024
Merged

Server: add test for num slots, fails on master #6950

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
88 changes: 56 additions & 32 deletions examples/server/tests/features/results.feature
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,44 +7,16 @@ Feature: Results
And a model file tinyllamas/split/stories15M-00001-of-00003.gguf from HF repo ggml-org/models
And a model file test-model-00001-of-00003.gguf
And 128 as batch size
And 256 KV cache size
And 1024 KV cache size
And 128 max tokens to predict
And continuous batching
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Minor: continuous batching is enabled by default (and cannot be disabled BTW :) )


Scenario Outline: Multi users completion
Scenario Outline: consistent results with same seed
Given <n_slots> slots
And continuous batching
Then the server is starting
Then the server is healthy

Given 42 as seed
And a prompt:
"""
Write a very long story about AI.
"""

Given 42 as seed
And a prompt:
"""
Write a very long story about AI.
"""

Given 42 as seed
And a prompt:
"""
Write a very long story about AI.
"""

Given 42 as seed
And a prompt:
"""
Write a very long story about AI.
"""

Given 42 as seed
And a prompt:
"""
Write a very long story about AI.
"""
Given 4 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 42

Given concurrent completion requests
Then the server is busy
Expand All @@ -55,3 +27,55 @@ Feature: Results
| n_slots |
| 1 |
| 2 |

Scenario Outline: different results with different seed
Given <n_slots> slots
Then the server is starting
Then the server is healthy

Given 1 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 42
Given 1 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 43
Given 1 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 44
Given 1 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 45

Given concurrent completion requests
Then the server is busy
Then the server is idle
And all slots are idle
Then all predictions are different
Examples:
| n_slots |
| 1 |
| 2 |

Scenario Outline: consistent results with same seed and varying batch size
Given 4 slots
And <temp> temperature
# And 0 as draft
Then the server is starting
Then the server is healthy

Given 1 prompts "Write a very long story about AI." with seed 42
And concurrent completion requests
# Then the server is busy # Not all slots will be utilized.
Then the server is idle
And all slots are idle

Given <n_parallel> prompts "Write a very long story about AI." with seed 42
And concurrent completion requests
# Then the server is busy # Not all slots will be utilized.
Then the server is idle
And all slots are idle

Then all predictions are equal
Examples:
| n_parallel | temp |
| 1 | 0.0 |
| 2 | 0.0 |
| 4 | 0.0 |
| 1 | 1.0 |
# FIXME: These tests fail on master. The problem seems to be the unified KV cache.
# See https://github.com/ggerganov/whisper.cpp/issues/1941#issuecomment-1986923227
# and https://github.com/ggerganov/llama.cpp/pull/6122#discussion_r1531405574 .
# | 2 | 1.0 |
# | 4 | 1.0 |
Loading
Loading