Spaces:
Runtime error
Runtime error
meg-huggingface
commited on
Commit
·
c5729e2
1
Parent(s):
258cdcb
Changing names to AIEnergyScore
Browse files- Dockerfile +1 -1
- create_results.py +17 -13
- entrypoint.sh +2 -2
- failed_run.py +3 -3
- parse_requests.py +4 -6
- pause_space.py +1 -1
Dockerfile
CHANGED
|
@@ -76,4 +76,4 @@ RUN --mount=type=secret,id=DEBUG,mode=0444,required=true \
|
|
| 76 |
git init && \
|
| 77 |
git remote add origin $(cat /run/secrets/DEBUG)
|
| 78 |
|
| 79 |
-
|
|
|
|
| 76 |
git init && \
|
| 77 |
git remote add origin $(cat /run/secrets/DEBUG)
|
| 78 |
|
| 79 |
+
ENTRYPOINT ["/entrypoint.sh"]
|
create_results.py
CHANGED
|
@@ -10,28 +10,32 @@ api = HfApi(token=TOKEN)
|
|
| 10 |
|
| 11 |
out_dir = sys.argv[1]
|
| 12 |
|
| 13 |
-
#Uploading results
|
| 14 |
api.upload_folder(
|
| 15 |
folder_path=out_dir,
|
| 16 |
repo_id="AIEnergyScore/results_debug",
|
| 17 |
repo_type="dataset",
|
| 18 |
)
|
| 19 |
|
| 20 |
-
#Updating requests
|
| 21 |
-
requests = load_dataset("AIEnergyScore/requests_debug", split="test",
|
|
|
|
| 22 |
requests_dset = requests.to_pandas()
|
| 23 |
|
| 24 |
-
models_ran=[]
|
| 25 |
for f in os.scandir(out_dir):
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
|
|
|
| 32 |
print("Models ran are: " + str(models_ran))
|
| 33 |
|
| 34 |
-
requests_dset.loc[
|
| 35 |
-
|
| 36 |
-
updated_dset
|
|
|
|
|
|
|
| 37 |
print("Updated model status")
|
|
|
|
| 10 |
|
| 11 |
out_dir = sys.argv[1]
|
| 12 |
|
| 13 |
+
# Uploading results
|
| 14 |
api.upload_folder(
|
| 15 |
folder_path=out_dir,
|
| 16 |
repo_id="AIEnergyScore/results_debug",
|
| 17 |
repo_type="dataset",
|
| 18 |
)
|
| 19 |
|
| 20 |
+
# Updating requests
|
| 21 |
+
requests = load_dataset("AIEnergyScore/requests_debug", split="test",
|
| 22 |
+
token=TOKEN)
|
| 23 |
requests_dset = requests.to_pandas()
|
| 24 |
|
| 25 |
+
models_ran = []
|
| 26 |
for f in os.scandir(out_dir):
|
| 27 |
+
if f.is_dir():
|
| 28 |
+
for s in os.scandir(f):
|
| 29 |
+
if s.is_dir() and s.name not in ['hooks', 'info', 'objects', 'refs',
|
| 30 |
+
'logs']:
|
| 31 |
+
for m in os.scandir(s):
|
| 32 |
+
models_ran.append(s.name + '/' + m.name)
|
| 33 |
+
|
| 34 |
print("Models ran are: " + str(models_ran))
|
| 35 |
|
| 36 |
+
requests_dset.loc[
|
| 37 |
+
requests_dset["model"].isin(models_ran), ['status']] = "COMPLETED"
|
| 38 |
+
updated_dset = Dataset.from_pandas(requests_dset)
|
| 39 |
+
updated_dset.push_to_hub("AIEnergyScore/requests_debug", split="test",
|
| 40 |
+
token=TOKEN)
|
| 41 |
print("Updated model status")
|
entrypoint.sh
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
#!/bin/bash
|
| 2 |
|
| 3 |
-
export SPACE="
|
| 4 |
|
| 5 |
echo "Not checking h100 -- already know it's not there."
|
| 6 |
#python /check_h100.py
|
|
@@ -31,4 +31,4 @@ python /upload_run_folder.py --run_dir "/runs"
|
|
| 31 |
# Pausing space
|
| 32 |
echo "Pausing space."
|
| 33 |
python /pause_space.py
|
| 34 |
-
echo "Done."
|
|
|
|
| 1 |
#!/bin/bash
|
| 2 |
|
| 3 |
+
export SPACE="AIEnergyScore/launch-computation-example"
|
| 4 |
|
| 5 |
echo "Not checking h100 -- already know it's not there."
