-
Notifications
You must be signed in to change notification settings
- Fork 1.1k
Expand file tree
/
Copy patheval.py
More file actions
79 lines (72 loc) · 2.73 KB
/
Copy patheval.py
File metadata and controls
79 lines (72 loc) · 2.73 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
# Copyright (c) 2025, EleutherAI
# This file is based on code by the authors denoted below and has been modified from its original version.
#
# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Evaluation tasks - modified from https://github.com/EleutherAI/lm-evaluation-harness"""
import os
import sys
sys.path.append(
os.path.abspath(os.path.join(os.path.dirname(__file__), os.path.pardir))
)
from megatron.training import forward_step
from megatron.utils import setup_for_inference_or_eval, init_wandb
from megatron.logging import tb_wandb_log
from eval_tasks import run_eval_harness
from pprint import pprint
from datetime import datetime
import json
def main(input_args=None, overwrite_values=None):
model, neox_args = setup_for_inference_or_eval(
use_cache=False, input_args=input_args, overwrite_values=overwrite_values
)
results = run_eval_harness(
model,
forward_step,
neox_args,
eval_tasks=neox_args.eval_tasks,
bootstrap_iters=10000,
)
if neox_args.rank == 0:
init_wandb(neox_args=neox_args)
# log to wandb
for k, v in results["results"].items():
if isinstance(v, dict):
for k2, v2 in v.items():
k3 = "_".join([k, k2])
tb_wandb_log(
f"eval/{k3}",
v2,
neox_args.iteration,
use_wandb=neox_args.use_wandb,
comet_experiment=neox_args.comet_experiment,
)
else:
tb_wandb_log(
f"eval/{k}",
v,
neox_args.iteration,
use_wandb=neox_args.use_wandb,
comet_experiment=neox_args.comet_experiment,
)
pprint(results)
results_path = (
f'eval_results_{datetime.now().strftime("%m-%d-%Y-%H-%M-%S")}.json'
)
if neox_args.eval_results_prefix:
results_path = f"{neox_args.eval_results_prefix}_{results_path}"
with open(results_path, "w") as f:
json.dump(results, f, indent=4)
if __name__ == "__main__":
main()