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.

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

Modify the check_results.py to support batch 2&4 #11133

Merged
merged 45 commits into from
Jun 5, 2024

Conversation

MargarettMao
Copy link
Contributor

Modify the check_results.py. Counting batch_size when checking the total test cases. Also comment the 'batch_szie == 4' exclude cases to support the checking-result process.

.github/workflows/llm_performance_tests.yml Show resolved Hide resolved
Comment on lines 10 to 11
batch_size: 1 # default to 1
batch_size:
- 1 # default to 1
Copy link
Contributor

Choose a reason for hiding this comment

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

revert this?

@@ -23,7 +23,8 @@ warm_up: 1
num_trials: 3
num_beams: 1 # default to greedy search
low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
batch_size: 1 # default to 1
batch_size: # default to 1
Copy link
Contributor

Choose a reason for hiding this comment

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

revert this file

@@ -9,11 +9,13 @@ warm_up: 1
num_trials: 3
num_beams: 1 # default to greedy search
low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
batch_size: 1 # default to 1
batch_size: # default to 1
Copy link
Contributor

Choose a reason for hiding this comment

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

revert this file

current_csv_model_input_2nd=current_csv_model+'-'+current_csv_input_output_pairs+'-'+'2nd'
add_to_dict(csv_dict, current_csv_model_input_1st, current_csv_row[latency_1st_token])
add_to_dict(csv_dict, current_csv_model_input_2nd, current_csv_row[latency_2_avg])
try:
Copy link
Contributor

Choose a reason for hiding this comment

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

under what case this try will fail?

for batch_size in conf["batch_size"]:
global csv_name
batch = str(batch_size)
csv_name = f'{current_dir}/test_batch{batch}/{api}-results-{today}-batch-{batch}.csv'
Copy link
Contributor

Choose a reason for hiding this comment

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

better not add a folder in this file.

Comment on lines 1826 to 1827
batch = str(conf["batch_size"])
csv_name = f'{current_dir}/{api}-results-{today}-batch-{batch}.csv' # mwj don't add directly / change the directory
Copy link
Contributor

Choose a reason for hiding this comment

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

do not modify the csv name
Support multiple batch in one config

@@ -17,3 +17,4 @@ in_out_pairs:
test_api:
- "transformer_int4_gpu" # on Intel GPU
cpu_embedding: False # whether put embedding to CPU (only avaiable now for gpu win related test_api)
exclude:
Copy link
Contributor

Choose a reason for hiding this comment

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

revert this if no exclude?

Comment on lines 35 to 36
- 'baichuan-inc/Baichuan2-13B-Chat-4bit:2048:1'
- 'bigscience/bloomz-7b1:2048:1'
Copy link
Contributor

@hkvision hkvision Jun 3, 2024

Choose a reason for hiding this comment

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

revert this? make run.py compatible with the original yaml.

Comment on lines 108 to 109
except NameError:
pass
Copy link
Contributor

Choose a reason for hiding this comment

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

add a comment here

cd ../../../../../
cp python/llm/test/benchmark/arc-perf-test-batch2.yaml python/llm/dev/benchmark/all-in-one/config.yaml
cd python/llm/dev/benchmark/all-in-one
# changr csv name
Copy link
Contributor

Choose a reason for hiding this comment

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

typo change

@@ -36,3 +36,4 @@ exclude:
# - 'fnlp/moss-moon-003-sft-4bit:2048'
- 'baichuan-inc/Baichuan2-13B-Chat-4bit:2048'
- 'bigscience/bloomz-7b1:2048'

Copy link
Contributor

Choose a reason for hiding this comment

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

revert this file?

else:
batch_list = conf["batch_size"]
for batch_size in batch_list:
csv_name = f'{current_dir}/{api}-results-{today}.csv'
Copy link
Contributor

Choose a reason for hiding this comment

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

put this out of the for loop?

@hkvision hkvision merged commit 231b968 into intel-analytics:main Jun 5, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants