TGI benchmark with llmperf #564
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dacorvo
force-pushed
the
tgi-benchmark-with-llmperf
branch
from
b6696a4 to
b8f310f
Compare
michaelbenayoun
approved these changes
Apr 12, 2024
| @@ -0,0 +1,29 @@ | |||
| #!/bin/bash | |||
|
|
|||
| model=${1:-NousResearch/Llama-2-7b-chat-hf} | |||
There was a problem hiding this comment.
Do we really want a default model for the benchmark?
There was a problem hiding this comment.
It also serves as an explaination of the args that can be passed
Comment on lines
7
to
24
| filenames = glob.glob("tgi_bench_results/*/*summary.json") | ||
|
|
||
| results = [] | ||
|
|
||
| for filename in filenames: | ||
| with open(filename) as f: | ||
| summary = json.load(f) | ||
| d = { | ||
| "model_id": summary["model"], | ||
| "concurrent requests": summary["num_concurrent_requests"], | ||
| "throughput (t/s)": summary["results_mean_output_throughput_token_per_s"], | ||
| "Time-to-first-token @ P50 (s)": summary["results_ttft_s_quantiles_p50"], | ||
| "average latency (ms)": summary["results_inter_token_latency_s_quantiles_p50"] * 1000, | ||
| } | ||
| results.append(pd.DataFrame.from_dict(d, orient="index").transpose()) | ||
|
|
||
| df = pd.concat(results).sort_values(by="concurrent requests") | ||
| df.to_csv("tgi-results.csv", index=False) |
There was a problem hiding this comment.
nit: I would just guard that with a if __name__ == "__main__".
| MODEL_ID='NousResearch/Llama-2-7b-chat-hf' | ||
| HF_BATCH_SIZE=32 | ||
| HF_SEQUENCE_LENGTH=4096 | ||
| HF_AUTO_CAST_TYPE='fp16' |
Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
This adds scripts to test TGI deployments using several TGI servers on the same host and a load-balancer to achieve Data Parallelism.
The test client is llmperf.
It also includes results for LLama 7b and Mistral v2 deployed on a inf2.48xlarge in a DP3 TP8 configuration.