Logic-LLM/evaluation.sh at main · Devesh-Maheshwari/Logic-LLM · GitHub

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
80
81
# 1. Navigate and activate environment
cd /users/devesh/Logic-LLM
source venv/bin/activate
source .env

# For GPT-3.5-turbo
python models/logic_program.py \
    --api_key "${OPEN_API_KEY}" \
    --dataset_name "FOLIO" \
    --split dev \
    --model_name "gpt-3.5-turbo" \
    --max_new_tokens 1024

# For GPT-4 (recommended for best results)
python models/logic_program.py \
    --api_key "${OPEN_API_KEY}" \
    --dataset_name "FOLIO" \
    --split dev \
    --model_name "gpt-4" \
    --max_new_tokens 1024


# 2. Run Logic Inference (executes logic programs with symbolic solver):
# For GPT-3.5-turbo
python models/logic_inference.py \
    --model_name "gpt-3.5-turbo" \
    --dataset_name "FOLIO" \
    --split dev \
    --backup_strategy LLM \
    --backup_LLM_result_path ./baselines/results/CoT_FOLIO_dev_gpt-3.5-turbo.json

# For GPT-4
python models/logic_inference.py \
    --model_name "gpt-4" \
    --dataset_name "FOLIO" \
    --split dev \
    --backup_strategy LLM \
    --backup_LLM_result_path ./baselines/results/CoT_FOLIO_dev_gpt-4.json

# 3. Evaluate Logic-LLM Results:

# For GPT-3.5-turbo
python models/evaluation.py \
    --dataset_name "FOLIO" \
    --model_name "gpt-3.5-turbo" \
    --split dev \
    --backup LLM

# For GPT-4
python models/evaluation.py \
    --dataset_name "FOLIO" \
    --model_name "gpt-4" \
    --split dev \
    --backup LLM

DATASET="FOLIO"
SPLIT="dev"
MODEL="gpt-4"
BACKUP="LLM"
python models/logic_inference.py \
    --model_name ${MODEL} \
    --dataset_name ${DATASET} \
    --split ${SPLIT} \
    --backup_strategy ${BACKUP} \
    --backup_LLM_result_path ./baselines/results/CoT_${DATASET}_${SPLIT}_${MODEL}.json

python models/self_refinement.py \
    --model_name "gpt-4" \
    --dataset_name "FOLIO" \
    --split dev \
    --backup_strategy "LLM" \
    --backup_LLM_result_path ./baselines/results/CoT_FOLIO_dev_gpt-4.json \
    --api_key "${OPEN_API_KEY}" \
    --maximum_rounds 3

# 4. Evaluate baseline results (need to be in baselines directory)
cd baselines
python evaluation.py --dataset_name "FOLIO" --model_name "gpt-3.5-turbo" --split "dev" --mode "Direct" | grep "EM:"
python evaluation.py --dataset_name "FOLIO" --model_name "gpt-3.5-turbo" --split "dev" --mode "CoT" | grep "EM:"
python evaluation.py --dataset_name "FOLIO" --model_name "gpt-4" --split "dev" --mode "Direct" | grep "EM:"
python evaluation.py --dataset_name "FOLIO" --model_name "gpt-4" --split "dev" --mode "CoT" | grep "EM:"