|
| 6 |
#python /check_h100.py
|
|
|
|
| 31 |
# Pausing space
|
| 32 |
echo "Pausing space."
|
| 33 |
python /pause_space.py
|
| 34 |
+
echo "Done."
|
failed_run.py
CHANGED
|
@@ -26,7 +26,7 @@ parser.add_argument(
|
|
| 26 |
args = parser.parse_args()
|
| 27 |
|
| 28 |
# Updating request
|
| 29 |
-
dataset = load_dataset("
|
| 30 |
token=TOKEN).to_pandas()
|
| 31 |
|
| 32 |
# Set benchmark to failed
|
|
@@ -48,5 +48,5 @@ except FileNotFoundError as e:
|
|
| 48 |
print(f"Could not find {args.run_dir}/error.log")
|
| 49 |
|
| 50 |
updated_dataset = Dataset.from_pandas(dataset)
|
| 51 |
-
updated_dataset.push_to_hub("
|
| 52 |
-
token=TOKEN)
|
|
|
|
| 26 |
args = parser.parse_args()
|
| 27 |
|
| 28 |
# Updating request
|
| 29 |
+
dataset = load_dataset("AIEnergyScore/requests_debug", split="test",
|
| 30 |
token=TOKEN).to_pandas()
|
| 31 |
|
| 32 |
# Set benchmark to failed
|
|
|
|
| 48 |
print(f"Could not find {args.run_dir}/error.log")
|
| 49 |
|
| 50 |
updated_dataset = Dataset.from_pandas(dataset)
|
| 51 |
+
updated_dataset.push_to_hub("AIEnergyScore/requests_debug", split="test",
|
| 52 |
+
token=TOKEN)
|
parse_requests.py
CHANGED
|
@@ -1,10 +1,10 @@
|
|
| 1 |
import os
|
| 2 |
from datasets import load_dataset, Dataset
|
| 3 |
|
| 4 |
-
skip_statuses = ['COMPLETED', 'FAILED', 'RUNNING']
|
| 5 |
TOKEN = os.environ.get("DEBUG")
|
| 6 |
requests_dataset = load_dataset("AIEnergyScore/requests_debug", split="test")
|
| 7 |
|
|
|
|
| 8 |
def normalize_task(task):
|
| 9 |
# Makes assumption about how the task names are being written, and called.
|
| 10 |
return '_'.join(task.split()).lower()
|
|
@@ -12,9 +12,7 @@ def normalize_task(task):
|
|
| 12 |
|
| 13 |
requests_dset = requests_dataset.to_pandas()
|
| 14 |
|
| 15 |
-
for model, task in requests_dset[['model','task']].loc[
|
|
|
|
|
|
|
| 16 |
print("%s,%s" % (model, normalize_task(task)))
|
| 17 |
-
|
| 18 |
-
# Custom errors we will rerun.
|
| 19 |
-
for model, task in requests_dset[['model','task']].loc[~requests_dset['status'].isin(skip_statuses)].to_dict(orient= 'split', index=False)['data']:
|
| 20 |
-
print("%s,%s" % (model, normalize_task(task)))
|
|
|
|
| 1 |
import os
|
| 2 |
from datasets import load_dataset, Dataset
|
| 3 |
|
|
|
|
| 4 |
TOKEN = os.environ.get("DEBUG")
|
| 5 |
requests_dataset = load_dataset("AIEnergyScore/requests_debug", split="test")
|
| 6 |
|
| 7 |
+
|
| 8 |
def normalize_task(task):
|
| 9 |
# Makes assumption about how the task names are being written, and called.
|
| 10 |
return '_'.join(task.split()).lower()
|
|
|
|
| 12 |
|
| 13 |
requests_dset = requests_dataset.to_pandas()
|
| 14 |
|
| 15 |
+
for model, task in requests_dset[['model', 'task']].loc[
|
| 16 |
+
requests_dset['status'] == 'PENDING'].to_dict(orient='split', index=False)[
|
| 17 |
+
'data']:
|
| 18 |
print("%s,%s" % (model, normalize_task(task)))
|
|
|
|
|
|
|
|
|
|
|
|
pause_space.py
CHANGED
|
@@ -3,4 +3,4 @@ import os
|
|
| 3 |
|
| 4 |
TOKEN = os.environ.get("DEBUG")
|
| 5 |
API = HfApi(token=TOKEN)
|
| 6 |
-
API.pause_space("
|
|
|
|
| 3 |
|
| 4 |
TOKEN = os.environ.get("DEBUG")
|
| 5 |
API = HfApi(token=TOKEN)
|
| 6 |
+
API.pause_space("AIEnergyScore/launch-computation-example")
|